Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

First steps in developing clinical practice guidelines for post-acute rehabilitation after primary total… Westby, Marie D. 2010

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2010_spring_westby_marie.pdf [ 2.62MB ]
Metadata
JSON: 24-1.0070954.json
JSON-LD: 24-1.0070954-ld.json
RDF/XML (Pretty): 24-1.0070954-rdf.xml
RDF/JSON: 24-1.0070954-rdf.json
Turtle: 24-1.0070954-turtle.txt
N-Triples: 24-1.0070954-rdf-ntriples.txt
Original Record: 24-1.0070954-source.json
Full Text
24-1.0070954-fulltext.txt
Citation
24-1.0070954.ris

Full Text

 FIRST STEPS IN DEVELOPING CLINICAL PRACTICE GUIDELINES FOR POST-ACUTE REHABILITATION AFTER PRIMARY TOTAL HIP AND KNEE ARTHROPLASTY  by  Marie D. Westby BSc. (PT), University of British Columbia, 1988   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Rehabilitation Sciences)    THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2010 © Marie D. Westby, 2010 ii Abstract Total hip (THA) and total knee arthroplasty (TKA) are cost-effective interventions for advanced osteoarthritis (OA) of the knee and hip. With the rapidly growing number of these procedures performed annually in Canada and the United States, greater attention needs to be directed to identify rehabilitation practices that optimize outcomes and minimize cost. Currently, there is no consensus on rehabilitation best practice and no evidence-based clinical practice guidelines to inform clinical decision-making on post-acute rehabilitation following THA and TKA. A multi-phase, mixed-method project integrated stakeholder perspectives, research evidence and expert opinion to develop best practice recommendations for THA and TKA rehabilitation. Chapter 2 involved 11 focus groups and eight interviews to identify key themes related from North American patients and health care professionals on rehabilitation practices and outcomes. Chapters 3-4 are Cochrane systematic reviews examining the strength of the evidence for post- acute physiotherapy after THA and TKA. Chapter 5-6 involved two parallel Delphi surveys with consumers, clinicians and researchers to develop consensus on a range of rehabilitation topics to inform best practice for THA and TKA rehabilitation. Chapter 2: Six key themes emerged relating to communication, patient expectations, patient attitude, forms of support, barriers to recovery, and diversity of outcomes. Chapters 3-4: Systematic reviews of THA (n=8) and TKA (n=7) trials revealed limited, low to high quality evidence with mixed findings for various forms of post-acute physiotherapy on pain, function and health-related quality of life. Trial heterogeneity prevented meta-analysis. Chapters 5-6: Consensus (80% agreement) was reached on the need for post-acute rehabilitation, types of interventions, rehabilitation providers, treatment settings, outcomes and outcome measurement. Consensus was not reached regarding timing and dosage of rehabilitation. Sub-group analysis revealed few differences comparing responses by profession, primary role and country. iii This thesis has taken important first steps in identifying appropriate rehabilitation interventions and health care resources to optimize individuals‟ activity, participation and health-related quality of life after THA and TKA.  Further, it highlights the need for more high quality research to address the knowledge gaps and inform policy on this important and understudied aspect of arthroplasty surgery. iv Table of contents  Abstract .......................................................................................................................................... ii Table of contents .......................................................................................................................... iv List of tables.................................................................................................................................. ix List of figures ................................................................................................................................. x List of abbreviations .................................................................................................................... xi Acknowledgements .................................................................................................................... xiii Dedication ................................................................................................................................... xvi Co-authorship statement .......................................................................................................... xvii Chapter 1: Introduction and overview of rehabilitation after total hip and knee arthroplasty ................................................................................................................................... 1 1.1 Background .......................................................................................................................... 1 1.2 Relevance ............................................................................................................................. 2 1.3 Rehabilitation practices ........................................................................................................ 4 1.4 Functional impairments, activity limitations and participation restrictions following THA and TKA ............................................................................................................................... 5 1.5 Canadian versus American health care systems ................................................................... 6 1.6 Clinical practice guidelines .................................................................................................. 7 1.7 Guideline development framework ...................................................................................... 8 1.8 Other guidelines initiatives ................................................................................................... 9 1.9 Purpose ............................................................................................................................... 10 1.10 Overview of thesis chapters ............................................................................................. 11 1.11 References ........................................................................................................................ 13 Chapter 2: Patient and healthcare professional views on total hip and knee replacement rehabilitation and outcomes: A focus group study .................................................................. 20 2.1 Background ........................................................................................................................ 20 2.2 Purpose ............................................................................................................................... 21 2.3 Methods .............................................................................................................................. 21 2.3.1 Sampling frame ............................................................................................................ 21 2.3.2 Recruitment .................................................................................................................. 22 2.3.3 Focus groups/interviews ............................................................................................... 22 2.3.4 Ethics ............................................................................................................................ 23 2.3.5 Data analysis ................................................................................................................. 24 2.4 Results ................................................................................................................................ 25 2.4.1 Key themes ................................................................................................................... 27 2.4.1.1 Theme 1: Let‟s talk .............................................................................................................. 27 2.4.1.2 Theme 2: Expecting the unexpected .................................................................................... 29 2.4.1.3 Theme 3: It‟s attitude that counts ........................................................................................ 31 2.4.1.4 Theme 4: It takes all kinds of support ................................................................................. 31 2.4.1.5 Theme 5: Barriers to recovery ............................................................................................. 33 2.4.1.6 Theme 6: Back to normal .................................................................................................... 36 2.5 Discussion .......................................................................................................................... 38 2.5.1 Strengths of the study ................................................................................................... 43 2.5.2 Limitations .................................................................................................................... 43 v 2.5.3 Clinical implications ..................................................................................................... 44 2.5.4 Future research directions ............................................................................................. 45 2.6 Conclusions ........................................................................................................................ 45 2.7 References .......................................................................................................................... 47 Chapter 3: Post-acute physiotherapy following primary total hip arthroplasty for osteoarthritis: A Cochrane systematic review .......................................................................... 52 3.1 Background ........................................................................................................................ 52 3.2 Purpose ............................................................................................................................... 55 3.2.1 Objectives ..................................................................................................................... 55 3.3 Methods .............................................................................................................................. 55 3.3.1 Criteria for considering studies for this review ............................................................ 55 3.3.2 Types of participants .................................................................................................... 55 3.3.3 Types of interventions .................................................................................................. 56 3.3.3.1 Experimental ....................................................................................................................... 56 3.3.3.2 Control ................................................................................................................................. 57 3.3.4 Types of outcome measures ......................................................................................... 58 3.3.5 Search methods for identification of studies ................................................................ 59 3.3.6 Data collection and analysis ......................................................................................... 60 3.3.7 Assessment of methodological quality ......................................................................... 60 3.3.8 Data analysis and synthesis .......................................................................................... 61 3.4 Results ................................................................................................................................ 63 3.4.1 Description of studies ................................................................................................... 63 3.4.2 Methodological quality ................................................................................................. 65 3.4.3 Participants ................................................................................................................... 66 3.4.4 Interventions ................................................................................................................. 70 3.4.5 Controls ........................................................................................................................ 70 3.4.6 Outcomes ...................................................................................................................... 71 3.4.7 Risk of bias in included studies .................................................................................... 72 3.4.8 Effects of interventions................................................................................................. 73 3.4.8.1 Pain ...................................................................................................................................... 73 3.4.8.2 Function ............................................................................................................................... 74 3.4.8.3 Health-related quality of life ............................................................................................... 80 3.5 Discussion .......................................................................................................................... 81 3.5.1 Summary of main results .............................................................................................. 81 3.5.2 Quality of the evidence ................................................................................................. 82 3.5.3 Strengths of this review ................................................................................................ 82 3.5.4 Weaknesses and potential biases of this review ........................................................... 83 3.5.5 Agreements and disagreements with other studies or reviews ..................................... 83 3.6 Conclusions ........................................................................................................................ 85 3.6.1 Implications for practice ............................................................................................... 85 3.6.2 Implications for research .............................................................................................. 86 3.7 References .......................................................................................................................... 88 Chapter 4: Post-acute physiotherapy following primary total knee replacement for osteoarthritis: A Cochrane systematic review .......................................................................... 93 4.1 Background ........................................................................................................................ 93 4.2 Purpose ............................................................................................................................... 96 4.2.1 Objectives ..................................................................................................................... 96 4.3 Methods .............................................................................................................................. 96 4.3.1 Criteria for considering studies for this review ............................................................ 96 vi 4.3.2 Types of participants .................................................................................................... 97 4.3.3 Types of interventions .................................................................................................. 97 4.3.3.1 Experimental ....................................................................................................................... 97 4.3.3.2 Control ................................................................................................................................. 99 4.3.4 Types of outcome measures ......................................................................................... 99 4.3.5 Search methods for identification of studies .............................................................. 100 4.3.6 Data collection and analysis ....................................................................................... 101 4.3.7 Assessment of methodological quality ....................................................................... 102 4.3.8 Data analysis and synthesis ........................................................................................ 103 4.4 Results .............................................................................................................................. 104 4.4.1 Description of studies ................................................................................................. 104 4.4.2 Methodological quality ............................................................................................... 106 4.4.3 Participants ................................................................................................................. 107 4.4.4 Interventions ............................................................................................................... 111 4.4.5 Controls ...................................................................................................................... 111 4.4.6 Outcomes .................................................................................................................... 112 4.4.7 Risk of bias in included studies .................................................................................. 113 4.4.8 Effects of interventions............................................................................................... 114 4.4.8.1 Pain .................................................................................................................................... 115 4.4.8.2 Function ............................................................................................................................. 116 4.4.8.3 HRQoL .............................................................................................................................. 121 4.5 Discussion ........................................................................................................................ 123 4.5.1 Summary of main results ............................................................................................ 123 4.5.2 Quality of the evidence ............................................................................................... 124 4.5.3 Strengths of this review .............................................................................................. 124 4.5.4 Weaknesses and potential biases of this review ......................................................... 125 4.5.5 Agreements and disagreements with other studies or reviews ................................... 125 4.6 Conclusions ...................................................................................................................... 127 4.6.1 Implications for practice ............................................................................................. 127 4.6.2 Implications for research ............................................................................................ 129 4.7 References ........................................................................................................................ 131 Chapter 5: Developing consensus on best practice recommendations for total hip arthroplasty rehabilitation: A Delphi study ........................................................................... 138 5.1 Background ...................................................................................................................... 138 5.1.1 Clinical practice guidelines ........................................................................................ 139 5.1.2 Rationale for Delphi survey ........................................................................................ 140 5.2 Purpose ............................................................................................................................. 141 5.3 Methods ............................................................................................................................ 141 5.3.1 Participants ................................................................................................................. 141 5.3.2 The Delphi questionnaires (rounds) ........................................................................... 142 5.3.3 Procedures .................................................................................................................. 143 5.3.4 Ethics approval ........................................................................................................... 144 5.3.5 Data collection and analysis ....................................................................................... 144 5.4 Results .............................................................................................................................. 147 5.4.1 Panelist demographics ................................................................................................ 147 5.4.2 Response rates ............................................................................................................ 148 5.4.3 Response times ........................................................................................................... 149 5.4.4 Round one results ....................................................................................................... 149 5.4.5 Round two results ....................................................................................................... 154 vii 5.4.6 Round 3 results ........................................................................................................... 154 5.4.7 Results by survey sections .......................................................................................... 156 5.4.7.1 Rehabilitation phases ......................................................................................................... 156 5.4.7.2 Need for post-acute rehabilitation ..................................................................................... 156 5.4.7.3 Rehabilitation providers .................................................................................................... 156 5.4.7.4 Rehabilitation format ......................................................................................................... 157 5.4.7.5 Timing of post-acute rehabilitation ................................................................................... 158 5.4.7.6 Rehabilitation setting ......................................................................................................... 159 5.4.7.7 Rehabilitation interventions ............................................................................................... 159 5.4.7.8 Dosage of rehabilitation .................................................................................................... 161 5.4.7.9 Rehabilitation outcomes .................................................................................................... 162 5.4.7.10 Rehabilitation outcome measurement ............................................................................. 164 5.4.7.11 Follow-up care ................................................................................................................. 165 5.4.8 Thematic analysis of comments ................................................................................. 167 5.5 Discussion ........................................................................................................................ 168 5.5.1 Strengths and limitations ............................................................................................ 171 5.6 Conclusion ........................................................................................................................ 174 5.6.1 Clinical implications ................................................................................................... 174 5.6.2 Implications for future research.................................................................................. 176 5.7 References ........................................................................................................................ 179 Chapter 6: Developing consensus on best practice recommendations for total knee arthroplasty rehabilitation: A Delphi study ........................................................................... 183 6.1 Background ...................................................................................................................... 183 6.1.1 Clinical practice guidelines and the Delphi method ................................................... 184 6.2 Purpose ............................................................................................................................. 185 6.3 Methods ............................................................................................................................ 185 6.3.1 Participants ................................................................................................................. 185 6.3.2 Sampling, recruitment and panel composition ........................................................... 185 6.3.3 The Delphi questionnaires (rounds) ........................................................................... 186 6.3.4 Ethics .......................................................................................................................... 187 6.3.5 Data collection and analysis ....................................................................................... 187 6.4 Results .............................................................................................................................. 188 6.4.1 Panelist demographics ................................................................................................ 188 6.4.2 Response rates ............................................................................................................ 189 6.4.3 Response times ........................................................................................................... 190 6.4.4 Round one results ....................................................................................................... 190 6.4.5 Round two results ....................................................................................................... 191 6.4.6 Round three results ..................................................................................................... 195 6.4.7 Results by sections ..................................................................................................... 197 6.4.7.1 Rehabilitation phases ......................................................................................................... 197 6.4.7.2 Need for post-acute rehabilitation ..................................................................................... 197 6.4.7.3 Rehabilitation providers .................................................................................................... 197 6.4.7.4 Rehabilitation format ......................................................................................................... 198 6.4.7.5 Timing of rehabilitation ..................................................................................................... 199 6.4.7.6 Rehabilitation setting ......................................................................................................... 200 6.4.7.7 Rehabilitation interventions ............................................................................................... 200 6.4.7.8 Dosage of rehabilitation .................................................................................................... 202 6.4.7.9 Rehabilitation outcomes .................................................................................................... 203 6.4.7.10 Rehabilitation outcome measurement ............................................................................. 205 6.4.7.11 Follow-up care ................................................................................................................. 206 6.4.8 Thematic analysis of comments ................................................................................. 208 viii 6.5 Discussion ........................................................................................................................ 209 6.5.1 Strengths and limitations of research ......................................................................... 212 6.6 Conclusion ........................................................................................................................ 213 6.6.1 Clinical implications ................................................................................................... 214 6.6.2 Implications for future research.................................................................................. 216 6.7 References ........................................................................................................................ 217 Chapter 7: Conclusions and implications for future research.............................................. 221 7.1 Summary of thesis objectives ........................................................................................... 221 7.2 Summary of study findings .............................................................................................. 224 7.3 Implications for clinical practice ...................................................................................... 227 7.4 Implications for future research ....................................................................................... 229 7.5 Implications for health care policy ................................................................................... 231 7.6 Strengths and limitations .................................................................................................. 233 7.7 Next steps ......................................................................................................................... 235 7.8 Thesis contribution ........................................................................................................... 238 7.9 References ........................................................................................................................ 239 Appendix A: Multidisciplinary advisory group ..................................................................... 243 Appendix B: Focus group key questions ................................................................................. 244 Appendix C: UBC Ethics certificates of approval ................................................................. 246 Appendix D: Sample search strategy for THA ...................................................................... 249 Appendix E: Sample search strategy for TKA....................................................................... 251 Appendix F: Letter of introduction for professional organizations ..................................... 253 Appendix G: Letter of introduction for consumer organizations ........................................ 254 Appendix H: Delphi questionnaire sections ........................................................................... 255 Appendix I: Delphi round 3 questionnaire ............................................................................. 256 Appendix J: THA Delphi panelists .......................................................................................... 293 Appendix K: Delphi panelists by geographic location (THA and TKA) ............................. 294 Appendix L: TKA Delphi panelists ......................................................................................... 295 Appendix M: Professional and consumer organizations’ Delphi participation .................. 296 ix List of tables Table 2.1 Patient participant demographics .................................................................................. 26 Table 2.2 Health professional participant demographics .............................................................. 26 Table 3.1 Quality Assessment and Risk of Bias Checklist ........................................................... 61 Table 3.2 GRADE approach to assessing overall levels of quality of a body of evidence .......... 62 Table 3.3 Excluded studies and main reason for exclusion .......................................................... 65 Table 3.4 Methodological quality and risk of bias in included trials ............................................ 66 Table 3.5 Summary of study characteristics ................................................................................. 67 Table 4.1 Quality Assessment and Risk of Bias Checklist ......................................................... 102 Table 4.2 GRADE approach to assessing overall levels of quality of a body of evidence ........ 103 Table 4.3 Excluded studies and main reason for exclusion ........................................................ 106 Table 4.4 Methodological quality and risk of bias of included trials ......................................... 107 Table 4.5 Summary of study characteristics ............................................................................... 109 Table 5.1 Sample controlled feedback ........................................................................................ 146 Table 5.2 THA panelists‟ demographics .................................................................................... 148 Table 5.3 THA panelists‟ response rates by rounds ................................................................... 149 Table 5.4 Level of agreement by Delphi round .......................................................................... 151 Table 5.5 Influence of personal and external factors by section ................................................. 155 Table 5.6 Appropriate and important post-acute rehabilitation interventions after THA........... 160 Table 5.7 Important outcomes to routinely assess and/or monitor after THA ............................ 163 Table 5.8 Measures and tools considered feasible and important for routine clinical outcome evaluation and/or monitoring after primary THA ........................................................ 165 Table 5.9 Sample comments from thematic analysis.................................................................. 168 Table 6.1 TKA panelists‟ demographics .................................................................................... 189 Table 6.2 TKA panelists‟ response rates by rounds ................................................................... 190 Table 6.3 Level of agreement by Delphi round .......................................................................... 192 Table 6.4 Influence of personal and external factors by section ................................................. 196 Table 6.5 Appropriate and important post-acute rehabilitation interventions after TKA........... 201 Table 6.6 Important outcomes to routinely assess and/or monitor after TKA ............................ 204 Table 6.7 Measures and tools considered feasible and important for routine clinical outcome evaluation and/or monitoring after primary TKA ........................................................ 206 Table 6.8 Sample comments from thematic analysis.................................................................. 209   x List of figures Figure 1.1 Guideline development cycle ........................................................................................ 9 Figure 2.1 Data analysis flow chart .............................................................................................. 25 Figure 3.1 PRISMA flow chart for THA systematic review ........................................................ 64 Figure 3.2 Forest plot for WOMAC pain comparing mean post-test values ................................ 74 Figure 3.3 Forest plot for WOMAC pain comparing mean change scores .................................. 74 Figure 3.4 Forest plot for Harris Hip Score comparing mean post-test values ............................. 76 Figure 3.5 Forest plot for Harris Hip Score comparing mean change scores ............................... 76 Figure 3.6 Forest plot for walking speed comparing mean post-test values ................................. 77 Figure 3.7 Forest plot for walking speed comparing mean change scores ................................... 77 Figure 4.1 PRISMA flow chart for TKA review ........................................................................ 105 Figure 4.2 Forest plot for WOMAC pain comparing mean 8-month follow-up values ............. 116 Figure 4.3 Forest plot for WOMAC pain comparing mean 8-month change scores .................. 116 Figure 4.4 Forest plot for WOMAC function comparing mean post-test values ........................ 118 Figure 4.5 Forest plot for WOMAC function comparing mean post-test change scores ........... 118 Figure 4.6 Forest plot for SF-36 PCS comparing mean 2-month follow-up values ................... 122 Figure 4.7 Forest plot for SF-36 PCS comparing mean 2-month follow-up change scores ....... 123 Figure 7.1 Phases of clinical guideline development for THA and TKA rehabilitation ............ 223 Figure 7.2 The Revised Ottawa Model of Research Use ............................................................ 237   xi List of abbreviations AAOS  American Academy of Orthopaedic Surgeons  AHP  Allied health professional  APTA  American Physical Therapy Association  CCT  Clinical controlled trial  CI  Confidence interval  CJRR  Canadian Joint Replacement Registry  COA  Canadian Orthopaedic Association  CPA  Canadian Physiotherapy Association  CPG  Clinical practice guideline  CRA  Canadian Rheumatology Association  DVT  Deep vein thrombosis  FP  Family practitioner  HRQoL Health-related quality of life  LOS  Length of stay  NIH  National Institutes of Health  NMES  Neuromuscular electrical stimulation  OA  Osteoarthritis  OT  Occupational therapy or occupational therapist  PAC  Post-acute care  PRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analyses  xii PT  Physiotherapist or physical therapist  RA  Rheumatoid arthritis  ROM  Range of motion  RCT  Randomized controlled trial  SNF  Skilled nursing facility  TENS  Transcutaneous electrical nerve stimulation  THA  Total hip arthroplasty  THR  Total hip replacement  TJA  Total joint arthroplasty  TJR  Total joint replacement  TKA  Total knee arthroplasty  TKR  Total knee replacement  WMD  Weighted mean difference  WOMAC Western Ontario and McMaster Universities Osteoarthritis Index xiii Acknowledgements I would like to thank my thesis supervisor Catherine Backman, PhD, OT for allowing me to work independently yet with just the right amount of guidance, questioning and expert editing to keep me on track and enable me to complete a project that has been very important to me for more than a decade. I so appreciate my committee members Donna MacIntyre, PhD, PT and Matt Liang, MD, MPH, for both supporting and challenging me throughout the process by bringing their different perspectives and vast experience into our discussions. I appreciate the help of the following individuals for their assistance in the conceptualization, planning and conducting of various aspects of the focus groups in chapter two: Carol Travis, MSc, RD; Catherine Morley, PhD, RD; Deborah Kennedy, MSc, PT; Susan Carr, Dip PT; and Susan Robarts, BSc, BHSc PT, MSc. In chapters three and four, I am grateful to Michael Schulzer, MD, PhD for his statistical advice, my co-authors, and Lara Maxwell and Elizabeth Ghogomu of the Cochrane Musculoskeletal Review Group for the support and guidance in designing and conducting the systematic reviews. For chapters five and six, I thank Rebecca Tunnacliffe, Executive Director, Physiotherapy Association of BC and Pam Montie, Chair, Consumer Advisory Board, Arthritis Research Centre of Canada for their enthusiastic support of the Delphi study. Several key people contributed to the success of the Delphi study: Asuko Brittain, BSc(PT), BSc(OT), co-author; Jonathan Berkowitz, PhD, statistical consultant; and Eric Sayre, PhD, Arthritis Research Centre of Canada, on-line survey system consultant. I was privileged to have the support of two research assistants over the course of my PhD: Osita Hibbert, MPT, MSc. and Michelle (Raglin) Block, Dip.RA. Michelle your dedication, contributions to every stage of this project and friendship have meant a great deal to me and I can‟t thank you enough. xiv Members of the North American Joint Replacement Rehabilitation Guidelines working group have provided encouragement and guidance through all phases of this thesis and helped me gain a better understanding of the Canadian and US healthcare systems. Members are listed in Appendix A. This project would not have been successful without the valuable contributions and efforts of all of the consumers and health professionals that shared their time, their experiences and expertise through participating in focus groups and serving as Delphi panelists. Many colleagues and patients also provided valuable input to the pilot testing phases and thoughtful feedback on drafts of this thesis and its manuscripts. I thank my fellow graduate students in Rehabilitation Sciences, University of BC for their support and kindness over the past several years and for sharing their strategies for completing a doctoral thesis. I also want to acknowledge the guidance and camaraderie provided by fellow trainees and researchers in the Arthritis Research Centre of Canada, in particular Linda del-Fabro and Allen Lehman, and trainees and staff of the Vancouver Coastal Research Institute and Centre for Hip Health and Mobility. Finally, I am truly grateful for the support and encouragement of my clinical colleagues, supervisors and managers in the Mary Pack Arthritis Program, Vancouver Coastal Health - many of whom took on extra tasks and responsibilities while I took time off from my position to pursue my doctoral studies. I thank Catherine McAuley and Bev Hills for all of their support and patience, and Susan Carr and Asuko Brittain in particular, for their contributions to various stages of the project. My doctoral training and research was made possible through the generous financial support from a number of organizations: UBC Rehabilitation Sciences Alumni Jane Hudson Scholarship; xv Canadian Institutes of Health Research Graduate Scholarship Master‟s Award and Strategic Training Fellowship in Quality of Life Research; The John Insall Foundation for Orthopaedics; UBC Cordula and Gunter Paetzold Graduate Fellowship; UBC Faculty of Medicine‟s Roman M. Babicki Fellowship in Medical Research (Rheumatology); and the BC Medical Services Foundation. xvi Dedication This thesis is dedicated to my family. To my husband Dale who endured the long hours, a messy dining room table and sandwiches for dinner over the past 5 years. I look forward to returning to a sense of normalcy and putting my family first again. To my children Mattias and Delaney for putting up with mom always doing “homework” and for being my cheerleaders. You both helped to keep me grounded and remind me of what really matters. To my parents, Ina and Weldon Wallace, who have always believed in me and supported my efforts along the way. Mom, I am so grateful for your help in the days leading up to completing this thesis. To my mother and father-in-law, Greta and Ole Westby, who stepped in to support our family in so many ways over the past 5 years. I so appreciate every one of you and thank you for your love and support through this journey.  xvii Co-authorship statement Sections of this thesis have been submitted or will be submitted as multi-authored papers in refereed journals. Details of co-authors‟ contributions are provided. Chapter 2: Westby MD, Backman CL. Patient and health professional views on rehabilitation practices and outcomes following total hip and knee arthroplasty for osteoarthritis: A focus group study. BMC Health Services Research. Accepted for publication February 2, 2010. I conceptualized and designed the study, conducted a majority of the focus groups, performed the thematic analysis and prepared and revised the manuscript. Dr. Backman contributed to the conceptual development and design of the study, provided guidance on and helped perform the data analysis, and assisted with preparing the manuscript draft and revision. Chapter 3: Westby MD, Carr S, Kennedy D, Brander V, Bell M, Doyle-Waters M, Backman C. Post-acute physiotherapy following primary total hip arthroplasty for osteoarthritis: A Cochrane systematic review. A version of this chapter will be submitted to The Cochrane Library of Systematic Reviews. I designed the study with input from all co-authors. M. Doyle-Waters performed the electronic search. S. Carr, D. Kennedy, V. Brander and I abstracted the data. I conducted the data analysis and prepared the manuscript. All co-authors contributed to the interpretation of results and provided feedback on drafts of the manuscript. Chapter 4: Westby MD, Kennedy D, Jones D, Jones A, Doyle-Waters M, Backman C. Post- acute physiotherapy following primary total knee arthroplasty for osteoarthritis: A Cochrane systematic review. xviii A version of this chapter will be submitted to The Cochrane Library of Systematic Reviews. I designed the study with input from all co-authors. M. Doyle-Waters performed the electronic search. D. Kennedy, D. Jones, A. Jones and I abstracted the data. I conducted the data analysis and prepared the manuscript. All co-authors contributed to the interpretation of results and provided feedback on drafts of the manuscript. Chapter 5: Westby MD, Brittain A, Liang M, Raglin Block M, Backman CL. Best practices for post-acute rehabilitation following primary total hip arthroplasty for osteoarthritis: A Delphi study. I designed this study with input from A. Brittain, C. Backman and M. Liang. The on-line version of the survey was created and administered by M. Raglin Block. A. Brittain and I performed the data analysis and all authors contributed to the interpretation of data and preparation of the manuscript. Chapter 6: Westby MD, Brittain A, Liang M, Raglin Block M, Backman CL. Best practices for post-acute rehabilitation following primary total knee arthroplasty for osteoarthritis: A Delphi study. I designed this study with input from A. Brittain, C. Backman and M. Liang. The on-line version of the survey was created and administered by M. Raglin Block. M. Westby and A. Brittain performed the data analysis and all authors contributed to the interpretation of data and preparation of the manuscript. 1 Chapter 1: Introduction and overview of rehabilitation after total hip and knee arthroplasty 1.1 Background Total hip (THA) and total knee (TKA) arthroplasty surgeries are elective surgical procedures recommended for individuals with advanced end-stage osteoarthritis (OA) that is not responsive to conservative therapies such as analgesics, anti-inflammatory agents, physiotherapy, activity modification, weight loss and use of walking aids (1). Approximately 62,000 1  primary THA and TKA surgeries were performed in Canada in 2006-07 (2) and more than ten times that in the United States (US) (3, 4). The volume of THA and TKA procedures in Canada grew by 60% and 141% respectively in the preceding ten-year period (2). The largest ten-year growth occurred in the 45-54 years and 85+ years age groups, although the majority of these procedures continue to be performed on patients aged 65-74 years (2, 4). This rapid rise in total joint arthroplasty (TJA) surgeries is projected to continue in an almost exponential fashion with the aging population and the increased prevalence of OA (5-8). Other factors thought to be contributing to the rise include the increasing prevalence of obesity, a leading risk factor in the development of knee OA, public demand and higher public expectations for enhanced quality of life, and improved surgical and anaesthetic techniques making these surgeries appropriate and safer for both older and younger patients (9-11). Data from the US National Hospital Discharge Survey (1996-1999) suggested that THA and TKA combined will equate to more than 748,000 procedures by the year 2030 (12), a figure that has already been surpassed (3, 4). More recent projections forecast that the  1  For the 2006-07 fiscal year, the CJRR annual report does not include data from Quebec and therefore under represents the total number of THA/TKA procedures performed in Canada that year. 2 demand for primary THA procedures in the US will grow to 572,000 and TKA to 3.48 million by 2030 (13). Osteoarthritis is the primary reason for 81% (2) to 94% (14) of all THA and TKA procedures. The Canadian Joint Replacement Registry (CJRR) reports that both procedures were more common in women in 2006-07 with age-standardized THA rates in women and men of 86 and 76 per 100,000 and TKA rates of 148 and 110 per 100,000 (2). The number of overweight and obese individuals undergoing THA and TKA is on the rise with 73% of THA and 87% of TKA patients being overweight or obese (2). 1.2 Relevance Joint replacement surgery is a cost-effective treatment for advanced OA of the hip and knee joints and decreases pain, increases mobility and function, and improves health-related quality of life (HRQoL) (7, 15-19). High rates of patient satisfaction are also reported (20). However, due to the rapid growth in primary and revision procedures, TJA surgeries place a significant burden on healthcare budgets (8, 9). The American Academy of Orthopaedic Surgeons (AAOS) using 2002 data reported total hospitalization costs for primary THA and TKA procedures were $5.91 and $11.38 billion US dollars respectively (3). Just three years later, the estimated in-hospital costs had risen to more than $9.2 and $17.7 billion US dollars for THA and TKA (3). An earlier cost comparative study based on 1997 - 2001 data from Canadian and US teaching hospitals revealed that in-hospital costs in US dollars per THA procedure were $6,766  $119 and $13,339  $131 at Canadian and US hospitals respectively (21). In 2005, mean hospitalization charges per THA and TKA procedure in the US had grown to $39,000 and $36,000 respectively (3). Exact figures are not available in Canada; however recent data suggests costs associated with THA and TKA surgery and the period up to six months post-op average more than $14,700 per 3 patient leading to an informal estimate of current Canadian expenditures in excess of $911 million dollars annually (2, 22). Approximately one-third of total costs of THA procedures in Canada are related to the post- operative rehabilitation phase (21). Based on 2003 data, direct and indirect costs associated with primary THA and TKA rehabilitation were estimated to be $3.4 billion annually in the US (23). Significant out-of-pocket costs have also been reported by patients during the first year following surgery (24). In the past decade, length of hospitalization following THA and TKA has decreased by 38% and 50% respectively in Canada with 2006-07 data showing the average acute hospital length of stay (LOS) to be seven days for THA procedures and six days for TKA procedures including revisions (2). International data confirm a shorter LOS when only primary procedures are considered (3, 9). Earlier discharge to home or community services decreases the time available for physical recuperation, acute rehabilitation, patient and family education and counseling, and discharge planning. A reduced LOS places additional burden and responsibility on the patients, their family and their post-acute health care providers to monitor for and address post-operative complications such as inadequate pain control, infection and deep vein thromboses. Further, the shortened acute care stay puts more emphasis on preadmission education and the role and timing of post-discharge physiotherapy and other rehabilitation interventions (11). In an editorial examining LOS after THA and TKA, Johanson comments that "it remains unclear just how far this process [of reducing LOS] can be taken without either compromising quality of care or simply shifting costs to a less regulated outpatient environment" (25)(p.1). 4 1.3 Rehabilitation practices The majority of patients undergoing TJA receive some form of post-operative rehabilitation to facilitate functional recovery and optimal surgical outcomes; however, there are no North American estimates of the actual proportion. Rehabilitation is defined as “a process aimed at enabling people with disabilities to reach or maintain their optimal physical, sensory, intellectual, psychological and social functional levels” (26). Rehabilitation after TJA is available through inpatient, outpatient or home care services. Inpatient rehabilitation settings include inpatient rehabilitation facilities (IRFs) and transitional care or skilled nursing facilities (SNFs). Discharge data for Ontario hospitals for the one-year period 2001/02 revealed that 44% and 42% of patients undergoing THA and TKA respectively received post-acute rehabilitation from an inpatient rehabilitation hospital (27). Since then, efforts have been made to reduce the number of patients discharged to inpatient rehabilitation; however, there remain marked differences amongst provinces (28, 29). Following an inpatient rehabilitation stay, a majority of patients go on to receive further therapy in outpatient settings (30). Variation in discharge destinations and availability of inpatient rehabilitation is reported worldwide (31-35). Clinical and non-clinical factors influence discharge destinations and setting for post-acute rehabilitation (36) including older age, length of acute care stay, co-morbidity (27, 34), functional independence, cognitive function, marital status (34), race and ethnicity (36), obesity (37), living alone (37), patient preference, gender and knowledge of TJA care (38). Worldwide, there is tremendous variation in the type, timing, frequency and duration of rehabilitation treatments and outcome assessment (31, 33, 34, 39). Therapeutic exercises and rehabilitation protocols are largely based on clinical experience and preferences (40-42), local customs and facility-based protocols (43-45), available resources, the acute care phase of recovery (46, 47) and outdated approaches (48). Essentially there is no guidance or evidence- 5 based practice guidelines to inform clinical decision-making and identify best practices for post- acute rehabilitation after THA and TKA surgery. Finally „best practice‟ has been confused by the pressure to reduce LOS from both acute and post-acute care, without adequate evidence regarding the downstream effect of these decisions. There has been no prospective, systematic follow up regarding patients‟ long-term functioning and surgical outcomes following different rehabilitation practices. Rehabilitation interventions may enhance surgical outcomes; however, their exact contribution to long-term outcomes such as functional mobility, participation and HRQoL is not clear. The National Institutes of Health (NIH) conferences on TJA concluded that “the use of rehabilitation services is perhaps the most understudied aspect of the peri-operative management of TKA patients” (49) (p.6) and also recommended that: “The contribution of pre-hospital, in-hospital and post-hospital education and rehabilitation programs to the eventual outcome of the surgical procedure deserves an organized, in-depth study to determine optimum regimen, duration of treatment, and expected outcomes” (50) (p.8). 1.4 Functional impairments, activity limitations and participation restrictions following THA and TKA Despite the fact that outcomes of TJA are amongst the few surgical procedures that are both cost-effective and improve quality of life, persistent pain, prolonged physical impairments, gait abnormalities, activity limitations and participation restrictions are evident two or more years following THA and TKA in some individuals even though some form of postoperative rehabilitation was undertaken (15, 51-57). It has been suggested that current exercise programs performed during the early phase of (55) rehabilitation are insufficient to restore muscle strength, balance, proper gait pattern and more complex functional activities (51, 54). As well, reduced physical capacity coupled 6 with further reduction of physiological reserve capacity seen with normal aging may lead to declining independence in daily living for older adults (55). Multiple studies have demonstrated that weakness in the lower extremities is a major risk factor for falls in the geriatric age group (50) and that falls remain an issue even after joint replacement surgery (58). More than a decade has passed since the original NIH conference on THA and few studies have been published addressing these concerns. There continues to be a notable lack of consensus regarding which rehabilitative peri-operative practices should be used and few well-designed studies testing the efficacy and effectiveness of such practices (49). 1.5 Canadian versus American health care systems Our work is intended for North America but it is important to acknowledge the differences between the American and Canadian health care systems. The Canadian health care system is characterized by universal access and government funded health care for physician and hospital- based services, few for-profit providers, and lower national health care expenditures than in the US (59), with its varied access to public and private providers depending on one‟s insurance. In the first half of 2008 in the US, 42.8 million or 14.3 % of individuals of all ages had no health insurance (60). Thus, a large number of patients may have limited access to surgical and rehabilitation services (61). In 1965, the US Government established the Medicare program, a health insurance program for people age 65 or older and some disabled individuals under age 65 (www.Medicare.gov). That same year, Medicaid was started to provide medical benefits to eligible groups of low-income people, some who may have no medical insurance or inadequate medical insurance. Based on 2001-2003 data, approximately 80% of primary THA and TKA procedures on adults aged 65 years and older in the US are covered by Medicare and Medicaid programs (62). 7 To address escalating costs incurred by Medicare and Medicaid, a prospective payment system (PPS) was introduced in 1983 for acute hospital care and later for post-acute care (PAC) with SNFs in 1998, home health services in 2000, and inpatient rehabilitation in 2002 (36, 63). The PPS established new rules governing access to inpatient rehabilitation (i.e. Medicare 75% Rule) and resulted in fewer patients with a primary THA or TKA for OA being eligible for this form of rehabilitation (36) yet increased overall spending for PAC (11, 63). Under the PPS, Medicare caps payments to home health providers based on a patient‟s classification in one of 80 home health resource groups and a 60-day episode of care. Rehabilitation professionals providing post- arthroplasty care have had to adapt to these reimbursement changes and find ways to optimize the allotted 10-visit therapy sessions (64). In Canada, roles and responsibilities for health care are shared between the federal and provincial-territorial governments according to the Canada Health Act (CHA) (www.hc- sc.gc.ca). Using funds transferred from the Federal Government, provincial and territorial governments are responsible for the management, organization and delivery of health services for their residents. Addressing wait times for elective TJA surgery has been a national and provincial priority for most of the past decade (65-67). However, national median/mean wait times of 127/182 and 169/237 days for THA and TKA surgery respectively are still reported (2). These are far longer than the three to four weeks waiting times reported in the US (68). In turn, these differences in funding schemes and access to surgical and rehabilitation care influence health outcomes (59). These fundamental differences and factors need to be integrated into practice guidelines that will be relevant to both countries. 1.6 Clinical practice guidelines Clinical practice guidelines (CPGs) are “systematically developed statements about specific clinical problems to assist practitioners and patients in making decisions about appropriate health 8 care” (69)(screen 3). Guidelines are “generated from a systematic, rigorous and explicit methodology, including a review of all available scientific evidence” (69)(screen 3). Clinical practice guidelines aim to close the gap between evidence and practice, address unexplained variations in clinical practice and improve the quality, efficiency and effectiveness of health care by using the best available scientific evidence and expert opinion to make clinical recommendations (70, 71). Criteria for selecting a clinical topic or question for guideline development include prevalence of the clinical condition, variations in current health practices, cost of current practice, availability of high quality evidence to support practice, potential to change health outcomes and costs, and feasibility (70), all of which are met by our topic of THA and TKA rehabilitation. 1.7 Guideline development framework A number of theoretical models and frameworks on the development, dissemination and implementation of guidelines were examined for their applicability to this project (72-75). Davis and colleagues identified a four-stage approach in the development of clinical practice guidelines (74): 1. Select clinical problem 2. Synthesize data 3. Develop guidelines 4. Disseminate, implement and evaluate guidelines. 9 Figure 1.1 Guideline development cycle  Identify current practice Consensus process to supplement evidence   Our multidisciplinary advisory group (Appendix A) felt at least two additional steps were important for developing guidelines for THA and TKA rehabilitation: 1) identify current rehabilitation practices, issues and post-operative outcomes from various stakeholder perspectives; and 2) conduct a formal consensus process to complement the scientific data in areas where the evidence is weak, inconsistent or nonexistent (Figure 1.1). 1.8 Other guidelines initiatives The French Society of Physical and Rehabilitation Medicine (SOFMER) attempted to develop evidence-based guidelines for ambulatory physiotherapy following THA (76) and TKA (77). Following a systematic review of the French and English literature, however, they restricted their work to orthopaedic surgeon, physiatrist and rheumatologist expert opinion, with little patient or allied health professional input and used a survey of current practice patterns rather than a formal consensus process, to establish best practice and supplement those areas where there was little or no published evidence (78). Stakeholder involvement is recognized as being extremely important to the guideline process and ensures that differing perspectives, expectations and practice variation are addressed and that complete data are available (71). 10 In 2005, two separate reviews were conducted to inform rehabilitation practice and education in Ontario. The review by the Medical Advisory Secretariat of the Ministry of Health and Long- Term Care examined the effects of pre- and post-operative physiotherapy on functional outcomes after THA and TKA. While practice guidelines were not their objective, their final recommendations were designed to inform THA and TKA referral, discharge and rehabilitation practice patterns in order to increase efficiencies in the delivery of rehabilitation services in that province (79). Several focus groups with patients and health professionals followed by an English-only literature review informed the Greater Toronto Area Rehabilitation Network's development of an online discussion forum to promote collaboration and education among patients and health professionals throughout the continuum of TJA care (80). Based on this review spanning pre- operative, acute and post-operative care, rehabilitation was found to be an effective component in TJA management at various stages of recovery. However, best practice recommendations were not clearly differentiated for hip versus knee replacement. Soever and MacKay concluded that comparison of outpatient care to other rehabilitation approaches warranted further study to determine the optimal processes of care following TJA (80). 1.9 Purpose The overall aim of this thesis was to complete the first step in the development of multidisciplinary clinical practice guidelines for the post-acute rehabilitation of patients undergoing primary THA and TKA for OA. The post-acute phase extends from immediately following discharge from the acute care setting and up to 12 months post-surgery (Figure 1.2). 11 Figure 1.2 TJA Continuum of care  1.10 Overview of thesis chapters This thesis consists of five manuscript chapters, placed between this introductory chapter and a concluding chapter. Together, they address the initial three phases of clinical guideline development. The primary aim of each manuscript chapter is briefly described below: Chapter 2: Using focus group methodology key themes were identified related to current rehabilitation practices and outcome assessment following THA and TKA surgery across different stakeholder groups in Canada and the United States. Chapters 3 and 4: The evidence for post-acute physiotherapy on pain, function and HRQoL following primary THA (Chapter 3) and TKA (Chapter 4) for osteoarthritis was examined following Cochrane systematic review methods. 12 Chapters 5 and 6: Delphi surveys with separate expert panels were conducted to develop best practice recommendations for post-acute rehabilitation after primary THA (Chapter 5) and TKA (Chapter 6). While the long term goal is comprehensive rehabilitation guidelines, the thesis is limited to sysnthesizing the physical aspects of rehabilitation. However, recognizing the multidisciplinary nature of rehabilitation, I used broad inclusion criteria and incorporated other professions in the focus groups and Delphi processes. 13 1.11 References 1. Katz JN. Total joint replacement in osteoarthritis. Best Pract Res Clin Rheumatol 2006;20(1):145-53. 2. Canadian Institute for Health Information (CIHI). Hip and total knee replacements in Canada - Canadian Joint Replacement Registry (CJRR). 2008-2009 Annual Report. [cited 2009 Sep 17]. Available from: URL: www.cihi.ca/cjrr. 3. American Academy of Orthopedic Surgeons. Facts on hip replacements and facts on knee replacements. [cited 2009 Dec 9]. Available from: URL: http://www.aaos.org/research/stats/patientstats.asp. 4. DeFrances CJ, Lucas CA, Verita CB, Golosinskiy A. 2006 National Hospital Discharge Survey. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Center for Health Statistics; 2008 Jul [cited 2009 Nov 21]. Report No.: 5. Available from URL: http://www.cdc.gov/nchs/data/nhsr/nhsr005.pdf. 5. Birrel F, Johnell O, Silman A. Projecting the need for hip replacement over the next three decades: Influence of changing demography and threshold for surgery. Ann Rheum Dis 1999;58(9):569-72. 6. Badley EM, Crotty M. An international comparison of the estimated effect of the aging of the population on the major cause of disablement in musculoskeletal disorders. J Rheumatol 1995;22(10):1934-40. 7. Jones CA, Beaupre LA, Johnston DWC, Suarez-Almazor ME. Total joint arthroplasties: Current concepts of patient outcomes after surgery. Rheum Dis Clin North Am 2007;33(1):71- 86. 8. Kim S. Changes in surgical loads and economic burden of hip and knee replacements in the US: 1997 - 2004. Arthritis Rheum 2008;59(4):481-8. 9. March LM, Bagga H. Epidemiology of osteoarthritis in Australia. Med J Aust 2004;180(5 Suppl):S6-S10. 10. Karrholm J, Garellick G, Rogmark C, Herberts P. Swedish hip arthroplasty register Annual Report 2007. [cited 2009 May 25]. Available from: URL: http://www.jru.orthop.gu.se. 11. Tian W, DeJong G, Brown M, Hsieh CH, Zamfirov ZP, Horn SD. Looking upstream: factors shaping the demand for postacute joint replacement rehabilitation. Arch Phys Med Rehabil 2009;90(8):1260-8. 12. Frankowski JJ, Watkins-Castillo S. Primary total knee and hip arthroplasty projections for the U.S. population to the year 2030. 2002 Aug [cited 2005 Dec 6]:1-8. Available from: URL: https://www.aaos.org/wordhtml/research/stats/TJR_projections.pdf. 14 13. Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 2007;89(4):780- 5. 14. New Zealand Orthopaedic Association. The New Zealand Joint Registry, Nine year report. [cited 2009 Jan 15]. Available from: URL: http://www.cdhb.govt.nz/NJR/reports/A2D65CA3.pdf. 15. Brander VA, Stulberg SD. Rehabilitation after hip- and knee-joint replacement: An experience- and evidence-based approach to care. Am J Phys Med Rehabil 2006;85(Suppl):S98- S118. 16. Chang RW, Pellisier JM, Hazen GB. A cost-effectiveness analysis of total hip arthroplasty for osteoarthritis of the hip. J Am Med Assoc 1996;275(11):858-65. 17. Ethgen O, Bruyere O, Richy F, Dardennes C, Reginster JY. Health-related quality of life in total hip and total knee arthroplasty. A qualitative and systematic review of the literature. J Bone Joint Surg Am 2004;86-A(5):963-74. 18. George LK, Ruiz D, Jr., Sloan FA. The effects of total hip arthroplasty on physical functioning in the older population. J Am Geriatr Soc 2008;56(6):1057-62. 19. Salmon P, Hall GM, Peerbhoy D, Shenkin A, Parker C. Recovery from hip and knee arthroplasty: Patients' perspective on pain, function, quality of life, and well-being up to 6 months postoperatively. Arch Phys Med Rehabil 2001;82(3):360-6. 20. Ozawa T, Shimizu K. Evaluation of patients' satisfaction after total hip arthroplasty (THA). J Phys Ther Sci 2007;19(2):139-43. 21. Antoniou J, Martineau P, Filion K, Haider S, Zukor D, Huk OL, et al. In-hospital cost of total hip arthroplasty in Canada and the United States. J Bone Joint Surg Am 2004;86-A(11):2435-9. 22. Hawker GA, Badley EM, Croxford R, Coyte PC, Glazier RH, Guan J, et al. A population- based nested case-control study of the costs of hip and knee replacement surgery. Med Care 2009;47(97):732-41. 23. Lavernia CJ, D‟Apuzzo MR, Hernandez VH, Lee DJ, Rossi MD. Postdischarge costs in arthroplasty surgery. J Arthroplasty 2006;21(6 Suppl 2):144-50. 24. March LM, Cross M, Tribe K, Lapsley H, Courtenay B, Brooks P. Cost of joint replacement surgery for osteoarthritis: The patients' perspective. J Rheumatol 2002;29(5):1006-14. 25. Johanson NA. The hospital course in total hip and knee arthroplasty: Which length of stay is right? J Rheumatol 1997;24(1):1-2. 26. World Health Organization. Health topics - Rehabilitation. [cited 2008 Dec 6]. Available from: URL: http://www.who.int/topics/rehabilitation/en/. 15 27. Jaglal SB, MacKay C, Corrigan L. Rehabilitation for total joint replacement. In: Badley E, Glazier R, editors. Arthritis and related conditions in Ontario. ICES research atlas. 2nd ed. Institute for Clinical Evaluative Sciences; 2004. p. 133 - 146. [cited 2008 Oct 1]. Available from: URL: http://www.ices.on.ca/file/ACF13B.pdf. 28. Nigam M. 2006_13 UBC_CSI completed cases by period additional cases 1. Centre for Surgical Innovation, Vancouver Coastal Health Authority; 2007 Apr. 29. Alberta Bone and Joint Institute. Alberta hip and knee replacement project interim report, December 19, 2005. [cited 2009 May 6]. Available from: URL: http://www.albertaboneandjoint.com. 30. DeJong G, Tian W, Smout RJ, Horn SD, Putman K, Smith P, et al. Use of rehabilitation and other health care services by patients with joint replacement after discharge from skilled nursing and inpatient rehabilitation facilities. Arch Phys Med Rehabil 2009;90(8):1297-305. 31. Lingard EA, Berven S, Katz JN, Kinemax Outcomes Group. Management and care of patients undergoing total knee arthroplasty: Variations across different health care settings. Arthritis Care Res 2000;13(3):129-36. 32. Mauer KA, Abrahams EB, Arslanian C, Schoenly L, Taggart HM. National practice patterns for the care of the patient with total joint replacement. Orthop Nurs 2002;21(3):37-47. 33. Müller E, Blatt K, Gülich M, Jäckel WH. Treatment schemes and guidelines for rehabilitation after total hip replacement (THR): Results of an international survey and a multicenter study in Germany. Int J Rehabil Res 2007;30(1):86-7. 34. Roos EM. Effectiveness and practice variation of rehabilitation after joint replacement. Curr Opin Rheumatol 2003;15(2):160-2. 35. Viliani T, Huber U, Pasquetti P, Poli P, Marcucci M, Popolizio A. Rehabilitation after primary total hip replacement. Comparison between Italian and international protocols. Eura Medicophys 2004;40(2):67-74. 36. Buntin MB. Access to postacute rehabilitation. Arch Phys Med Rehabil 2007;88(11):1488- 93. 37. de Pablo P, Losina E, Phillips CB, Fossel AH, Mahomed N, Lingard EA, et al. Determinants of discharge destination following elective total hip replacement. Arthritis Rheum 2004;51(6):1009-17. 38. Mahomed NN, Koo Seen Lin MJ, Levesque J, Lan S, Bogoch ER. Determinants and outcomes of inpatient versus home based rehabilitation following elective hip and knee replacement. J Rheumatol 2000;27(7):1753-8. 39. Oatis CA, Brennan JL, Kushman S, Werner A, Franklin PD. Describing "usual" physical therapy practice following total knee replacement in Montgomery County, Pennsylvania. Arthritis Rheum 2009;60(10 Suppl):S425. 16 40. Ganz SB, Viellion G. Pre and post surgical management of the hip and knee. In: Wegener ST, Belza BL, Gall EP, editors. Clinical Care in the Rheumatic Diseases. Atlanta: American College of Rheumatology; 1996. p. 103-6. 41. Kampa K. Orthopedic assessment and treatment of the geriatric patient. Lewis CB and Knortz KA, editors. St. Louis (MO): CV Mosby Co.; 1993. 42. Naylor J, Harmer A, Fransen M, Crosbie J, Innes L. Status of physiotherapy rehabilitation after total knee replacement in Australia. Physiother Res Int 2006;11(1):35-47. 43. Cahill JB, Koxman LM. Total knee arthroplasty. In: Cioppa-Mosca J, Cahill JB, Cavanaugh JT, Corradi-Scalise D, Rudnick H, Wolff AL, editors. Postsurgical Rehabilitation Guidelines for the Orthopedic Clinician. St. Louis: Hospital for Special Surgery, Mosby Elsevier; 2006. p. 17- 28. 44. Kennedy DM, Walsh M, Woodhouse LJ. Development of a functional independence training rehabilitation program for patients following total hip or knee arthroplasty. Physiother Can 1999;51(4):292-9. 45. Young Tucker C, Diamond A. Total hip arthroplasty. In: Cioppa-Mosca J, Cahill JB, Cavanaugh JT, Corradi-Scalise D, Rudnick H, Wolff AL, editors. Postsurgical Rehabilitation Guidelines for the Orthopedic Clinician. St. Louis: Hospital for Special Surgery, Mosby Elsevier; 2006. p. 3-16. 46. Bukowski E. Practice guideline: Acute care management following total hip arthroplasty (postoperative days 1-4). Orthop Phys Ther Pract 2005;17(3):10-4. 47. Enloe LJ, Shields RK, Smith K, Leo K, Miller B. Total hip and knee replacement treatment programs: A report using consensus. J Orthop Sports Phys Ther 1996;23(1):3-11. 48. Mont MA, Seyler TM. Advanced techniques for rehabilitation after total hip and knee arthroplasty. Clin Orthop Relat Res 2009 Jun [cited 2009 Aug 13];467(6). Available from: URL: http://www.springerlink.com/content/5j7t7h3510r203t5/. 49. National Institutes of Health. NIH consensus development conference on total knee replacement. NIH Consens State Sci Statements 2003 Dec 8-10 [cited 2004 Mar 17];20(1):1-34. Available from: URL: http://consensus.hin.gov/2003/2003TotalKneeReplacement117PDF.pdf. 50. National Institutes of Health. Total hip replacement. NIH Consens State Sci Statement 1994 Sep 12-14 [cited 2004 Apr 20];12(5):1-31. Available from: URL: http://consensus.nih.gov/1994/1994HipReplacement098html.htm. 51.  Meier W, Mizner RL, Marcus RL, Dibble LE, Peters C, Lastavo PC. Total knee arthroplasty: Muscle impairments, functional limitations, and recommended rehabilitation approaches. J Orthop Sports Phys Ther 2008;38(5):246-56. 52. Shih CH, Du YK, Lin YH, Wu CC. Muscular recovery around the hip joint after total hip arthroplasty. Clin Orthop Relat Res 1994;302:115-20. 17 53. Rossi MD, Hasson S. Lower-limb force production in individuals after unilateral total knee arthroplasty. Arch Phys Med Rehabil 2004;85(8):1279-84. 54. Trudelle-Jackson E, Emerson R, Smith S. Outcomes of total hip arthroplasty: A study of patients one year postsurgery. J Orthop Sports Phys Ther 2002;32(6):260-7. 55. Walsh M, Woodhouse LJ, Thomas SG, Finch E. Physical impairments and functional limitations: A comparison of individuals 1 year after total knee arthroplasty with control subjects. Phys Ther 1998;78(3):248-58. 56. McClelland JA, Webster KE, Feller JA. Gait analysis of patients following total knee replacement: A systematic review. Knee 2007;14(4):253-63. 57. Mizner RL, Stevens JE, Snyder-Mackler L. Voluntary activation and decreased force production of the quadriceps femoris muscle after total knee arthroplasty. Phys Ther 2003;83(4):359-65. 58. Swinkels A, Newman JH, Allain TJ. A prospective observational study of falling before and after knee replacement surgery. Age Ageing 2008;38(2):175-81. 59. Guyatt GH, Devereaux PJ, Lexchin J, Stone SB, Yalnizyan A, Himmelstein D, et al. A systematic review of studies comparing health outcomes in Canada and the United States. Open Med 2007;1(1):E27-36. 60. Martinez ME, Cohen RA. Health insurance coverage: Early release of estimates from the National Health Interview Survey, January-June 2008. [cited 2009 Dec 10]. Available from: URL: http://www.cdc.gov/nchs/nhis.htm. 61. Hinman H, Bozic K. Impact of payer type on resource utilization, outcomes and access to care in total hip arthroplasty. J Arthroplasty 2008;23(6 Suppl 1):9-14. 62. Matlock D, Earnest M, Epstein A. Utilization of elective hip and knee arthroplasty by age and payer. Clin Orthop Relat Res 2008;466(4):914-9. 63. Kaplan SJ. Growth and payment adequacy of Medicare postacute care rehabilitation. Arch Phys Med Rehabil 2007;88(11):1494-9. 64. Collins T, Herness J, Martenas J, Roberson A. Medicare prospective payment before and after implementation: A review of visits and physical performance among Medicare home health patients after total knee replacements. Home Healthc Nurse 2007;25(6):401-7. 65. Arnett G, Hadorn D, the Steering Committee of the Western Canada Waiting List Project. Developing priority criteria for hip and knee replacement surgery: Results from the Western Canada Waiting List Project. Can J Surg 2003;46(4):290-6. 66. Mascarenhas R. The Manitoba Arthroplasty Waiting List: Impact on health-related quality of life and initiatives to remedy the problem. J Eval Clin Pract 2009;15(1):208-11. 18 67. Frank C, Dick D, Smith D, Wasylak T, Gooch K, Zernicke R. The Alberta Bone and Joint Health Institute: Creating sustainable accountability through collaboration, relevant measurement and timely feedback. Healthc Pap 2006;7(1):34-9. 68. Westby MD, Backman CL. Patient and health professional views on rehabilitation practices and outcomes following total hip and knee arthroplasty: A focus group study. BMC Health Services Research. Accepted for publication 2010 Feb 2. 69. Bigos S, Bowyer O, Braen G, Brown K, Deyo R, Haldeman S, et al. Acute low back problems in adults. Rockville (MD): US Department of Health and Human Services; 1994 Dec. Report No.: Publication No. 95-0642. Clinical Practice Guideline No. 14. 70. Canadian Physiotherapy Association. Clinical practice guidelines: A discussion paper for the Canadian Physiotherapy Association. 1996 May. 71. Liang MH. The American College of Rheumatology Practice Guidelines Strategy. Arthritis Care Res 1995;8(3):132-3. 72. Grol R, Grmishaw J. From best evidence to best practice: Effective implementation of change in patients‟ care. Lancet 2003;362(9391):1225-30. 73. Moulding NT, Silagy CA, Weller DP. A framework for effective management of change in clinical practice: Dissemination and implementation of clinical practice guidelines. Qual Health Care 1999;8(3):177-83. 74. Davis DA, Taylor-Vaisey A. Translating guidelines into practice: A systematic review of theoretic concepts, practical experience and research evidence in the adoption of clinical practice guidelines. Can Med Assoc J 1997;157(4):408-16. 75. Koutsavlis AT, Paradis G, Renaud L. Disseminating practice guidelines to physicians. [cited 2004 Sep 16]. Available from: URL: http://www.inspq.qc.ca. 76. Genêt F, Gouin F, Coudeyre E, Revel M, Rannou F. The benefits of ambulatory physiotherapy after total hip replacement. Clinical practice recommendations. Ann Readapt Med Phys 2007;50(9):776-82. 77. Genêt F, Mascard E, Coudeyre E, Revel M, Rannou F. The benefits of ambulatory physiotherapy after total knee replacement. Clinical practice recommendations. Ann Readapt Med Phys 2007;50(9):793-801. 78. Rannou F, Coudeyre E, Ribinik P, Macé Y, Poiraudeau S, Revel M. Establishing recommendations for rehabilitation interventions: The SOFMER methodology. Ann Readapt Med Phys 2007;50(2):106-10. 79. Medical Advisory Secretariat, Ontario Ministry of Health and Long-Term Care. Physiotherapy rehabilitation after total knee or hip replacement. Health technology literature review. 2005 May [cited 2005 Dec 6]. Available from: URL: http://www.health.gov.on.ca/english/providers/program/mas/tech/ohtas/tech_rehabtkr_061705.ht ml. 19 80. Soever L, MacKay C. Best practices across the continuum of care for total joint replacement. Greater Toronto Area Rehabilitation Network; 2005 Jul. [cited 2005 Dec 6]. Available from: URL: http://www.gtarehabnetwork.ca/downloads/report-totaljointreplacement-2005.pdf.  20 Chapter 2: Patient and healthcare professional views on total hip and knee replacement rehabilitation and outcomes: A focus group study2 2.1 Background Total hip arthroplasty (THA) and total knee arthroplasty (TKA) surgeries are highly successful orthopaedic procedures for more than 62,000 Canadians (1) and 773,000 Americans (2) each year. The growth in number of THAs and TKAs exceeds the aging of our population due in part to both younger and older individuals electing joint replacement surgery as a feasible option for their advanced hip and knee osteoarthritis (OA) (3). Most patients receive post-operative physical therapy and/or other rehabilitative services in the hospital, as an outpatient or through home care services (4). However, the setting, timing, amount and treatment approaches differ widely (5-8). Despite the cost effectiveness of THA and TKA, in-hospital and rehabilitation costs associated with these surgeries place significant burdens on North American healthcare systems (2, 9-11). Rehabilitation interventions (e.g., physical therapy, occupational therapy, nursing care) may enhance surgical outcomes; however, their precise contribution to long-term outcomes such as physical function, mobility, participation in life roles and health-related quality of life (HRQoL) is not clear. A National Institutes of Health (NIH) conference concluded that “…rehabilitation services are perhaps the most understudied aspect of the peri-operative management of TKA patients” (12)(p.6). Disparate views on need for total joint arthroplasty (TJA) surgery, expectations and outcomes of surgery have been reported for physicians and patients (13-15), and between surgeons and other health professionals (16). Hewlett suggests that patients‟ assessments may differ from those of  2  A version of this chapter has been accepted for publication. Westby MD, Backman CL. Patient and health professional views on rehabilitation practices and outcomes following total hip and knee arthroplasty for osteoarthritis: A focus group study. BMC Health Services Research. Accepted February 2, 2010. 21 health professionals due to the influence of needs, attitudes, priorities, experiences and expectations (17). It is therefore necessary to explore patient and provider expectations to inform clinical practice guidelines. The Canadian health care system is characterized by universal access and government funded health care for physician and hospital-based services, few for-profit providers, and lower national health care expenditures than in the US (18), with its varied access to public and private providers depending on one‟s insurance. These differences in turn influence surgical wait times (1), access to and funding for rehabilitation services, and health outcomes (18); thus the need to incorporate both perspectives. 2.2 Purpose The purpose of this study was to move beyond the existing literature and explore patient and health professional experiences with current rehabilitation practices and outcomes following THA and TKA to inform the development of clinical practice guidelines applicable for North America. 2.3 Methods 2.3.1 Sampling frame We were interested in perspectives from four stakeholder groups: 1) individuals who had a primary THA or TKA for OA within the past year; 2) allied health professionals (AHP, e.g., physical therapist (PT), occupational therapist (OT), nurse, medical social worker) currently providing THA or TKA rehabilitative care, education or counselling; 3) physicians (e.g., rheumatologist, physiatrist, family practitioner) who provide THA or TKA care; or 4) orthopaedic surgeons currently performing THA or TKA. Patients were excluded if they were less than 19 years of age, could not converse in English; or had undergone THA or TKA surgery 22 for inflammatory arthritis, acute fracture/trauma or tumour. Spouses were permitted to join the patient discussion groups. 2.3.2 Recruitment We therefore used strategies to accrue a purposive sample across stakeholder group, demographics, geographic areas (e.g. rural (<10,000 inhabitants) and urban areas) and level of experience. Notices, inviting interested individuals to contact the local study coordinator, were posted in clinics, waiting rooms, seniors‟ centers and arthritis consumer groups‟ newsletters as applicable to each stakeholder group. E-mail notices were distributed using staff directories for all types of health professionals. 2.3.3 Focus groups/interviews Focus groups are particulary suited to studying diverse perspectives to gain insight into participants‟ experiences (19, 20) and were the primary means of gathering data, where possible. Focus groups encourage contributions from less verbal individuals who feel supported by other group members with shared experiences (21). However, individual interviews were conducted when a participant was unable to attend their group. Both focus groups and interviews have been used previously in studying various aspects of THA and TKA care, patient experiences and expectations (22-27), but we are not aware of studies that examine THA and TKA rehabilitation practices and outcomes from multiple stakeholders‟ perspectives. A discussion guide was developed with input from a multi-disciplinary group of clinicians experienced in THA and TKA rehabilitation and researchers experienced in focus group methodology. Open-ended questions progressed from general and uncued to more specific questions with accompanying probes (20, 28). The discussion guide was tested twice and revised to improve clarity based on health professional and patient feedback. Key questions and probes 23 (Appendix B) were rephrased for each stakeholder group to ensure relevance to participants (20).  Separate focus groups were conducted with each set of stakeholders to avoid a perceived hierarchy among mixed professional and professional-patient participants (29). A pair of moderators led each focus group using the standardized discussion guide. The four moderators were female PTs with experience in TJA rehabilitation and group process and included the lead author. Prior to the first focus group, moderators were given written and videotaped instructions on focus group methodology, moderating tips and use of the data collection forms, and each pair conducted a pilot session to gain skill and confidence in moderating sessions and trouble shoot problems related to audiotaping, timing and logistics. Focus group sessions lasted 90 minutes for health professionals and 120 minutes for patient groups (allowing for a stretch break). Individual semi-structured interviews (face-to-face or telephone) of 30-60 minutes were conducted with participants unable to participate in a focus group; they followed the discussion guide. Sessions were audiotaped and transcribed verbatim for analysis. Participants recorded thoughts on a response form prior to sharing their perspectives with other group members. Forms were collected and together with the moderators' field notes served to enrich transcripts and study rigor (30). Member checking was incorporated into focus groups and interviews by inviting participant feedback on the moderator‟s summary of the session (21). Immediately following each focus group, the moderators met to debrief, identify issues that may influence analysis and suggest possible modifications to the discussion guide (21). 2.3.4 Ethics Approval was received from the UBC Behavioral Research Ethics Board (Appendix C) and the Vancouver Coastal Health Research Institute for the primary site and as required by institutional 24 policy for each of the other sites. All participants provided informed consent prior to participation, and were offered a small token ($10 gift certificate). 2.3.5 Data analysis A thematic content analysis occurred concurrently with data collection to allow for revision of questions and development of new lines of inquiry (20, 21, 29, 31). After checking transcripts for accuracy, the two authors independently read the transcripts and performed line-by-line, open coding (29), and, following the process outlined in Figure 1, developed sub-themes for „within group analysis‟ and subsequently refined these into key themes for „across group analysis‟. Disagreements in coding and categorization were discussed and the coding framework refined as necessary using a constant comparison approach (29). Minority opinions or outliers (negative cases) were identified and discussed (30). Data collection was discontinued when it was agreed that no new ideas or issues were likely to be raised (20, 29). A decision audit trail was maintained throughout the data collection and analysis phases. Once key themes were identified, transcripts were reviewed and representative quotes selected for each theme. Portions of the coding framework and final analysis were shared with an independent, experienced qualitative researcher for peer checking (30). (See Figure 2.1) 25 Figure 2.1 Data analysis flow chart  Verbatim transcription Independent, open line- by-line coding Clustering of common codes (consensus) Formation & labeling of categories Formation of concepts/subthemes (within group analysis) Identification of major themes (across group analysis)  Audit Trail    Data Source Triangulation     Reflective Journal     Member Checking     Peer Checking STEPS PROCESS TO ENSURE STUDY RIGOUR  2.4 Results Eleven focus groups and eight semi-structured interviews were conducted in five Canadian and one US site. Participants included 32 patients and four spouses, 30 AHPs, five physicians and nine surgeons (Tables 2.1 and 2.2). Focus groups ranged in size from four to 10 participants. 26  Table 2.1 Patient participant demographics  Patients* (Type of surgery)  Age (Range, years)  Gender (♀/♂)  English as first language   Education (No. some college or higher)  Post-op stage (Range, months)  Rehab status (Completed rehab)  Work status (Retired)  Lives in urban community THA n = 13  46 - 81 7/6 13 10 1 – 11  7 8 9 TKA n = 19  46 - 78 11/8 18 15 1 – 10  9 10 14 * - patient participants only (does not include the 4 spouses)     Table 2.2 Health professional participant demographics  Professions  Age (Range, years)  Gender (♀/♂)  English as first language  TJA experience 1 (Range, years)  TJA patient volume 2 (Cases/year)  Practice setting 3   Urban- based practice   AHPs n = 30  28 – 62  26/4  25  1 – 35 <50/yr = 7  50-100/yr = 8  >100/yr = 15 Inpt acute = 4 Inpt rehab = 4 Outpatients = 15 Home care = 5 Other = 2  22   Surgeons n = 9  33 – 64  0/9  7  1 – 30 50-100/yr = 1  >100/yr = 8 Teaching hospital = 8 Regional hospital = 1   9  Physicians n = 5   41 - 60  1/4  4  6 – 35 <50/yr = 1  50-100/yr = 2  >100/yr = 2 Inpt acute = 1 Inpt rehab = 1 Private practice = 3  5   Legend: AHPs = allied health professionals; TJA = total joint arthroplasty 1 - Years of experience providing surgical, treatment or counseling services to patients with THA or TKA 2 - Number of combined THA and TKA patients treated or operated on each year 3 - Number of professionals practicing in each setting; for AHPs “Other” = recreational setting  27 2.4.1 Key themes Subthemes for each stakeholder group were compared and contrasted across the groups and six major themes emerged. Quotes are attributed to participants by noting their age, sex and group, e.g., 41, F, FP is a 41 year old, female, family practitioner. Other participants are identifed as follows: PT=physiotherapist; OT=occupational therapist; RN=nurse; SW=social worker; SURG=surgeon; PHYS=physiatrist; RHEUM=rheumatologist; THA=individual with a THA; TKA=individual with a TKA. 2.4.1.1 Theme 1: Let‟s talk A substantial amount of focus group time was spent discussing communication issues. The greatest energy and strongest group interaction occurred over the issues of inter-professional communication and collaboration across settings and throughout the continuum of care. While participants offered descriptions of both positive and negative patient-provider and inter-provider communication, most examples described how poor or lack of communication decreased efficiency, effectiveness and collaboration. “Communication amongst all the people involved is pretty much non-existent. There‟s no communication between surgeons and family doctors anymore.” [41, F, FP] “So we have this parade of people with total hips, for example, coming through as though they‟re all the same and they‟re not. And I think there‟s a real need for us all to get better information from the surgeon and I‟ve crowed about this for a long time and it hasn‟t yet happened, but I think that‟s a major weakness…. I think this lack of information leads to rote [physiotherapy] procedures that don‟t have very much thinking going on with them.” [62, M, PT] Poor communication across settings (e.g., from in-patient rehab to family practitioner or private PT) was believed to contribute to inconsistent and poorly coordinated services and negatively impact clinical outcomes and patient satisfaction. Centralized information, a communication 28 form that stays with the patient, better links between facilities and providers, and practice guidelines were suggestions shared by AHPs and physicians as ways to address this issue. „Team care‟ was another approach to enhance communication and was acknowledged as more feasible in inpatient rehabilitation settings where different healthcare providers were housed under the same roof, shared charting and participated in regular team meetings. Inadequate staffing, part time positions and staff turnover negatively impacted team dynamics and consistencies in care. A lack of a collaborative, multidisciplinary approach was felt to lead to inefficiencies, duplication of services and patient dissatisfaction. “The problem in our health care system is that the bureaucratic aspect of things precludes us from being efficient…” [36, M, SURG] Patients suggested that surgeons could improve their communication and understanding of what is important to patients by: “Giving more time and listening to the patient. Assessing what they‟re saying, what the patient is saying. To give the patient time so that they feel comfortable enough to really express themselves.” [73, F, TKA] A good patient-provider relationship and open communication were believed to motivate the patient and facilitate recovery. Suggestions for opening channels of communication included providing patients with contact phone numbers, calling them when they had missed appointments and liaising with the next health care provider in the rehabilitation continuum to ensure timely and efficient „hand offs‟. “One of the things I feel is really important is that physiotherapy departments and physicians don‟t forget their patients. …call and see what‟s going on. Many people seem to feel like they were forgotten and that after physio and they were out on their own, nobody cared.” [73, M, TKA] 29 2.4.1.2 Theme 2: Expecting the unexpected Patients identified a number of unexpected challenges in the post-operative period for which they felt inadequately prepared: pain management, sleep disturbances, psychological issues and unrealistic activity expectations. “Nobody said how much pain and swelling there was going to be.” [76, F, TKA] “I think a lot of surgeons forget you‟ve got to sleep – honest to God, they should have to go through it.  The first thing is you‟d be offered, you know, adequate pain medication post-operative and then that sleep is the biggest factor that you‟re faced with.” [73, M, TKA and retired health professional] “I don‟t know how many people [with TKA] I‟ve had in the last little while that come in and they‟re stunned that they have pain postoperatively…They‟re so not prepared for the amount of pain they have.” [43, F, PT] “…after surgery I felt like the bull AND the china shop. Like I feel I am potentially the source of my demise and I feel fragile.” [57, F, THA] Of equal concern to many patients and health professionals were the issues of who to go to when post-operative pain was not well-managed and inconsistent advice on whether additional analgesics (e.g. narcotics) were appropriate after the initial acute care period. “I don‟t think anybody tells the patients, so they go home, they‟ll be getting some T3‟s or something by their surgeon or surgical RN and sometimes that‟s enough, but usually it‟s not enough… and they just don‟t think to call or they don‟t know who to call.” [41, F, FP] “…the knowledge of pain management from the patient‟s perspective and their primary care provider‟s perspective is very poor.” [55, M, SURG] All study participants viewed the pre-operative education and preparatory phase as being critical for clarifying expectations and empowering the patient. 30 “What I‟ve noticed is the [acute care] discharges tend to go better if patients are clear on the expectations, you know, that they‟re informed of the possible date of discharge so psychologically they can start to prepare themselves.  Involving social work early on to assist with addressing the barriers or obstacles I find goes well. [42, F, SW] Unclear or unrealistic patient expectations were felt to lead to greater post-operative pain, significant anxiety and depression, and disappointment around the rate of recovery. “I think my expectations on the recovery period were overly optimistic.” [57, M, TKA] “They should be realistic in what they project for you.” [69, M, THA] Differing expectations and views between surgeons and rehabilitation providers on patients‟ functional status, ongoing need for supervised physical therapy and achievable outcomes lead to inconsistent advice, patient confusion, premature discontinuation of therapy and less than optimal outcomes. A PT described a common scenario whereby the surgeon‟s assessment differed from that of the treating therapist. [The surgeon tells the patient at the 6-8 week follow-up visit] “„Oh, you‟re doing great. You don‟t need to do any more (physical therapy).‟ Well, they‟re not doing great. I don‟t think they‟re gotten the best bang for their buck as far as the surgery, and … you‟d like to see them progress a lot further than they are…” [43, F, PT] Health professionals voiced concerns about misinformation available through the popular press and commercial Internet sites and said this was a growing problem leading to unrealistic expectations and a negative impact on patient recovery. “Patients learn just enough to be dangerous [from the Internet].” [39, M, PT] 31 2.4.1.3 Theme 3: It‟s attitude that counts Health providers and patients alike stressed the importance of the patient‟s attitude when it came to being an active participant in the rehabilitation process and remaining motivated during the typical ups and downs of recovering from TJA surgery. Physicians and AHPs felt a key part of their roles was to help the patient in this regard: “I like to empower the patient first and foremost.” [47, M, PHYS] Patients were considered an integral part of the team and their active participation in the rehabilitation process vital to good outcomes and greater satisfaction. “I tell them „This is what you need to do at home‟ and they go home and don‟t practice, definitely that makes a huge difference when you see the patient next time. People are afraid to move or people are really reluctant to do it, so I think patient compliance with home exercises is very effective, it‟s huge.” [42, M, PT] “I would think that a person should be checked to make sure that they are continuing to exercise, they are using the leg. I think it‟s such a waste of money and time if you don‟t become better.” [61, F, TKA] Having a positive attitude and taking a proactive approach to the surgery and subsequent rehabilitation phase while acknowledging the mind-body connection were strategies used by many patient participants. “I learned to recognize that my body was wiser and far cleverer than I was so I had better just obey it.” [77, M, THA] “You have to be willing to give not just 100 percent but 150 percent to your own recovery.” [46, F, THA] 2.4.1.4 Theme 4: It takes all kinds of support Participants reported how different „facets‟ of support contributed to health outcomes and overall satisfaction with the surgery and rehabilitation process. Patients and AHPs were more likely than 32 physicians to describe peer and spousal/family support as having favorable effects on an individual‟s rehabilitation process. “Hearing from another patient first hand and how they experienced it really helps the fear part of it.” [43, F, RN] “The support from my spouse and my family immediately after surgery was the most invaluable and wonderful. Because we are sent out of the hospital faster now and you‟ve just got to have that help at home.” [64, F, TKA] The important role of family was acknowledged in descriptions of one health care facility where a family member was given the designation of „coach‟ and encouraged to participate in all aspects of the patient‟s rehabilitation. Involving a family member was also ideal in cases where cultural differences and language barriers impeded rehabilitation instruction. When spousal and/or family support was lacking, there was greater need for home support services. In several communities, a lack of such services coupled with few transitional care units/beds was felt to contribute to longer acute hospital stays and a group of patients “who fall into the cracks”. Patients wanted to be recognized as a whole person and valued a holistic approach, which was sometimes lacking. Patients shared stories of how feeling supported enhanced their recovery and coping. “After the [physical therapy] program she phoned me and asked me how I was doing, so that was pretty good. It gives a little bit of feedback to the people and they feel inside that at least somebody cares about them.” [58, M, THA] Physicians discussed their role in supporting and counseling TJA patients, however, both family practitioners and specialists expressed concerns over their ability to spend sufficient time with patients. The „system‟ was most often blamed for not allowing for protracted conversations with patients: “Physicians don‟t get paid adequately to provide counseling on an ongoing basis to 33 patients.” [62, M, FP]  Patients also expressed their frustration in accessing their surgeon post- operatively. “Does anyone find it important to have access to your surgeon, which is almost impossible? Anything, just hearing him, you know, on the phone even. Maybe you want to say something that‟s been bothering you and I‟m sure you‟re not the only one that‟s ever bothered, but you feel reassured.” [75, M, TKA] Another area of professional support overlapped with communication concerns; it was believed that health professional advice and guidance should be more consistent to be helpful: “… and I know that we can‟t all give the same exercises but I think everyone – we all have slightly different messages, we say slightly different things as to how long it‟s going to take or talking about the wound or talking about pain management. It would be really good if we could have some sort of education or something that‟s a little bit more consistent as far as the message that‟s going out for people.” [43, F, PT] “They‟re not standardized. I‟m just thinking, there‟s all sorts of physiotherapy clinics around and they all do different kinds of things….” [57, F, THA] “… it‟s really inconsistent among physicians in terms of who gets referred to home care and who gets referred to outpatient. There‟s no consistency… especially between health regions.” [31, F, OT] 2.4.1.5 Theme 5: Barriers to recovery Participants identified patient, provider and system level factors as being barriers to recovery after TJA. Patient factors such as pain coping, motivation, attitude, state of readiness for treatment, psychological distress and self-efficacy were felt to influence the acute care hospital stay, course of recovery and participation in rehabilitation. “…pain management after total knee replacement is probably one of the biggest barriers to recovery.” [55, M, SURG] 34 “One of the most common [concurrent] diagnoses that gets noticed is depression in the patients … which hugely affects motivation, adherence to the protocols, and follow up, and it doesn‟t get addressed frequently because primary care physicians don‟t take the time to diagnose it appropriately.  It‟s probably the most widely under diagnosed and under treated condition.” [53, F, RN]  Physicians and surgeons saw the role of rehabilitation after TJA as being “to enhance the safety of the [surgical] procedure and make it easier for the patient to recover.” [64, M, SURG] However, the quality of rehabilitation, and in particular physical therapy services, was frequently thought to be poorly administered and therefore more detrimental to patients‟ recovery than helpful. “I have little faith in the ability of the external providers to provide appropriate care for my patients and I tend to dissuade them from pursuing outpatient physical and occupational therapy after surgery. …my experience has been that they [therapists] tend to do more harm than good.” [55, M, SURG] While several surgeons described having a good relationship with rehabilitation professionals and expressed confidence in their referral to post-operative physical therapy services, others did not: “We are sending them into a dark, black hole.” [60, M, SURG] At both patient and provider levels, language barriers and lack of translated educational materials were believed to compromise AHPs‟ ability to provide effective and timely education and support in a variety of rehabilitation settings. At the system level, issues related to access to rehabilitation were common to both Canadian and American participants; however, the contributing factors differed in important ways. Prolonged waits for surgical consultation, TJA surgery and in some cases, outpatient rehabilitation were unique to Canadian experiences. “…the Canadian system should be very clearly differentiated from the American. Their healthcare system is totally different. There‟s no similarity at all …we have the longest waiting list in the Western world.” [60, M, SURG] 35 “…when it comes to the physio after, there don‟t seem to be more physiotherapy spaces. We all experienced longer waits. And we‟ve all felt we‟ve developed slower because of this extra wait.” [64, F, TKA] Caps on physical therapy and rehabilitation services through private health insurers and managed health care practices were at issue in the American experience. Limited healthcare resources, ever-changing funding formulas and costs of rehabilitation services concerned all stakeholder groups in both countries.  “If [patients] don‟t do physio it‟s usually because it‟s going to be expensive and they don‟t have extended health [insurance].” [41, F, FP] “Medicare has put a cap on the amount of money that you can get in terms of the physical therapy and I think that‟s wrong. People vary too much in how they respond to surgery and to put a dollar value on that is totally crazy.” [72, F, TKA] With limited access to supervised rehabilitation, patients and providers had to decide how and when to use their „allotment‟. While some surgeons routinely sent people for physical therapy before surgery (pre-hab), others felt that rehabilitation postoperatively was of greater value. Barriers to rehabilitation services included limited access outside urban settings and larger hospitals. Patients typically had fewer if any options for publicly funded therapy in more remote areas of Canada. Travel and associated costs with receiving rehabilitation outside of their home community were problematic for patients. “It‟s been hard because I live so far away. It‟s about a two and half hour drive from here to [my rehabilitation setting].” [51, M, TKA] “I think the farther you get away from a hospital and whether you‟re talking doctors or physiotherapists, oftentimes you do move away from evidenced based practices…” [63, M, PHYS] 36 Suggestions for addressing issues related to access and quality of care in rural communities included greater use of tele-rehab and enhanced training for rehabilitation providers. 2.4.1.6 Theme 6: Back to normal This final theme reflects the common view that patients wanted nothing more than to return to a sense of normalcy after surgery. While being pain-free and mobile was of primary importance, a more holistic view of „normal‟ was repeatedly expressed. “I want to get back to be able to walk distances and participate in cross-country skiing, snow shoeing and hiking and fitness class, you know, things I did before.” [76, F, THA] “…I can only think of emerging from this cocoon of pain, which pulls you into a very small horizon. And so I really just wanted to get my vitality back.” [77, M, THA] “…to do my work is really just life‟s blood to me.” [64, F, TKA] “I was on crutches for 4 years and I have an 8-year- old daughter, so she‟d never really seen me walk without crutches and now I don‟t have them. So that was really important.  She sees me more as a normal person – now I can be the parent again.” [46, F, TKA] “…you don‟t want people losing their independent community skills so that they can stay out of nursing homes.” [53, F, RN] There was strong support for a holistic approach to conceptualizing and measuring outcomes from the patients‟ perspective. “Look at the whole person.  The psychosocial aspect is not always surgeons‟ strong suit.” [63, F, TKA] 37 “…I‟ve had times where I felt that everyone had an area of expertise and that me as a whole person, nobody was addressing or even wanting to hear about the total person going through this.” [64, F, TKA] Consistent with the diverse conceptualization of  „normal‟ as the desired outcome, ways of measuring outcomes varied greatly with no agreement on measurement approaches or the value of using standardized tools in clinical settings. Measures that could be used throughout the rehabilitation continuum were thought to be ideal. “…it would be nice for people to actually use the same outcome measures pre- operatively, immediately post-op… so you could actually see a difference.” [43, F, PT] “Some people I believe use the WOMAC. Some people use the Oxford. Some people have their own little compilation of different things, and I really don‟t know what they use off in private practice frankly. So big weaknesses and we don‟t have a standardized approach to this yet.” [63, M, PHYS] Others questioned the value of administering outcome tools and questionnaires. “…I think that you have got to be very careful about trying to quantify it at all. Questionnaires, I‟ve come across them before and I think this is stupid! And you put something down, you don‟t know how it‟s going to be interpreted.” [81, M, TKA] “I don‟t ask patients to fill out questionnaires. That‟s highly inefficient.” [56, M, SURG] When prompted to discuss the need for ongoing follow-up or long term monitoring of patients‟ outcomes, with the exception of surgeons, most felt that surgeons, primary care physicians and AHPs should be involved in follow-up care. Physical therapists were named most often as being able to offer an important complementary role to the surgeon‟s evaluation. 38 “The same team should follow the same patient, because the [surgeon] now, what‟s the first thing he does? “Okay, your x-ray looks great.” But the patient says, “I‟m not walking good.” We don‟t treat x-rays, we treat people, right?” [42, M, PT] 2.5 Discussion This paper describes the results of the initial exploratory phase of a mixed method project to develop practice guidelines for THA and TKA rehabilitation. A pragmatic approach was used to identify recurrent issues and important concepts for each of the broad discussion points in order to inform guideline development and ensure stakeholders‟ views were captured at the outset. A lack of communication coupled with poor appreciation for each other‟s roles and expertise appeared to be major issues among our study participants. This was most apparent with family physician-surgeon and PT-surgeon dyads, in less rural communities and between health care settings. Trust was also a dominant factor with many surgeons sharing concerns about the quality and safety of treatment approaches thought to be provided by outpatient PTs; PTs also lacked trust about other PT providers. Lack of trust could potentially be alleviated by improved communication to reduce the misunderstandings, conflicts, inefficiencies and role confusion that may arise and severely hamper patient care and outcomes (15, 32-34). Different professional training and cultures may explain some of the disparity in how health professionals communicate. For change to occur, it will require support at both the provider and system level (35). Greater opportunity for inter-professional dialogue is needed to truly enact team care within programs and across the continuum of care. Patients‟ perception of poor and inconsistent communication among their healthcare providers can negatively impact patient adherence, confidence, outcomes and satisfaction (36). Participants in this study spoke to the need for patient-provider communication to improve professionals‟ understanding of patients‟ beliefs and preferences and 39 clear, shared expectations regarding rehabilitation outcomes of TJA surgery. Disparities in expectations and evaluation of surgical outcomes are well documented with surgeons tending to rate outcomes more favorably than patients (13, 14). The intensity and duration of post-operative pain was common yet unexpected among patient participants in our study. Despite this information being readily available through previous studies (37-41), inadequate provider-patient communication and education may once again be at fault. We found there was both uncertainty and disagreement among patient and physician participants concerning professional responsibility for ensuring adequate pain control beyond the immediate post-operative period. Similarly, sleep disturbances described by patients in our study have been previously reported (39, 42) yet not adequately covered in pre-operative education sessions and virtually ignored during the early recovery phase. This was problematic for both patients and their spouses. Post-operative anxiety, depression, fear and vulnerability were widely reported by patients and of concern to many AHPs. While pre-operative psychological factors were not specifically probed in our study, the literature suggests that pre-operative depressive symptoms are strongly related to post-operative outcomes and satisfaction (37, 38, 41, 43). Study participants recommended pre-operative screening for depression and other factors that may contribute to protracted pain and psychological distress and improved surgeon awareness of such psychological factors. Emotional well being including more positive attributes (e.g., self-worth, hope, confidence, empowerment) is increasingly recognized as an important factor in coping and health outcomes of a number of chronic conditions and surgical procedures (36). Few studies examining the role of patient factors‟ in determining TJA need and outcomes have included these attitudinal factors in their analyses of important characteristics. Further, current orthopaedic outcome tools fail to capture the concepts of patient attitude, self-efficacy and empowerment (38) despite the evidence suggesting self-efficacy, for example, impacts patient expectations (44), long-term functional 40 outcome (45) and adherence to prescribed exercise (46). Poor adherence was a commonly held assumption of AHPs in our study and felt to be strongly related to patients‟ overall attitude about their role and outcome expectations of rehabilitation. Adherence to therapeutic protocols is problematic in many studies of TJA rehabilitation and warrants subgroup analysis to determine whether higher adherence (e.g., greater treatment dosage) results in larger treatment effects. These findings support adoption of self-efficacy theory to guide interventions, such as adopting efficacy enhancing strategies like contracting and role-modeling to enhance patient‟s confidence regarding the adoption of habits that will support their recovery (47). Our findings show that undergoing TJA surgery magnifies the need for support in the short term, consistent with other qualitative reports regarding the value of family and peer support to patients post-operatively (23). Better social support is associated with lower complication rates, better functional outcomes and higher post-operative quality of life (48). Patients described feeling supported by health professionals when they were „heard‟ and given sufficient time to have their questions and concerns addressed. Similarly, health professionals were most satisfied with their support efforts when they had adequate time to spend with the patient. Surgeons on the whole admitted to having little time to provide the support and guidance sought by most patients and this is equally problematic in Canada and the US. Concerns about poor health professional support were linked mostly to the follow-up (FU) phase, once supervised rehabilitation was completed. While the patients in our study had a TJA within the past year, several had undergone TJA surgery on another joint previously and expressed their dissatisfaction and feelings of being forgotten after rehabilitation ended. In a survey and chart review of 622 THA patients from three US states, only 41% reported consistent FU visits with their orthopaedic surgeons over a 6-year period and 16% reported they had no FU care (49). Older individuals and those with lower socioeconomic status were less likely to receive regular 41 FU. Our study patients suggested they would feel more supported in the year following TJA with regular phone calls, drop-in FU clinics with both surgeons and PTs, and group classes to review exercises, monitor progress and address any concerns. Personal, provider and system-level factors were identified by our study participants as creating barriers to patients‟ recovery after TJA. Hoppe et al. acknowledged rehabilitation as an important tool in reducing costs of disability regardless of cause (50). However, with “the rapid proliferation of private rehabilitation services currently operating with little regulation” (p.18), those using, prescribing and paying for the services are finding it increasingly difficult to determine if in fact, these services are of good quality, justified and cost-effective (50). In addition to other strategies, routine use of outcome measures and practice guidelines is suggested as a means of justifying and standardizing treatment approaches to address the structure, process and outcomes of the rehabilitation system. Capping the number of visits or duration of rehabilitation may help to control costs but as identified in our study, such limits were felt to hinder the rehabilitation process, ignore individual patient needs, and potentially lead to poorer outcomes and an overall increase in direct and indirect costs (50). The issue of timely access to surgical care has been a priority of provincial healthcare ministries in Canada for several years and the focus of several innovative quality improvement strategies (34, 51, 52). However, little attention and additional funding have been directed toward addressing barriers to quality rehabilitative care following surgery. Access, including transportation concerns, to rehabilitation services continues to be problematic for Canadians and Americans living in more rural settings. Greater use of technology including telerehabilitation (e.g., videoconferencing, remote monitoring) was voiced as a possible solution and deserves further investigation in this patient population (53). 42 Sanderson et al. reported clinicians and patients have different perspectives on outcomes and whereas patients‟ conceptualization of valued outcomes is broad, health professionals tend to focus on pathology and functional disability (54). We found a similar trend with patients describing a wide range of anticipated and expected outcomes covering many dimensions of health and psychosocial well-being while health professionals, in particular physicians and surgeons, focused more on impairment, basic function (e.g. walking, using stairs) and surgical parameters (e.g., fixation of implant). These incongruent views may play a role in the reported discrepancies between patients‟ and health professionals‟ evaluation of surgical outcomes in which there are moderate correlations at best between patient and clinician assessment of symptoms and disability (55). Few health professionals reported routinely using standardized outcome measures in their surgical and clinical practices, despite considerable support for their use. Participants‟ negative views on the utility (e.g., meaningfulness of numerical scores) and feasibility of using such instruments in clinical practice (e.g., time to administer and score) contributed to the low rate of standardized outcome evaluation. Jette et al. reported that a lack of support (e.g., technology, staffing) and irrelevant and confusing questions were barriers to routine use (56). Further, the apparent confusion among health professionals regarding what constituted an outcome measure may have led to underreporting and suggests more education is needed. Across all themes was the overarching view that “hips and knees are two different beasts” with different patterns of recovery and rehabilitation needs. Equally stressed was that younger and more active individuals have different outcome expectations and rehabilitation needs than older or more sedentary patients. Study participants saw little value in a „one size fits all‟ approach when designing rehabilitation programs, identifying recovery milestones and determining 43 outcomes. This is important to keep in mind as we design rehabilitation practice guidelines for a broad target audience. 2.5.1 Strengths of the study The credibility and trustworthiness of findings were enhanced by using a standardized discussion guide, multiple data sources, peer and member checking, independent coding and maintenance of an audit trail throughout the data collection and analysis phases. This study provides new data on specific inter-professional communication issues and barriers to recovery after TJA and shares insight from two vastly different health care systems. Further, it adds to the research on protracted post-operative pain, sleep disturbance and anxiety well beyond the immediate post- operative stage, which all stakeholders agree are inadequately and inconsistently managed. The perspectives of patients and health care providers alike are important to ensuring the relevance of practice guidelines, which are extremely time-consuming and expensive to produce (57) and it is imperative to guideline adoption that all viewpoints be carefully considered. The patient perspective was maintained by having patient-only focus groups and using their language to label the themes. Racial differences in patient-provider communication and the expectations and utilization of joint replacement therapy have been described elsewhere (58, 59), however, we could find no published information specifically on the experiences of Native North Americans undergoing TJA. The issues of geographical isolation and access to TJA rehabilitation care in a remote First Nation community identified in this study warrant further exploration. 2.5.2 Limitations Due to delays in the ethical review process incompatible with project timelines, only one US site was involved. It is unlikely that we heard the diversity of experiences and health care delivery 44 issues that are inherent in a country with no universal healthcare program and varied access to health insurance. As well, the attitudes, functional limitations, access to specialty care, and rehabilitation experiences of uninsured individuals were not captured and may differ from the individuals in our study. Secondly, physician/surgeon focus groups were challenging to organize and did not include as much practice settings diversity as intended. Physicians‟ views may not be transferable to those practicing in more rural settings with less access to rehabilitation resources for their patients. Similarly, despite efforts to ensure maximum diversity in patient participants, the experiences of less educated individuals and those not receiving formal rehabilitation services were underrepresented. 2.5.3 Clinical implications There are several take home messages for clinicians, most of which are directly aligned with principles of client-centered practice (60) aiming to individualize intervention for optimal client outcomes as well as best use of therapeutic resources:  Prior to surgery, ensure patient and provider expectations are clearly communicated and realistic;  Prior to surgery, develop a plan for addressing post-acute pain management, psychological distress and sleep disturbances for several weeks following surgery;  Use strategies to enhance self-efficacy and empower patients to adopt a positive attitude and take an active role in their rehabilitation;  Incorporate efficient approaches to optimize health professional support and follow-up care beyond three months after TJA; 45  Where possible, engage family members and peers in education, counseling and exercise instruction;  Select meaningful outcome measures and consistently use to evaluate effects of interventions throughout the care continuum and across health care settings. 2.5.4 Future research directions This study raises a number of questions that could be addressed through future research including an examination of communication and information technologies (e.g., telerehabilitation) on patient-provider and inter-provider communication and delivery of TJA rehabilitation services. Development and testing of a decision aide or screening tool would assist health care providers in identifying patients at risk for protracted pain, emotional distress and functional impairment. Further, there is a need to design, implement and evaluate the effects of a range of FU programs on patient satisfaction and long-term outcomes after TJA. 2.6 Conclusions This qualitative, exploratory study provides valuable insight into rehabilitation experiences, attitudes and expectations of individuals who have undergone THA or TKA surgery and the health professionals directly involved in their care. Patients offered a perspective that differed, but overlapped, with the perspectives of health professionals regarding rehabilitation practices and outcomes. Themes arising from all stakeholder groups related to communication, unexpected events, importance of patient attitude and active involvement, professional and social support, barriers to recovery and a return to normalcy. Awareness of the facilitators and barriers to achieving optimal outcomes that emerged from this study will help clinicians and administrators in the design and delivery of pre- and post-operative interventions aimed at helping patients reach their desired goals after TJA. Stakeholders‟ views on rehabilitation for TJA will inform the 46 next phases of guideline development and ensure all perspectives shape guideline priorities, scope, and format. 47 2.7 References 1. Canadian Institute for Health Information (CIHI). Hip and total knee replacements in Canada - Canadian Joint Replacement Registry (CJRR). 2008-2009 Annual Report. [cited 2009 Sep 17]. Available from: URL: www.cihi.ca/cjrr. 2. American Academy of Orthopedic Surgeons. Facts on hip replacements and facts on knee replacements. [cited 2009 Dec 9]. Available from: URL: http://www.aaos.org/research/stats/patientstats.asp. 3. Kim S. Changes in surgical loads and economic burden of hip and knee replacements in the US: 1997 - 2004. Arthritis Rheum 2008;59(4):481-8. 4. Jaglal SB, MacKay C, Corrigan L. Rehabilitation for total joint replacement. In: Badley E, Glazier R, editors. Arthritis and related conditions in Ontario. ICES research atlas. 2nd ed. Institute for Clinical Evaluative Sciences; 2004. p. 133 - 146. [cited 2008 Oct 1]. Available from: URL: http://www.ices.on.ca/file/ACF13B.pdf. 5. Lingard EA, Berven S, Katz JN, Kinemax Outcomes Group. Management and care of patients undergoing total knee arthroplasty: Variations across different health care settings. Arthritis Care Res 2000;13(3):129-36. 6. Mahomed NN, Lau JTC, Lin MKS, Zdero R, Davey JR. Significant variation exists in home care services following total joint arthroplasty. J Rheumatol 2004;31(5):973-5. 7. Mauer KA, Abrahams EB, Arslanian C, Schoenly L, Taggart HM. National practice patterns for the care of the patient with total joint replacement. Orthop Nurs 2002;21(3):37-47. 8. Roos EM. Effectiveness and practice variation of rehabilitation after joint replacement. Curr Opin Rheumatol 2003;15(2):160-2. 9. Antoniou J, Martineau P, Filion K, Haider S, Zukor D, Huk OL, et al. In-hospital cost of total hip arthroplasty in Canada and the United States. J Bone Joint Surg Am 2004;86-A(11):2435-9. 10. Lavernia CJ, D‟Apuzzo MR, Hernandez VH, Lee DJ, Rossi MD. Postdischarge costs in arthroplasty surgery. J Arthroplasty 2006;21(6 Suppl 2):144-50. 11. Medical Advisory Secretariat, Ontario Ministry of Health and Long-Term Care. Physiotherapy rehabilitation after total knee or hip replacement. Health technology literature review. 2005 May [cited 2005 Dec 6]. Available from: URL: http://www.health.gov.on.ca/english/providers/program/mas/tech/ohtas/tech_rehabtkr_061705.ht ml. 12. National Institutes of Health. NIH Consensus Development Conference on total knee replacement. NIH Consens State Sci Statements 2003 Dec 8-10 [cited 2004 Mar 17];20(1):1-34. Available from: URL: http://consensus.hin.gov/2003/2003TotalKneeReplacement117PDF.pdf. 13. Brokelman RB, van Loon CJ, Fijnberg WJ. Patient versus surgeon satisfaction after total hip arthroplasty. J Bone Joint Surg Br 2003;85(4):495-8. 48 14. Lieberman JR, Dorey F, Shekelle P, Schumacher L, Thomas BJ, Kilgus DJ, et al. Differences between patients‟ and physicians‟ evaluations of outcome after total hip arthroplasty. J Bone Joint Surg Am 1996;78-A(6):835-8. 15. Street RL, Richardson MN, Cox V, Suarez-Almazor ME. (Mis)understanding in patient- health care provider communication about total knee replacement. Arthritis Rheum 2009;61(1):100-7. 16. Haworth RJ, Hopkins J, Ells P, Ackroyd CE, Mowat AG. Expectations and outcome of total hip replacement. Rheumatol Rehabil 1981;20(2):65-70. 17. Hewlett SA. Patients and clinicians have different perspectives on outcomes in arthritis. J Rheumatol 2003;30(4):877-9. 18. Guyatt GH, Devereaux PJ, Lexchin J, Stone SB, Yalnizyan A, Himmelstein D, et al. A systematic review of studies comparing health outcomes in Canada and the United States. Open Med 2007;1(1):E27-36. 19. Kitzinger J. Qualitative research: Introducing focus groups. Br Med J 1995;311(7000):299- 302. 20. Morgan DL, Krueger RA. The focus group kit. Thousand Oaks (CA): Sage Publications; 1998. 21. Kidd PS, Parshall MB. Getting the focus and the group: Enhancing analytical rigor in focus group research. Qual Health Res 2000;10(3):293-308. 22. Showalter A, Burger S, Salyer J. Patients‟ and their spouses‟ needs after total joint arthroplasty: A pilot study. Orthop Nurs 2000;19(1):49-57. 23. Jacobson AF, Myerscough RP, DeLambo K, Fleming E, Huddleston AM, Bright N, et al. Patients‟ perspectives on total knee replacement. Am J Nurs 2008;108(5):54-63. 24. Fielden JM, Scott S, Horne JG. An investigation of patient satisfaction following discharge after total hip replacement surgery. Orthop Nurs 2003;22(6):429-36. 25. Mancuso CA, Sculco TP, Wickiewicz TL, Jones EC, Robbins L, Warren RF, et al. Patients' expectations of knee surgery. J Bone Joint Surg Am 2001;83-A(7):1005-12. 26. Middleton S, Lumby J. Measuring outcomes from the patients‟ perspective. Int J Nurs Pract 1999;5:143-6. 27. Heaton J, McMurray R, Sloper P, Nettleton S. Rehabilitation and total hip replacement: Patients‟ perspectives on provision. Int J Rehabil Res 2000;23(4):253-9. 28. Hollis V, Openshaw S, Goble R. Conducting focus groups: Purpose and practicalities. Br J Occup Ther 2002;65(1):2-8. 49 29. Krueger RA, Casey MA. Focus groups: A practical guide for applied research. 3rd ed. Thousand Oaks (CA): Sage Publications; 2000. 30. Mays N, Pope C. Assessing quality in qualitative research. Br Med J 2000;320(7226):50-2. 31. Pope C, Ziebland S, Mays N. Qualitative research in healthcare: Analyzing qualitative data. Br Med J 2000;320(7227):114-6. 32. Suter E, Arndt J, Arthur N, Parboosingh J, Taylor E, Deutschlander S. Role understanding and effective communication as core competencies for collaborative practice. J Interprof Care 2009;23(1):41-51. 33. Teutsch C. Patient-doctor communication. Med Clin North Am 2003;87(5):1115-45. 34. Frank C, Dick D, Smith D, Wasylak T, Gooch K, Zernicke R. The Alberta Bone and Joint Health Institute: Creating sustainable accountability through collaboration, relevant measurement and timely feedback. Healthc Pap 2006;7(1):34-9. 35. Weinberg DB, Gittell JH, Lusenhop RW, Kautz CM, Wright J. Beyond our walls: Impact of patient and provider coordination across the continuum on outcomes for surgical patients. Health Serv Res 2007;42(1):7-24. 36. Street RL, Makoul G, Arora NK, Epstein RM. How does communication heal? Pathways linking clinician-patient communication to health outcomes. Patient Educ Couns 2009;74(3):295-301. 37. Wylde V, Dieppe P, Hewlett S, Learmonth ID. Total knee replacement: Is it really an effective procedure for all? Knee 2007;14(6):417-23. 38. Dorr LD, Chao L. The emotional state of the patient after total hip and knee arthroplasty. Clin Orthop Relat Res 2007;463:7-12. 39. Hall-Lord ML, Steen B, Larsson G. Postoperative experiences of pain and distress in elderly patients. An explorative study. Aging Clin Exp Res 1999;11(2):73-82. 40. O‟Brien S, Bennett D, Doran E, Beverland DE. Comparison of hip and knee arthroplasty outcomes at early and intermediate follow-up. Orthopedics 2009;32(3):168. 41. Brander VA, Stulberg SD, Adams AD, Harden RN, Bruehl S, Stanos SP, et al. Predicting total knee replacement pain: A prospective, observational study. Clin Orthop Relat Res 2003;416:27-36. 42. Fielden JM, Gander PH, Horne JG, Lewer BM, Green RM, Devane PA. An assessment of sleep disturbance in patients before and after total hip arthroplasty. J Arthroplasty 2003;18(3):371-6. 43. Brander VA, Gondek S, Martin E, Stulberg SD. Pain and depression influence outcome 5 years after knee replacement surgery. Clin Orthop Relat Res 2007;464:21-6. 50 44. Mahomed NN, Liang MH, Cook EF, Daltroy LH, Fortin PR, Fossel AH, et al. The importance of patient expectations in predicting functional outcomes after total joint arthroplasty. J Rheumatol 2002;29(6):1273-9. 45. van den Akker-Scheek I, Stevens M, Groothoff JW, Bulstra SK, Zijlstra W. Preoperative or postoperative self-efficacy: Which is a better predictor of outcome after total hip or knee arthroplasty? Patient Educ Couns 2007;66(1):92-9. 46. Moon LB, Backer J. Relationships among self-efficacy, outcome expectancy, and postoperative behaviours in total joint replacement patients. Orthop Nurs 2000;19(2):77-85. 47. Nuñez DE, Keller C, Ananian CD. A review of the efficacy of the self-management model on health outcomes in community-residing older adults with arthritis. Worldviews Evid Based Nurs 2009;6(3):130-48. 48. Young NL, Cheah D, Waddell JP, Wright JG. Patient characteristics that affect the outcome of total hip arthroplasty: A review. Can J Surg 1998;41(3):188-95. 49. de Pablo P, Losina E, Mahomed N, Wright J, Fossel AH, Barrett JA, et al. Extent of follow- up care after elective total hip replacement. J Rheumatol 2006;33(6):1159-66. 50. Hoppe E, Rowat B, Verrier M. A black box: The rehabilitation clinic. Physiother Can 1996;48(1):16-26. 51. Markel F, Rafferty C, Rodgers J. Ontario Waiting List Project. Final Report. 2002 Aug [cited 2009 Feb 9]. Available from: URL: http://www.jppc.org. 52. Sanmartin C, Lewis S. From chaos to order: Making sense of waiting lists in Canada. Final Report. Western Canada Waiting List Project. 2001 [cited 2009 Feb 9]. Available from: URL: http://www.wcwl.org/. 53. Brennan DM, Mawson S, Brownsell S. Telerehabilitation: Enabling the remote delivery of healthcare, rehabilitation and self management. Stud Health Technol Inform 2009;145:231-48. 54. Sanderson T, Kirwan J. Patient-reported outcomes for arthritis: Time to focus on personal life impact measures? Arthritis Rheum 2009;61(1):1-3. 55. Bream E, Black N. What is the relationship between patients‟ and clinicians‟ reports of the outcomes of elective surgery. J Health Serv Res Policy 2009;14(3):174-82. 56. Jette DU, Halbert J, Iverson C, Miceli E, Shah P. Use of standardized outcome measures in physical therapist practice: Perceptions and applications. Phys Ther 2009;89(2):125-35. 57. Liang MH. The American College of Rheumatology practice guidelines strategy. Arthritis Care Res 1995;8(3):132-3. 58. Skinner J, Weinstein JN, Sporer SM, Wennberg JE. Racial, ethnic, and geographic disparities in rates of knee arthroplasty among Medicare patients. N Engl J Med 2003;349(14):1350-9. 51 59. Groeneveld P, Kwoh CK, Mor MK, Appelt CJ, Geng M, Gutierrez JC, et al. Racial differences in expectations of joint replacement surgery outcomes. Arthritis Rheum 2008;59(5):730-7. 60. Sumsion T, Law M. A review of evidence on the conceptual elements informing client- centred practice. Can J Occup Ther 2006;73(3):153-62.  52 Chapter 3: Post-acute physiotherapy following primary total hip arthroplasty for osteoarthritis: A Cochrane systematic review3 3.1 Background For individuals with advanced osteoarthritis (OA) that is not successfully managed with conservative measures such as use of analgesics, anti-inflammatory agents, walking aids, exercise and activity modifications, joint replacement surgery becomes the recommended treatment. Total hip arthroplasty (THA) surgery or total hip replacement (THR), is the most common surgical procedure for advanced end-stage OA of the hip. Approximately 22,000 primary THA surgeries were performed in Canada in 2006/07 (not including Quebec) (1). In the United States (US), this number is more than 10-fold greater (2). The number of THA procedures in Canada grew by 101% in the preceding 10-year period and by 7% in the preceding year (1). Similar one year growth rates are reported in other countries (2, 3). This number is projected to rise in an almost exponential fashion with the aging population, increasing prevalence of obesity and increased prevalence of OA (4-7). Other factors contributing to the increase are greater patient demand and public expectations for improved quality of life, shortened acute care length of stay (LOS) allowing for a greater number of procedures, advances in prosthetic design and materials, and improved surgical and anaesthetic techniques that have made these surgeries appropriate and safer for individuals who were not previously eligible (3, 5, 7, 8). In the US, an increase in number of orthopaedic surgeons with a focus on total joint arthroplasty (TJA) has further fueled the growth in number of these procedures (7). International data show that OA is the primary reason for surgery and contributes to between 81% (1) and 94% (9) of all primary THA procedures.  3  A version of this chapter will be submitted for publication. Westby MD, Carr S, Kennedy D, Brander V, Bell M, Doyle-Waters M, Backman C. Post-acute physiotherapy following primary total hip arthroplasty for osteoarthritis: A Cochrane systematic review. 53 While joint replacement surgery decreases pain, improves mobility and function, and improves quality of life (10-12) with minimal complications, THA surgeries place a significant burden on healthcare budgets (3, 5). Data from the American Academy of Orthopaedic Surgeons (AAOS) indicate that 2005 total hospitalization costs for primary THA procedures in the US was $9.2 billion (USD) (2). In the past decade, days spent in hospital for THA has decreased by 38% in Canada (1) and by 49% (1993 to 2005) in the US (7). Data from 2006/07 show that the average acute hospital length of stay (LOS) in Canada to be seven days (median five days) for all THA procedures, including revisions (1). A significantly shorter hospital stay (mean LOS 4.2 days) is reported in the US when only primary procedures are considered (2). Earlier discharge decreases the time available for recuperation, early rehabilitation, and patient and family education and counseling. Reduced LOS places additional burden and responsibility on the patients, their family and their post-acute health care providers to monitor for and address post-operative complications such as wound infection and deep vein thrombosis (DVT). The short stay puts more emphasis on preadmission education and the role and timing of post-acute physiotherapy interventions following discharge. In an editorial examining LOS after TJA, Johanson comments that "it remains unclear just how far this process [of reducing LOS] can be taken without either compromising quality of care or simply shifting costs to a less regulated outpatient environment" (13) (p.1). To date, no evidence-based practice guidelines exist to inform best practice for post-acute rehabilitation following THA in North America and specific rehabilitation practices vary greatly among providers (14-16). Current therapeutic interventions are based largely on clinical experience (17), clinician and surgeon preferences (18), the acute care phase of recovery (19, 20), facility-based protocols (21), health insurance funding schemes (7) and outdated approaches 54 (22). Furthermore, little or no guidance exists on recommended activities following discharge from the acute care setting (23). All but two international joint replacement registries (8, 24) limit their data collecting and reporting to surgical outcomes such as peri-operative complications and revision rates. Rehabilitation and its contribution to both surgical and patient- centred outcomes are an overlooked aspect of the total episode of care for THA patients. There are a number of reports suggesting that some patients are not reaching their physical potential following THA and that post-operative pain, physical impairments and functional limitations are still issues more than one year following THA (10, 25-27). Current exercise programs during the early phase of rehabilitation are insufficient to restore muscle strength, normal gait and more complex functional activities in some patients (25). However, we do not know exactly what kind of physiotherapy performed over what period of time is effective or necessary to enhance short- and long-term surgical outcomes and optimize patient activity, participation and health-related quality of life (HRQoL). The National Institutes of Health (NIH) conference on THA concluded: “The contribution of pre-hospital, in-hospital and post-hospital education and rehabilitation programs to the eventual outcome of the surgical procedure deserves an organized, in-depth study to determine optimum regimen, duration of treatment, and expected outcomes.” (28)(p. 8). As part of the larger project to develop evidence-based clinical practice guidelines for post-acute rehabilitation after primary THA and total knee arthroplasty (TKA), systematic reviews were undertaken to examine the current literature on various rehabilitation approaches and interventions in this patient population. Previous reviews have addressed multidisciplinary 55 rehabilitation for primary and revision THA and TKA (29), pre-operative education for THA and TKA (30) and most recently, therapeutic exercise after discharge from hospital for THA (31). 3.2 Purpose The primary aim of this review is to assess the effects of post-acute physiotherapy following primary THA for OA on patient-centred outcomes of pain, physical function and HRQoL. 3.2.1 Objectives The specific objectives are to: 1. Document the short-term and long-term effects of post-acute physiotherapy on pain, physical function and HRQoL in persons undergoing primary THA for OA 2. Compare outcomes across different clinical settings and treatment approaches, timing and dosages 3. Make recommendations for clinical practice based on the strength of the evidence 4. Identify research needs in post-acute physiotherapy in the THA population 3.3 Methods 3.3.1 Criteria for considering studies for this review We considered both randomized controlled trials (RCTs) and controlled clinical trials (quasi- randomized and controlled before-after designs) for inclusion in this review. Designs not included were cohort, case-control, single case studies, single subject, case series and pre/post studies with no control group. 3.3.2 Types of participants Primary trials of individuals aged 19 years and older who underwent an elective, primary THA for primary (idiopathic) or secondary OA due to chronic trauma, developmental or congenital 56 problems were considered. Surgical procedures included all forms of fixation (cemented, hybrid or cementless with/without porous coating), surgical approaches (anterolateral, lateral, posterolateral, minimally invasive) and types of prosthetic bearing surfaces (metal, polyethylene, ceramic). Studies were excluded if participants had:  undergone a revision THA, hemi-arthroplasty or resurfacing procedure;  an inflammatory arthritis diagnosis (i.e. rheumatoid arthritis, ankylosing spondylitis);  an acute fracture;  significant pre- or peri-operative complications (e.g., intraoperative fracture, nerve palsy, deep wound infection, severe anaemia, DVT)  an extended acute hospital stay beyond the standard LOS for that hospital facility. Studies with mixed pre-surgical diagnoses were included if 90% or more of the participants had an OA diagnosis. 3.3.3 Types of interventions 3.3.3.1 Experimental Post-acute physiotherapy includes physiotherapy provided in any setting (in-patient, outpatient or home/community). The 'post-acute' phase was determined to extend from immediately following discharge from the acute care ward or setting and up to 12 months post-surgery. The time frame selected reflects patients' views on the duration of rehabilitation services (32). Physiotherapy includes 1 to1 treatments and group interventions supervised by the physiotherapist (PT) and/or a trained physiotherapy assistant under the direct supervision of a PT. Interventions could include alone or in combination:  therapeutic exercise;  hydrotherapy; 57  use of thermal or electrical modalities (i.e. cryotherapy, heat, transcutaneous electrical nerve stimulation (TENS) and neuromuscular electrical stimulation);  postural, proprioceptive and balance training;  gait training;  functional exercises;  cardiovascular training;  manual therapy techniques (i.e. soft tissue massage, passive range of motion (ROM), hold-relax);  patient education and self-management techniques. Interventions not included in this review were splinting or bracing, and chiropractic or osteopathic manipulation. We characterized interventions by:  timing of when treatment initiated (immediate = < 2 weeks post-surgery, intermediate = between 2 weeks and 3 months post-surgery or delayed = > 3 months post-surgery)  setting (outpatient, inpatient or home-based)  dosage (lower = < 6 visits or 3 weeks of intervention and higher = ≥ 6 visits or 3 weeks of intervention) 3.3.3.2 Control Supervised physiotherapy was compared to control situations in which participants received „routine care‟ (standard or traditional physiotherapy of a given institution or surgeon), „attention only‟ (no active intervention), unsupervised home exercises or other operationalized comparators (which may include comparing one physiotherapy program or therapeutic setting to another). 58 3.3.4 Types of outcome measures The primary outcomes of interest were:  Post-operative pain (at rest or with activity);  Physical function (as measured by validated functional assessment self report tools, surgeon-rated tools or performance-based measures);  HRQoL (as measured by validated generic, condition-specific, individualized or preference-based instruments);  Adverse events due to rehabilitation (e.g. increased pain, fall).  Secondary outcomes were categorized according to the International Classification of Functioning (ICF) (33) but were not included in the analysis:  Body structure and function (e.g. hip ROM, lower extremity muscle strength and endurance, soft tissue flexibility, proprioception, cardiovascular fitness)  Activity (i.e. those not captured under physical function outcomes)  Participation (e.g. paid/unpaid work, care giving, leisure activities, social activities)  Personal contextual factors (e.g. age, gender, body mass index, patient satisfaction, self- efficacy, anxiety, motivation)  Environmental factors (e.g. health insurance coverage, waiting time for surgery)  Process measures:  Timing of treatment intervention  Timing of baseline assessment (pre-surgery or post-surgery/pre-intervention)  Treatment setting (inpatient, outpatient, home)  Treatment dosage (e.g. frequency, duration, number of sessions) 59  Adherence to treatment 3.3.5 Search methods for identification of studies We searched the following electronic databases to identify relevant studies published between January 1990 and August 31, 2008. The period of time covered was recommended by the expert panel and covers the period of time when inpatient LOS fell and diagnostic related groups (DRGs) became widespread basis for reimbursement in the US by this year.  Cochrane Musculoskeletal Group Trials Register (September 17, 2008)  Cochrane Central Register of Controlled Trials (CENTRAL)(Issue 3, 2008)  MEDLINE (OVID) (1950 to August 31, 2008)  EMBASE (OVID) (1980 to 2008 Week 39)  CINAHL (EBSCO) (1981 to August 31, 2008)  Cochrane Database of Systematic Reviews (Issue 3, 2008), ACP Journal Club and DARE (OVID)  PEDro (Physiotherapy Evidence Database) (August 31, 2008)  Web of Science (ISI) (August 31, 2008) Articles in any language were reviewed. See Appendix D for an example of the search strategy for MEDLINE using a combination of MeSH and key words. The quality filter recommended by the Cochrane Musculoskeletal Review Group was applied. We scanned articles, review papers and textbooks for additional papers. Known experts in the field and authors of existing, high quality studies were contacted for further references. A grey literature search was conducted to identify published and unpublished papers, reports and other documents. We searched papers and proceedings from congresses and symposiums through the databases PapersFirst and ProceedingsFirst. Additionally, we hand searched conference 60 proceedings of key professional organizations, highly relevant orthopaedic and rehabilitation journals, professional organizations' websites, and national joint arthroplasty registries. 3.3.6 Data collection and analysis A team of four primary reviewers (MW, SC, DK and VB) and two adjudicators (CB and MB) was assembled. Three of the four reviewers were trained in the Cochrane review methodology. The lead reviewer (MW) and one other primary reviewer (SC) independently screened all identified papers by scanning titles and abstracts for appropriateness (first level screening). Pairs of reviewers then evaluated potentially relevant studies to see if they met our inclusion criteria (second level screening). The same pair then independently extracted data from accepted papers using a standardized form. Any disagreement was resolved through discussion and consensus. There was no situation in which a tie breaker was needed. If necessary, study authors were contacted for additional information to complete the screening and abstraction process. A Swedish (E.R.) and German (M.B.) colleague reviewed and translated some papers and abstracted the necessary data. These were then checked by two independent reviewers and any further questions were discussed with the foreign language colleague. 3.3.7 Assessment of methodological quality We assessed the risk of bias and overall methodological quality of the studies using the scale by van Tulder et al. (34) for RCTs and controlled clinical trials (CCTs). This 11-item quality assessment scale developed by the Cochrane Back Review Group was chosen for its relevance (Table 3.1). While designed as a checklist for aggregating items and generating a single quality score, this process is no longer favoured (35) and we present the individual quality criteria results for each trial. 61 Criteria Scoring 1. Was the method of randomization adequate? yes no don't know 2. Was treatment allocation concealed? yes no unclear 3. Were groups similar at baseline regarding most important prognostic factors? yes no don't know 4. Was the patient blinded to the intervention? yes no don't know 5. Was the therapist/care provider blinded to the intervention? yes no don't know 6. Was the outcome assessor blinded to the intervention? yes no don't know 7. Were co-interventions avoided or similar? yes no don't know 8. Was compliance acceptable in all groups? yes no don't know 9. Was the drop-out rate described and acceptable (20% or less for post- test, 30% or less for long term follow up) yes no don't know 10. Was the timing of the outcome assessment similar in both groups? yes no don't know 11. Was data for at least one key outcome analyzed by intention-to-treat (ITT)? yes no don't know   Table 3.1 Quality Assessment and Risk of Bias Checklist Table 3.1 Quality Assessment and Risk of Bias Checklist  Initial inter-observer reliability of both the screening and quality assessment was determined on two papers prior to the actual review process. These related studies were drawn from a pool of articles that did not meet the inclusion criteria due to publication date or difference in intervention or patient population (34). Poor agreement on some items during this pilot test phase led to revisions to the screening and data extraction forms. After further discussion, agreement was reached on the screening and data abstraction process. Comprehensive guidelines for data abstraction were prepared and used by all reviewers. 3.3.8 Data analysis and synthesis Continuous data from individual trials were recorded as means and standard deviations. We contacted several authors for missing data. Evidence tables were created and a qualitative 62 analysis performed using the GRADE approach recommended by the Cochrane Musculoskeletal Group (36) (Table 3.2). Detailed methods for upgrading or downgrading studies are described in the Chapter 12.2.1 of the Cochrane Handbook for Systematic Reviews (35). Table 3.2 GRADE approach to assessing overall levels of quality of a body of evidence Underlying methodology Quality rating Randomized trials; or double-upgraded observational studies. High Downgraded randomized trials; or upgraded observational studies. Moderate Double-downgraded randomized trials; or observational studies. Low Triple-downgraded randomized trials; or downgraded observational studies; or case series/case reports. Very low   All raters independently examined the evidence tables for clinical and methodological heterogeneity taking into account the participants, interventions, controls (comparators), outcomes, timing of assessments, measurement tools and methodological quality of the original studies (34). Due to lack of both clinical and methodological homogeneity among trials, we were unable to pool the data and perform a meta-analysis. This is a common problem among Cochrane reviews of allied health interventions (37). We do, however, present the individual studies' treatment effect sizes using weighted mean differences (WMD) for those with mean post-test values and sufficient data. In addition, for those trials with small sample sizes (less than 50 subjects per group) (38) and differing baseline values, we also calculated WMDs using mean change scores to better reflect the true effect. Since original data sets were not available, we used the formula suggested in the Cochrane Handbook, Chapter 16.1.3.2 (39) and a conservative imputed value of 0.8 as the correlation between pre- and post-test scores. Smaller correlation coefficients are known to overestimate effect sizes and thus the more conservative value was 63 chosen (40). Forest plots were created to display effect estimates with 95% confidence intervals and samples are provided for select outcomes within individual trials. 3.4 Results 3.4.1 Description of studies A total of 1,538 English and non-English papers were identified. Of these, 98 were retained for second level screening and eight for full review (six RCTs, two CCTs) (Figure 3.1). Three of the eight articles were published in languages other than English (41-43) and all but one study was conducted outside of North America. All trials were published between 1998 and 2008. 64  Figure 3.1 PRISMA flow chart for THA systematic review  Potentially relevant RCTs & CCTs identified and screened for retrieval (n = 1, 538) RCTs & CCTs excluded with reasons (n = 1, 440) Main reasons:  Not THA-related  Not intervention study  Not a rehabilitation intervention  No control group RCTs & CCTs retrieved for more detailed evaluation (n = 98 ) RCTs & CCTs excluded with reasons (n = 90) Main reasons:  Multidisciplinary intervention  Significant pre-op component  Acute care intervention  No THA subgroup data  Mixed diagnoses (no OA subgroup data)  Not within review timeframe Potentially appropriate RCTs & CCTs to be included in the meta- analysis (n = 8) RCTs & CCTs excluded from meta-analysis with reasons (n = 8) (Due to clinical & methodological heterogeneity among trials) RCTs included in meta-analysis  (n = 0)  65 Several studies initially screened as being appropriate and subsequently excluded from this review are listed in Table 3.3. Table 3.3 Excluded studies and main reason for exclusion Study Main reason for exclusion Gilbey HJ et al. 2003 (44) Included significant pre-op component Jan M-H et al. 2004 (45) Initiated more than 1 year post-THA Mahomed NN et al. 2008 (46) No subgroup data for THA Weaver FM et al. 2003 (47) Included significant pre-op component Maire J et al. 2003 (48) Less than 10 subjects per group Sashika H et al. 1996 (49) Less than 10 subjects per group Siggeirsdottir K et al. 2005 (50) Included significant pre-op component   3.4.2 Methodological quality The range in methodological quality of the included studies was very low to moderate. Randomization procedures were adequate in three trials, inadequate in one, and not clear in four. Concealed allocation, where neither the researcher nor participants are aware of group assignment in advance, occurred in three studies. Only three trials reported blinding the outcome assessor. Blinding of the provider and patients is not possible in most cases in physiotherapy intervention studies; however, authors of one trial stated the participants were blind to the important details of the intervention. Short-term drop out rates were satisfactory in five studies and exceeded the acceptable short-term rate of 20% (34) in three. Two studies included a long- term follow-up phase (12 months or more) with satisfactory drop out rates in one and not in the other. (See Table 3.4) 66 Table 3.4 Methodological quality and risk of bias in included trials Criterion Galea 2008 Hesse 2003 Mayer 2005 Nyberg 2002 Scherak 1998 Ström 2006 Suetta 2004 Trudelle- Jackson 2004 1. Randomization ? Y Y* N ? ? Y Y 2. Concealment ? Y Y* N ? Y Y* Y 3. Groups similar Y Y N Y Y Y Y Y 4. Patient blinded N N N N N N N Y 5. Therapist blinded N N N N N N N N 6. Assessor blinded ? Y Y* N ? Y N N 7. Co-interventions/ contamination avoided ? Y ? ? ? N Y* ? 8. Acceptable compliance Y Y ? ? ? ? Y* Y 9. Acceptable drop-out rate (short term) Y Y Y N ? Y N Y 10. Timing of outcome assessment similar Y Y Y Y Y Y Y Y 11. ITT analysis NA Y NA N N ? N N Legend: Y = criterion was met; * = positive response was obtained by contacting the author and not from published article; ? = unclear, incomplete or no data provided; N = criterion was not met; NA = not applicable as there were no drop outs reported  3.4.3 Participants A total of 497 participants were included in this review. Individual study sample sizes ranged from 23 to 219 participants. Mean participant ages ranged from 54 to 79 years and a majority was female in all trials. Race was not identified in any studies. General health, presence of co- morbidities and details of the surgery were poorly described or not mentioned (Table 3.5). 67 Table 3.5 Summary of study characteristics Study/ language Design/ quality Setting/ country Participants (1° THA for OA) Exclusion criteria Intervention Comparison Outcomes Galea 2008  (English) RCT (n=23)  Low Outpatient rehabilitation centre, Australia TG = mean age 69 (10) yrs, 63% ♀. CG = mean age 67 (8) yrs, 50% ♀. Uncontrolled systemic disease, neurologic or orthopaedic condition affecting walking, > 4 wks of post-op physiotherapy Start: 8 wks post-op. Centre-based progressive exercise program, 2x/wk for 8 wks Home-based program with illustrated exercise guide, no further instruction Baseline: post-op TUG, stair climbing, 6MWT, WOMAC, AQoL  Adverse events not reported No LT follow-up Hesse 2003  (English) RCT (n=80)  Moderate Inpatient rehab facility, Germany Cementless THA. Combined mean age = 65 (13) yrs, 70% ♀. Orthopaedic/ neurologic disease impairing gait, history of DVT, symptomatic heart disease Start: 3 wks post-op. Body weight supported treadmill training + standard physio program, 5x/wk for 2 wks Standard inpatient physio program Baseline: post-op Harris Hip Score, walking speed, hip abduction strength.  Adverse events reported 12 month follow-up Mayer 2005  (German) RCT (n=24)  Low Outpatient physio dept, Germany Cementless THA. Combined mean age = 61 (range 40-85) yrs, 67% ♀. Hip fracture Start: 2 wks post-op Mental gait training + standard physio program, 3x/wk for 3 wks Standard outpatient physio program Baseline: post-op Walking speed, step length, stance phase.  No adverse events reported No LT follow-up Nyberg 2002  (Swedish)  CCT (n=45)  Very Low Outpatient dept, University Hospital, Sweden Cemented or hybrid THA. TG = mean age 69 (5) yrs, 65% ♀. CG = mean age 68 (5) yrs, 53% ♀. Revision surgery Start: 8 wks post-op Group “THR school” 2x/wk for 15 wks + standard home exercise program Standard home exercise program for 15 wks Baseline: post-op Pain (VAS), MACTAR, general health satisfaction (VAS), rating of exercise importance.  Adverse events reported No LT follow-up  68 Table 3.5 Summary of study characteristics (continued)  Study/ language Design/ quality Setting/ country Participants (1° THA for OA) Exclusion criteria Intervention Comparison Outcomes Scherak 1998  (German)  CCT, 3 treatment arms (n=219)  Low Inpatient rehab hospitals, Austria Combined mean age = 66 (8) yrs, 67% ♀. Unclear Start: Grp 1 = within 2 months post-op; Grp 2 = 3-4 months post-op; Grp 3 = 11 months post-op All grps received standard inpatient rehab program of exercises, gait training & hydrotherapy 5x/wk for 4 wks. Different start times for inpatient rehab program Baseline: post-op Walking speed, gait pattern, disablement tool, ROM.  Adverse events not discussed Partial LT follow-up Ström 2006  (English) RCT (n=36)  Low Setting not clear (gym- based?), Sweden Cementless THA. Combined mean age = 54 (range 25-63) yrs, 53% ♀ None given Start: immediately post-op Unrestricted weight bearing & supervised, progressive exercise program (gym + pool based) 2-3x/wk for 12 wks Partial weight bearing & unsupervised home exercise program for 12 wks Baseline: pre-op Merle d‟Aubigne Score, hip abduction strength & ROM, peak weight bearing load.  Adverse event reported 12 month follow-up Suetta 2004  (English) RCT, 3 arms (n=36)  Low Outpatient hospital & home, Denmark Cemented THA, ASA score I-II. Combined mean age = 69 (range 60-86) yrs. TG 1 = 46% ♀; TG 2 = 55%, CG = 58% female Cardiopulmonary, neurological or cognitive problems Start: immediately post-op TG 1: outpatient based progressive resistance training 3x/wk for 12 wks. TG 2: NMES to quadriceps starting post-op day 1 then at home 1-2x/day for 12 wks. CG 3: home based exercise program with weekly visits from PT for 12 wks. Baseline: pre-op Walking speed, timed stair climb, sit-to-stand test, quads strength (isokinetic), muscle cross sectional area  No adverse events reported No LT follow-up   69 Table 3.5 Summary of study characteristics (continued)     TG = treatment or experimental group; CG = control or comparator group; TUG = timed up & go test; 6MWT = 6 minute walk test; WOMAC = Western Ontario & McMaster Universities OA Index; AQoL = Assessment of Quality of Life instrument; LT = long term; DVT = deep vein thrombosis; VAS = visual analogue scale; MACTAR = McMaster Toronto Arthritis patient preference questionnaire; ROM = range of motion; ASA score = American Society of Anesthesiologists score; NMES = neuromuscular electrical stimulation.    Study/ language Design/ quality Setting/ country Participants (1° THA for OA) Exclusion criteria Intervention Comparison Outcomes Trudelle- Jackson 2004  (English) RCT, 2 treatment arms (n=34)  Moderate Home- based, Dallas, USA Adult of any age, race or gender,  1° THA 4 -12 months ago. Combined mean age = 60 (11) yrs, 54% ♀. Pain with weigh bearing, current low back pain, diagnosed vestibular problems, central or peripheral nervous system problems, dementia Start: 8-12 months post-op TG 1: Set of 7 progressive weight bearing exercises including functional strength, balance & trunk stability 3-4x/wk for 8 wks. TG 2: Set of 7 basic active ROM & isometric strength exercises commonly prescribed in acute phase of recover, 3-4x/wk for 8 wks. No other comparator Baseline: post-op Oxford Hip Score, hip strength, postural stability (portable force platform)  Adverse events reported No LT follow-up   70 3.4.4 Interventions No two trials evaluated similar post-acute physiotherapy interventions following primary THA with regards to timing, duration, dosage, setting and type of rehabilitation. The timing of the start of the intervention ranged from immediately after surgery to more than eight months following surgery. Most studies fell into the intermediate range delivering the intervention between two weeks and eight weeks post-operatively. Treatment duration ranged from two weeks to 15 weeks. Overall dosage was difficult to calculate in two trials and ranged from three times a week for three weeks (nine sessions in total) to three times a week for 12 weeks (36 sessions in total). All trials were classified as 'higher dosage' according to our categories established a priori. Treatment settings included inpatient rehabilitation facilities in two trials, outpatient departments/clinics in four, participants' homes in one, and not clear in one. Adherence to the physiotherapy protocol or intervention was rarely discussed and prospective adherence data were only collected in three trials. No studies included a subgroup analysis of high versus low adherers with regards to treatment effects. 3.4.5 Controls Control or comparator situations were not consistent among trials and included standard or routine physiotherapy in three studies, unsupervised home exercises in three, delayed treatment (wait list) in one, and a different type and intensity of exercise in one. No trials compared a physiotherapy intervention to a “no active intervention” or placebo control group. Baseline characteristics of the treatment and control groups were similar in all trials; however, only minimal baseline information was provided in most studies (Table 3.5). 71 3.4.6 Outcomes More than 23 different outcome tools were used in the eight trials again making comparison across studies difficult. Post-operative pain (ICF level: body function) was evaluated in three trials. Physical functioning (ICF level: activity) was assessed in seven studies and included performance measures (e.g. walking speed), self-report tools (e.g. WOMAC OA Index) and surgeon-rated tools (e.g. Harris Hip Score). One study used a self-report tool developed for rheumatoid arthritis (RA) and not validated in the OA population. Only one trial reported on HRQoL using the Assessment of Quality of Life (AQoL) measure. For the secondary outcomes of interest, data on hip muscle strength was most frequently reported; however, methods for assessing strength varied and included manual muscle testing using a 0 - 5 scale and dynamometry to evaluate isometric and isokinetic strength of a variety of lower extremity muscle groups. Additional secondary outcomes included gait parameters (e.g. step length, symmetry), passive and active ROM and self-rated exercise importance. We do not include any further discussion on secondary outcomes due to their inconsistent use and reporting methods. Baseline assessments occurred immediately prior to THA surgery in two studies and post- operatively, prior to the intervention in six. The timing of subsequent assessments was similar for treatment and control groups in all trials. Two trials included a long-term follow-up assessment at one year post-op. Adverse events during the intervention study period were documented in four studies:  Pulmonary embolus resulting in death in one treatment group participant (51)  Two DVTs and one hip dislocation in control group participants and one hip dislocation in a treatment group participant (41) 72  Pulmonary embolism in one participant (group not specified) (52)  Pain on non-surgical side in one standard care group participant and unspecified pain resulting in drop out in one functional exercise group participant (53) Authors of three of the trials reported the adverse events were not related to the treatment intervention while the other stated it was not clear. Three studies reported there were no adverse events while the remaining trial did not provide any information. Patient satisfaction with the physiotherapy intervention or outcomes was not reported in any trials. 3.4.7 Risk of bias in included studies All trials had design or methodological issues putting them at risk for bias. Allocation (selection) bias may have been present in five studies that did not describe adequate randomization with concealed allocation. Correspondence with the authors of two of these studies suggested appropriate randomization techniques may have been used; however, this information was not included in the published reports. Measurement bias is of concern in the five trials that did not report using a blinded outcome assessor. Performance bias, attributable to unblinded participants, may have been an issue in all but one trial. Attrition bias due to higher than anticipated dropouts was evident in one study and unclear in another. Another issue potentially contributing to performance bias is the lack of data on adherence to study protocols and exercise regimes. Only three trials reported monitoring participants' adherence to the intervention and/or comparator in a prospective fashion (e.g. attendances, exercise logs). This makes it difficult to determine the true effects of the intervention on the outcomes of interest or if there was a dose-response effect. That is, were there greater treatment effects in the subgroup of participants that adhered to the intervention compared to those who only partially completed the program? Similarly, the issue of therapist or provider adherence to the intervention protocol was not addressed in any of the trials. While there is always the need 73 for some individualization of therapeutic interventions to address the needs and preferences of patients, deviating from the study protocol complicates matters and introduces elements of co- intervention and even contamination that cannot be controlled for in the analysis if not monitored. Small sample sizes (less than the 50 participants per group standard) (38) in all but one trial may have contributed to underpowered studies and the risk of false conclusions (Type II error) about lack of treatment effect. Only one study reported performing a power calculation to determine an appropriate sample size a priori. A recent study reported that trial sample size (> 100 participants) more so than trial quality, predicted a positive study outcome (54). 3.4.8 Effects of interventions As noted earlier, we were unable to pool the data and perform a quantitative analysis due to trial heterogeneity. Therefore, we performed a qualitative analysis and best evidence synthesis for the primary outcomes using the GRADE approach (36). For trials with adequate data, WMDs and 95% confidence intervals for the primary outcomes are summarized below and sample forest plots are provided in Figures 3.2 to 3.7. Where appropriate, we present a comparison of effect sizes based on mean change versus those based on post-test values for select studies. 3.4.8.1 Pain Three studies reported on post-operative pain following different interventions and at two different time points. Galea et al. found no significant between group difference (WMD -16.80, [-5.13, 11.53]) after an 8-week supervised, centre-based physiotherapy program initiated eight weeks post-op compared to an unsupervised home program during the same time frame (55). This difference remained non-significant using mean change scores (Figures 3.2 and 3.3). Nyberg and colleagues reported no difference in pain at rest (WMD -3.20, [-11.78, 5.38]) or with 74 activity (WMD -0.50, [-10.23, 9.23]) after a 15-week group "THR school" when compared to home exercises alone at six months post-THA (41). These differences remained non-significant using mean change scores. We could not calculate an effect size for the third trial as the author did not respond to our request for missing data (52). There is consistent low quality evidence that:  Delayed outpatient physiotherapy of either eight or 15 weeks duration results in no differences in pain at rest or with activity when compared to home-based exercise programs. Figure 3.2 Forest plot for WOMAC pain comparing mean post-test values  Study or Subgroup Galea 2008 Total (95% CI) Heterogeneity: Not applicable Test for overall effect: Z = 1.16 (P = 0.25) Mean 39.5 SD 31.3 Total 11 11 Mean 56.3 SD 38.1 Total 12 12 Weight 100.0% 100.0% IV, Fixed, 95% CI -16.80 [-45.20, 11.60] -16.80 [-45.20, 11.60] Treatment (outpatient) Control (home-based) Mean Difference Mean Difference IV, Fixed, 95% CI -100 -50 0 50 100 Favours outpatient Favours home-based  Figure 3.3 Forest plot for WOMAC pain comparing mean change scores   3.4.8.2 Function Seven trials included one or more functional measures as outcomes. Function was assessed through self-report, surgeon-rated and performance-based tools. 75 Self-reported function Two studies reported on function using different self-report measures following different interventions. Galea et al. found no significant between-group difference (WMD -54.40, [- 168.08, 59.28]) in WOMAC function after eight weeks of supervised, centre-based physiotherapy initiated eight weeks post-op compared to an unsupervised home program (55). This difference remained non-significant using mean change scores. Trudelle-Jackson and colleagues reported that a late-phase (eight to 12 months post-op) functional exercise program resulted in a 5-point improvement in the Oxford Hip Score compared to a 2-point improvement with a standard home exercise program (53); however, only median values were provided and effect sizes could not be calculated. There is low quality evidence that:  Delayed outpatient physiotherapy results in no differences in self-reported function when compared to a home-based program.  Surgeon-rated function Two studies reported on function using surgeon-rated tools following different interventions. Hesse et al. found significant between-group differences for the Harris Hip Score at 10 days (WMD 13.60, [7.58, 19.62]), three months (WMD 8.90, [3.17, 14.63]) and 12 months (WMD 16.50, [11.55, 21.45]) after two weeks of body weight supported treadmill training as an adjunct to standard inpatient physiotherapy compared to standard physiotherapy alone (both initiated three weeks post-op) (51). Between-group differences at all measurement points remained significant when using change scores (Figures 3.4 and 3.5). Strom et al. reported no significant differences in the function subscale of the Merle d'Aubigne score at 12 weeks, six and 12 months when comparing an immediate, supervised progressive training program with unrestricted weight 76 bearing to an unsupervised, home exercise program with partial weight bearing instructions; however, missing data precluded calculation of effect sizes (52). There is moderate level evidence that:  Treadmill training with body weight support in addition to standard inpatient physiotherapy results in greater short- and long-term improvements in surgeon-rated function when compared to 'standard' care alone. Figure 3.4 Forest plot for Harris Hip Score comparing mean post-test values  Figure 3.5 Forest plot for Harris Hip Score comparing mean change scores  Performance-based function Six studies reported on function using six different performance measures (6-minute walk test (6MWT); walking speed over 10 to 15 metres; timed up-and-go (TUG); sit-to-stand test; timed stair climbing; and postural stability) and following varied interventions. 6MWT and walking speed Walking speed was assessed in five studies. Galea et al. reported no significant between group differences in 6-minute walking distance (WMD -30.50, [-1.09, 48.02) after an eight week 77 supervised physiotherapy program compared to unsupervised home exercises (55). This difference remained non-significant using change scores. There were no significant between group differences in self-selected walking speed as assessed in a gait lab at 10 days (WMD 0.10, [-0.02, 0.22]), three months (WMD 0.07, [-0.04, 0.18]) and 12 months (WMD 0.13, [-0.01, 0.27]) following two weeks of bodyweight supported treadmill training in addition to standard inpatient physiotherapy compared to standard physiotherapy alone (51). Using changes scores, a between group difference favouring the experimental group was found at 12 month follow-up only. Three weeks of mental gait training as an adjunct to standard out-patient physiotherapy resulted in no between group difference in walking speed (WMD 0.10, [-0.11, 0.31]) when compared to standard out-patient care alone (43). However, an improvement favouring the experimental group was significant (WMD 0.16, [0.02, 0.30]) when change scores were used (Figures 3.6 - 3.7). Figure 3.6 Forest plot for walking speed comparing mean post-test values  Figure 3.7 Forest plot for walking speed comparing mean change scores  78 Scherak et al. found significant between group differences in 15-metre walking speeds with four weeks of inpatient rehabilitation provided within two months of THA surgery compared to the same treatment started at three to four months post-op (WMD 2.30, [0.12, 4.48]) and at 11 months (WMD 3.70, [2.39, 5.01]) post-op (42). For this study, we calculated effect sizes using change scores as it was not appropriate to calculate the WMD using the pre and post-test data with the markedly different baseline values and time points. Significant between group differences in maximum walking speed were found after 12 weeks of daily home-based exercise program with electrical stimulation (WMD 0.40, [0.03, 0.77]) but not with out-patient resistance training (WMD 0.33, [-0.08, 0.74) when compared to 'standard' rehabilitation (56). Using mean change scores, also resulted in a non-significant difference with outpatient resistance training (WMD 0.23 [-0.01,0.47]) compared to 'standard' care. Interestingly, use of mean change scores to compare the electrical stimulation group and 'standard' care shows a non-significant between group difference (WMD 0.04, [-0.18, 0.26]) most likely due to the faster mean walking speed at baseline in the electrical stimulation group coupled with the very small sample size. Timed up and go test (TUG) The TUG test was used in one trial and significant between group differences (WMD 1.80, [0.15, 3.45]) were reported following an 8-week supervised, outpatient physiotherapy program provided eight weeks post-op compared to an unsupervised home program (55). This difference was not significant using mean change scores.  Sit-to-stand test Suetta et al. reported significant between group differences in the sit-to-stand test (5 repetitions) after 12 weeks of out-patient resistance training (WMD -4.50, [-7.09, -1.91]) and daily home- 79 based exercise program with electrical stimulation (WMD -3.00, [-5.97, -0.03]) compared to standard rehabilitation (56). Using mean change scores, the difference remained significant for the resistance training group (WMD 2.90, [0.93, 4.87]) but not the electrical stimulation group (WMD 2.10, [-0.22, 4.42] compared to 'standard' care. Timed stair climbing Timed stair climbing was assessed in two trials. Non-significant between group differences (WMD -0.20, [-0.17, 0.57]) in timed stair climbing were found following an 8-week supervised, outpatient physiotherapy program provided eight weeks post-op compared to an unsupervised home program (55). This difference remained non-significant using mean change scores. Timed to ascend 10 stairs did not differ after 12 weeks of daily home-based exercises with electrical stimulation (WMD -1.40, [-3.42, 0.62]) and outpatient resistance training (WMD -1.20, [-3.35, 0.95]) compared to 'standard' rehabilitation (56). This difference remained non-significant using mean change scores. Postural stability Postural stability improved after an eight week home-based functional exercise program delivered eight to 12 months post-op compared to a 'standard' exercise program; however, the difference was not statistically significant when effect sizes were calculated with post-test means (WMD 13.45, [-3.07, 29.97]). Using mean change scores, a significant difference was found (WMD 23.73, [4.47, 42.99]) (53). For performance-based function, there is conflicting, low to moderate level evidence that:  Daily home-based exercises plus electrical stimulation result in favourable outcomes for walking speed compared to 'standard' home exercises alone. 80  Early inpatient rehabilitation results in greater improvements in walking speed when compared to intermediate and delayed rehabilitation.  Outpatient physiotherapy started eight weeks post-op results in better TUG scores when compared to unsupervised home exercises.  Outpatient resistance training and home-based training with daily electrical stimulation result in better sit-to-stand scores than 'standard' home exercises.  Outpatient physiotherapy and home-based training with daily electrical stimulation result in no differences in timed stair climbing values when compared to 'standard' home exercises.  A late-phase home-based functional exercise program results in better postural stability compared to 'standard' exercises (based on mean change scores) Results mostly favoring the intervention groups differ in several trials when effect sizes are calculated using mean change scores compared to mean post-test values. 3.4.8.3 Health-related quality of life One study reported on HRQoL after an 8-week outpatient physiotherapy program and found no significant between group difference (WMD 0.05, [-0.13, 0.23]) when compared to unsupervised home exercises (55). The lack of between-group difference was also found using mean change scores. There is low quality evidence that:  Delayed, outpatient physiotherapy results in no short-term differences in HRQoL compared to an unsupervised home-based program. 81 3.5 Discussion 3.5.1 Summary of main results This review synthesized the research published between 1990 and August 2008 that examined the effects of post-acute physiotherapy after primary THA on pain, function and HRQoL. There was limited evidence of very low to moderate quality to support the use of various forms of post- acute physiotherapy for self-reported, surgeon-rated and performance-based function. We found favourable results for physical function with: body weight supported treadmill training as an adjunct to inpatient physiotherapy; a progressive, functional home exercise program initiated eight months post-op; outpatient resistance training and daily electrical muscle stimulation as an adjunct to 'standard' home exercises; inpatient rehabilitation initiated within two months of surgery; and outpatient physiotherapy initiated eight weeks post-surgery. We determined there were no significant between-group differences for pain and HRQoL based on mean post-test values. However, in several cases, differences were evident when mean change scores were used. Due to the heterogeneity among studies, it was not possible to perform a meta-analysis and draw conclusions across studies as to whether one form of physiotherapy was superior to another following THA surgery.  The limited between-group differences are largely explained by the fact that no trials compared a physiotherapy intervention to a „no treatment‟ or placebo control group. Clinically important within-group benefits were evident in most groups that received the targeted or experimental intervention, but one cannot draw conclusions about the magnitude of the added benefit of physiotherapy compared to the natural course of recovery after THA.  This lack of accumulated evidence is not evidence of a lack of effectiveness, but speaks to the need for high quality trials with carefully specified interventions and uniform outcome measures. 82 3.5.2 Quality of the evidence The overall quality of trials included in this review was very low to moderate with many trials being downgraded due to lack of concealed allocation and unblinded assessors (36). As noted earlier, it is almost impossible to blind patients and providers due to the nature of the interventions. It is troubling however, to see this number of studies with unblinded outcome assessors. A majority of trials were underpowered to detect meaningful between group differences. Reporting issues were also evident in the included trials with few studies identifying their primary outcome and only one study reporting they performed a power calculation using this outcome to determine necessary sample size. Participant demographics, particularly presence and severity of co-morbidities, were poorly described in most trials. A growing body of research demonstrates that co-existing health problems are prevalent in individuals undergoing THA (57, 58) and influence the post-operative course of recovery and THA outcomes (58, 59). Therefore, co-morbidities should be reported and considered in the analyses of outcomes after THA rehabilitation. 3.5.3 Strengths of this review To our knowledge, this is the most comprehensive systematic review to date examining the varied forms of physiotherapy commonly delivered during the first year after THA surgery. Our extensive search strategy resulted in a large body of published and unpublished studies and our inclusion of non-English trials reduced the risk of language bias. Five different countries and respective health care systems are represented in our review. However, it is possible that some relevant trials are missing and those published since August 2008 are not included. There remain large gaps in the evidence and a paucity of high quality trials related to physiotherapy after THA. 83 3.5.4 Weaknesses and potential biases of this review Due to trial heterogeneity, we were unable to perform the planned meta-analysis and make definitive statements about the strength of the evidence and magnitude of treatment effects for physiotherapy interventions after THA. Limiting our review to RCTs and CCTs precluded other study designs that may have contributed additional evidence on the topic of post-acute physiotherapy after THA. We made every effort to avoid various forms of bias by pursuing unpublished reports in our search strategy, including non-English studies, having two authors independently screen and review papers, and performing a rigorous and standardized quality assessment process. We did not create funnel plots to examine the possibility of publication bias due to the small number of included trials. 3.5.5 Agreements and disagreements with other studies or reviews A recently published Cochrane review compared the effect of multidisciplinary rehabilitation interventions to single discipline care following primary and revision THA and TKA for OA and RA (29). This review, however, focused on the acute care phase of rehabilitation and included use of clinical pathways, formal patient education programs, post-operative pain management and dietary interventions in its range of outcomes. Studies examining physiotherapy interventions alone were excluded. Due to clinical heterogeneity and low study quality, a meta- analysis was not performed. The authors concluded that there was silver level (low to moderate) evidence to support early multidisciplinary rehabilitation in improving activity and participation outcomes; however, optimal intensity, frequency and effects of rehabilitation warranted further study. In order to support their development of evidence-based guidelines for ambulatory physiotherapy following THA, the French Society of Physical and Rehabilitation Medicine (SOFMER) conducted a systematic review of French and English language articles from an unspecified time 84 frame (60). A meta-analysis was not possible due to trial heterogeneity. Based on their analysis of 16 studies of mixed methodologies and inclusion criteria, the authors concluded that there are some advantages of ambulatory physiotherapy on muscle strength and function. From their limited economic analysis, no differences were found comparing home therapy to other approaches. In 2005, two separate reviews were conducted to inform rehabilitation practice and education in Ontario, Canada. In a review by the Medical Advisory Secretariat of the Ministry of Health and Long-Term Care, it was concluded that there was limited high-quality evidence (from one published abstract) to support the use of home-based physiotherapy instead of inpatient treatment after THA and TKA (61). The English-only literature was reviewed up to 2004 to inform the Greater Toronto Area Rehabilitation Network's development of an online discussion forum to promote collaboration and education among patients and health professionals (62). Based on this review of experimental and observational studies spanning pre-operative, acute and post- operative care, rehabilitation was found to be an effective component in TJA management at various stages of recovery. Best practice recommendations were not clearly differentiated for hip versus knee replacement. Soever and MacKay also concluded that comparison of outpatient care to other rehabilitation processes warrants further study to determine the optimal processes of care following TJA (62). In the month prior to completion of our review, a systematic review on the effects of therapeutic exercise after discharge from hospital for primary THA surgery was published (31). This review of trials up to April 2007 included interventions delivered beyond one year post-op and excluded those that involved inpatient rehabilitation and electrical modalities as a treatment adjunct. While there is obvious overlap in the trials included in this and our review, our review is more comprehensive in that it includes inpatient treatment, which is routine rehabilitation practice in 85 many countries (14, 15, 63). The authors also determined a full meta-analysis was not appropriate; however, calculated pooled effect sizes for walking speed and hip abductor strength despite marked clinical heterogeneity among the four trials. Regardless of these differences, Minns Lowe et al. similarly concluded that trial availability, quality and diversity prevented definitive answers regarding effectiveness of post-discharge physiotherapy exercise for primary THA (31). 3.6 Conclusions 3.6.1 Implications for practice There is insufficient evidence from the studies included in this review, to suggest that any one type, setting, timing or amount of post-acute physiotherapy for THA is superior to another. The tremendous variety in the timing, duration, intensity and content of the interventions make developing clinical recommendations problematic. Additional high quality trials with sufficient power and standardized outcome assessment methods are needed to establish specific treatment recommendations for post-acute PT after THA. With the longstanding impairments, functional limitations and participation restrictions present in individuals awaiting primary THA (57, 59, 64-66) and still evident as much as a year post-op (10, 25, 26), it may be inappropriate and unethical to withhold post-operative physiotherapy all together. The physiotherapy setting may be less important than the type, quality, dosage and adherence to the intervention; however, dose-response calculations were not possible and limited adherence data were available in the trials selected, and further evidence is required to support this conclusion. Despite the limitations in the availability and quality of evidence, there are a number of implications for clinical practice: 86  There is low to moderate evidence to support the use of inpatient body weight supported treadmill training, outpatient resistance exercises, and home-based exercise with and without electrical muscle stimulation for improving surgeon-rated and performance-based function after THA;  There is limited, moderate evidence to support the use of a late-phase, functional and progressive exercise approach for improving balance and self-reported function;  There is limited, low quality evidence to suggest that earlier participation in inpatient rehabilitation results in greater walking speed than with delayed rehabilitation;  There is limited, low quality evidence to support the use of group exercise programs following THA surgery and conflicting evidence that PT supervised exercise programs result in better outcomes than unsupervised programs. 3.6.2 Implications for research The heterogeneity among trials regarding the timing and types of interventions, control situations, outcomes and measures highlights the need for greater standardization in the conducting, evaluating and reporting of future research trials in THA rehabilitation. Additional suggestions for future research in this field include:  Conduct trials with stronger study designs including concealed group allocation, blinded outcome assessors and larger sample sizes  Prospectively monitor participants' adherence to the study protocol (intervention) with subgroup analysis of treatment effects for high and low adherers (dose-response)  Prospectively monitor treatment providers' adherence to the study protocol  Provide more detailed descriptions of the therapeutic interventions including timing, duration, frequency, intensity and progressions 87  Provide more detailed descriptions of 'standard' or 'usual' care to ensure appropriate comparisons are made  Determine what baseline and post-operative contextual factors (personal and environmental) are most strongly related to rehabilitation outcomes and patient satisfaction and perform appropriate subgroup analyses  Assess how patient preference for type, timing and/or setting for rehabilitation affects outcomes  Identify a core set of outcome measures for each of the ICF domains to allow for consistent, standardized reporting of outcomes and comparison among interventions  Include measures that capture participation, HRQoL and patient satisfaction  Include a follow-up phase of at least 12 months  Monitor and report on physical activity levels and use of other medical or rehabilitative therapies during the follow-up period  Use multiple research sites to increase sample sizes and generalizability of findings  Include an economic analysis including both direct and indirect costs, and patient out-of- pocket expenses of varied physiotherapy interventions and follow-up services 88 3.7 References 1. Canadian Institute for Health Information (CIHI). Hip and total knee replacements in Canada - Canadian Joint Replacement Registry (CJRR). 2008-2009 Annual Report. [cited 2009 Sep 17]. Available from: URL: www.cihi.ca/cjrr. 2. American Academy of Orthopedic Surgeons. Facts on hip replacements and facts on knee replacements. [cited 2009 Dec 9]. Available from: URL: http://www.aaos.org/research/stats/patientstats.asp. 3. March LM, Bagga H. Epidemiology of osteoarthritis in Australia. Med J Aust 2004;180(5 Suppl):S6-S10. 4. Badley EM, Crotty M. An international comparison of the estimated effect of the aging of the population on the major cause of disablement in musculoskeletal disorders. J Rheumatol 1995;22(10):1934-40. 5. Kim S. Changes in surgical loads and economic burden of hip and knee replacements in the US: 1997 - 2004. Arthritis Rheum 2008;59(4):481-8. 6. Jones D, Westby M, Greidanus N, Johanson NA, Krebs DE, Robbins L, et al. Update on hip and knee arthroplasties: Current state of evidence. Arthritis Rheum 2005;53(5):772-80. 7. Tian W, DeJong G, Brown M, Hsieh CH, Zamfirov ZP, Horn SD. Looking upstream: Factors shaping the demand for postacute joint replacement rehabilitation. Arch Phys Med Rehabil 2009;90(8):1260-8. 8. Karrholm J, Garellick G, Rogmark C, Herberts P. Swedish Hip Arthroplasty Register Annual Report 2007. [cited 2009 May 25]. Available from: URL: http://www.jru.orthop.gu.se. 9. National Joint Registry for England and Wales. 6th Annual Report. 2009 [cited 2009 Dec 8]. Available from: URL: http://www.njrcentre.org.uk. 10. Brander VA, Stulberg SD. Rehabilitation after hip- and knee-joint replacement: An experience- and evidence-based approach to care. Am J Phys Med Rehabil 2006;85(Suppl):S98- S118. 11. Ethgen O, Bruyere O, Richy F, Dardennes C, Reginster JY. Health-related quality of life in total hip and total knee arthroplasty. A qualitative and systematic review of the literature. J Bone Joint Surg Am 2004;86-A(5):963-74. 12. Salmon P, Hall GM, Peerbhoy D, Shenkin A, Parker C. Recovery from hip and knee arthroplasty: Patients' perspective on pain, function, quality of life, and well-being up to 6 months postoperatively. Arch Phys Med Rehabil 2001;82(3):360-6. 13. Johanson NA. The hospital course in total hip and knee arthroplasty: Which length of stay is right? J Rheumatol 1997;24(1):1-2. 89 14. Müller E, Blatt K, Gülich M, Jäckel WH. Treatment schemes and guidelines for rehabilitation after total hip replacement (THR): Results of an international survey and a multicenter study in Germany. Int J Rehabil Res 2007;30(1):86-7. 15. Roos EM. Effectiveness and practice variation of rehabilitation after joint replacement. Curr Opin Rheumatol 2003;15(2):160-2. 16. Troughton D, Pearce J. Total hip and knee replacement survey: Rehabilitation practices through the continuum of care. Rehabilitation Services, Vancouver Island Health Authority; 2007. 17. Ganz SB, Viellion G. Pre and post surgical management of the hip and knee. In: Wegener ST, Belza BL, Gall EP, editors. Clinical Care in the Rheumatic Diseases. Atlanta: American College of Rheumatology; 1996. p. 103-6. 18. Kampa K. Orthopedic assessment and treatment of the geriatric patient. Lewis CB and Knortz KA, editors. St. Louis (MO): CV Mosby Co.; 1993. 19. Bitar AA, Kaplan RJ, Stitik TP, Shih VC, Vo AN, Kamen LB. Rehabilitation of orthopedic and rheumatologic disorders. 3. Total hip arthroplasty rehabilitation. Arch Phys Med Rehabil 2005;86(3 Suppl 1):S56-60. 20. Enloe LJ, Shields RK, Smith K, Leo K, Miller B. Total hip and knee replacement treatment programs: A report using consensus. J Orthop Sports Phys Ther 1996;23(1):3-11. 21. Young Tucker C, Diamond A. Total hip arthroplasty. In: Cioppa-Mosca J, Cahill JB, Cavanaugh JT, Corradi-Scalise D, Rudnick H, Wolff AL, editors. Postsurgical Rehabilitation Guidelines for the Orthopedic Clinician. St. Louis: Hospital for Special Surgery, Mosby Elsevier; 2006. p. 3-16. 22. Mont MA, Seyler TM. Advanced techniques for rehabilitation after total hip and knee arthroplasty. Clin Orthop Relat Res 2009 [cited 2009 Aug 13];467(6). Available from: URL: http://www.springerlink.com/content/5j7t7h3510r203t5/. 23. Givens-Heiss DL, Krebs DE, Riley PO, Strickland EM, Fares M, Hodge WA, et al. In vivo acetabular contact pressures during rehabilitation, part II: Postacute phase. Phys Ther 1992;72(10):700-10. 24. New Zealand Orthopaedic Association. The New Zealand Joint Registry, Nine year report. [cited 2009 Jan 15]. Available from: URL: http://www.cdhb.govt.nz/NJR/reports/A2D65CA3.pdf. 25. Trudelle-Jackson E, Emerson R, Smith S. Outcomes of total hip arthroplasty: A study of patients one year postsurgery. J Orthop Sports Phys Ther 2002;32(6):260-7. 26. Shih CH, Du YK, Lin YH, Wu CC. Muscular recovery around the hip joint after total hip arthroplasty. Clin Orthop Relat Res 1994;302:115-20. 90 27. McMeeken JM, Galea MP. Impairment of muscle performance before and following total hip replacement. Int J Ther Rehabil 2007;14(2):55-62. 28. National Institutes of Health. Total hip replacement. NIH Consens State Sci Statement 1994 Sep 12-14 [cited 2004 Apr 20];12(5):1-31. Available from: URL: http://consensus.nih.gov/1994/1994HipReplacement098html.htm. 29. Khan F, Ng L, Gonzalez S, Hale T, Turner-Stokes L. Multidisciplinary rehabilitation programmes following joint replacement at the hip and knee in chronic arthropathy. Cochrane Database Syst Rev 2008(2):CD004957. 30. McDonald S, Hetrick S, Green S. Pre-operative education for hip or knee replacement. Cochrane Database Syst Rev 2004(1):CD003526. 31. Minns Lowe CJ, Barker KL, Dewey ME, Sackley CM. Effectiveness of physiotherapy exercise following hip arthroplasty for osteoarthritis: A systematic review of clinical trials. BMC Musculoskelet Disord 2009;10:98. 32. Westby MD, Backman CL. Patient and health professional views on rehabilitation practices and outcomes following total hip and knee arthroplasty: A focus group study. Submitted 2009 Sep 3. 33. World Health Organization. International classification of functioning, disability and health (ICF). 2001 [cited 2004 Mar 3]. Available from: URL: http://www.who.int/classifications/icf/. 34. van Tulder M, Furlan A, Bombardier C, Bouter L. Updated method guidelines for systematic reviews in the Cochrane Collaboration Back Review Group. Spine 2003;28(12):1290-9. 35. Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions version 5.0.2 [updated 2009 Sep]. The Cochrane Collaboration [cited 2009 Dec 15]. Available from: URL: http://www.cochrane-handbook.org. 36. Schunemann HJ, Oxman AD, Vist GE, Higgins JPT, Deeks JJ, Glasziou P, et al. Interpreting results and drawing conclusions. In: Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions 5.0.2. The Cochrane Collaboration; 2009. 37. van den Ende CHM, Steultjens EMJ, Bouter LM, Dekker J. Clinical heterogeneity was a common problem in Cochrane reviews of physiotherapy and occupational therapy. J Clin Epidemiol 2006;59(9):914-9. 38. Tugwell P, Shea B, Boers M, Simons L, Strand V, Wells G. Evidence-based Rheumatology. London (ENG): BMJ Books; 2004. 39. Higgins JPT, Deeks JJ, Altman DG. Special topics in statistics. In: Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions 5.0.2. The Cochrane Collaboration; 2009. 40. Ray JW, Shadish WR. How interchangeable are different estimators of effect size? J Consult Clin Psychol 1996;64(6):1316-25. 91 41. Nyberg B, Kreuter M. A comparison between group physical therapy and individual training at home in patients with total hip replacement. Nordisk Fysioterapi 2002;6:82-8. 42. Scherak O, Kolarz G, Wottawa A, Maager M, El Shohoumi M. Comparison between early and late inpatient rehabilitation measures after implantation of total hip endoprostheses. Rehabilitation 1998;37(3):123-7. 43. Mayer J, Bohn J, Görlich P, Eberspächer H. Mental gait training - effectiveness of a therapy method in the rehabilitation after hip-replacement. Z Orthop Ihre Grenzgeb 2005;143(4):419-23. 44. Gilbey HJ, Ackland TR, Wang AW, Morton AR, Trouchet T, Tapper J. Exercise improves early functional recovery after total hip arthroplasty. Clin Orthop 2003(408):193-200. 45. Jan MH, Hung JY, Lin JC, Wang SF, Liu TK, Tang PF. Effects of a home program on strength, walking speed, and function after total hip replacement. Arch Phys Med Rehabil 2004;85(12):1943-51. 46. Mahomed NN, Davis AM, Hawker G, Badley E, Davey JR, Syed KA, et al. Inpatient compared with home-based rehabilitation following primary unilateral total hip or knee replacement: A randomized controlled trial. J Bone Joint Surg 2008;90(8):1673-80. 47. Weaver FM, Hughes SL, Almagor O, Wixson R, Manheim L, Fulton B, et al. Comparison of two home care protocols for total joint replacement. J Am Geriatr Soc 2003;51(4):523-8. 48. Maire J, Dugue B, Faillenet-Maire A-, Tordi N, Parratte B, Smolander J, et al. Recovery after total hip joint arthroplasty in elderly patients with osteoarthritis: Positive effect of upper limb interval-training. J Rehabil Med 2003;35(4):174-9. 49. Sashika H, Matsuba Y, Watanabe Y. Home program of physical therapy: Effect on disabilities of patients with total hip arthroplasty. Arch Phys Med Rehabil 1996;77(3):273-7. 50. Siggeirsdottir K, Olafsson O, Jonsson H, Iwarsson S, Gudnason V, Jonsson BY. Short hospital stay augmented with education and home-based rehabilitation improves function and quality of life after hip replacement: Randomized study of 50 patients with 6 months of follow- up. Acta Orthop 2005;76(4):555-62. 51. Hesse S, Werner C, Seibel H, von Frankenberg S, Kappel EM, Kirker S, et al. Treadmill training with partial body-weight support after total hip arthroplasty: A randomized controlled trial. Arch Phys Med Rehabil 2003;84(12):1767-73. 52. Ström H, Huss K, Larsson S. Unrestricted weight bearing and intensive physiotherapy after uncemented total hip arthroplasty. Scand J Surg 2006;95(1):55-60. 53. Trudelle-Jackson E, Smith SS. Effects of a late-phase exercise program after total hip arthroplasty: A randomized controlled trial. Arch Phys Med Rehabil 2004;85(7):1056-62. 54. Singh JA, Murphy S, Bhandari M. Trial sample size, but not trial quality, is associated with positive study outcome. J Clin Epidemiol 2010;63(2):154-62. 92 55. Galea MP, Levinger P, Lythgo N, Cimoli C, Weller R, Tully E, et al. A targeted home- and center-based exercise program for people after total hip replacement: A randomized clinical trial. Arch Phys Med Rehabil 2008;89(8):1442-7. 56. Suetta C, Magnusson SP, Rosted A, Aagaard P, Jakobsen AK, Larsen LH, et al. Resistance training in the early postoperative phase reduces hospitalization and leads to muscle hypertrophy in elderly hip surgery patients - A controlled, randomized study. J Am Geriatr Soc 2004;52(12):2016-22. 57. Philbin EF, Groff GD, Ries MD, Miller TE. Cardiovascular fitness and health in patients with end-stage osteoarthritis. Arthritis Rheum 1995;38(6):799-805. 58. Marks R, Allegrante JP. Comorbid disease profiles of adults with end-stage hip osteoarthritis. Med Sci Monit 2002;8(4):CR305-9. 59. Tuominen U, Blom M, Hirvonen J, Seitsalo S, Lehto M, Paavolainen P, et al. The effect of co-morbidities on health-related quality of life in patients placed on the waiting list for total joint replacement. Health Qual Life Outcomes 2007;15:5-16. 60. Genêt F, Gouin F, Coudeyre E, Revel M, Rannou F. The benefits of ambulatory physiotherapy after total hip replacement. Clinical practice recommendations. Ann Readapt Med Phys 2007;50(9):776-82. 61. Medical Advisory Secretariat, Ontario Ministry of Health and Long-Term Care. Physiotherapy rehabilitation after total knee or hip replacement. Health technology literature review. 2005 May [cited 2005 Dec 6]. Available from: URL: http://www.health.gov.on.ca/english/providers/program/mas/tech/ohtas/tech_rehabtkr_061705.ht ml. 62. Soever L, MacKay C. Best practices across the continuum of care for total joint replacement. Greater Toronto Area Rehabilitation Network; 2005 Jul. [cited 2005 Dec 6]. Available from: URL: http://www.gtarehabnetwork.ca/downloads/report-totaljointreplacement-2005.pdf. 63. Jaglal SB, MacKay C, Corrigan L. Rehabilitation for total joint replacement. In: Badley E, Glazier R, editors. Arthritis and related conditions in Ontario. ICES research atlas. 2nd ed. Institute for Clinical Evaluative Sciences; 2004. p. 133,- 146. [cited 2008 Oct 1]. Available from: URL: http://www.ices.on.ca/file/ACF13B.pdf. 64. Mahon JL, Bourne RB, Rorabeck CH, Feeny DH, Stitt L, Webster-Bogaert S. Health-related quality of life and mobility of patients awaiting elective total hip arthroplasty: A prospective study. Can Med Assoc J 2002;167(10):1115-21. 65. Salaffi F, Carotti M, Stancati A, Grassi W. Health-related quality of life in older adults with symptomatic hip and knee osteoarthritis: A comparison with matched healthy controls. Aging Clin Exp Res 2005;17(4):255-63. 66. van Dijk GM, Veenhof C, Schellevis F, Hulsmans H, Bakker JPJ, Arwert H, et al. Comorbidity, limitations in activities and pain in patients with osteoarthritis of the hip or knee. BMC Musculoskelet Disord 2008;9:95. 93 Chapter 4: Post-acute physiotherapy following primary total knee replacement for osteoarthritis: A Cochrane systematic review4 4.1 Background Total knee arthroplasty (TKA) surgeries are highly successful, commonly performed orthopaedic procedures for individuals with advanced knee osteoarthritis (OA). Also termed total knee replacement (TKR), more than 37,900 of these primary procedures were performed in Canada in 2006/07 (1). This number is far greater in the United States (US) with more than 0.6 million TKA surgeries reported in 2007(2). Total knee arthroplasty procedures out number total hip arthroplasty (THA) in most countries (1, 3-6) or are expected to surpass THA numbers within the next few years (7). The one-year increase in number of TKAs was 12.5% in Canada (1) and was similar to that reported in other international joint registries (8, 9). The number of TKA surgeries is projected to grow with the aging population, rising obesity rates and increased prevalence of OA (10-13). Other factors contributing to the increase are patient demand and higher expectations for improved quality of life, shorter acute care length of stay (LOS), technical advances, and safer anaesthetic techniques that have expanded the indications for joint replacement surgery (9, 11, 13, 14). In the US, an increase in number of orthopaedic surgeons with a focus on total joint arthroplasty (TJA) has further fueled the growth in these procedures (13). International data indicate that OA is the primary reason for TKA surgery (4-6, 15) and in Canada accounts for 93% of all TKAs (1). Total knee arthroplasty surgery decreases pain, restores mobility and function, improves health- related quality of life (HRQoL) and is associated with a high patient satisfaction rate (16-18).  4  A version of this chapter will be submitted for publication. Westby MD, Kennedy, Jones D, Jones A, Doyle- Waters M, Backman C. Post-acute physiotherapy following primary total knee arthroplasty for osteoarthritis: A Cochrane systematic review. 94 While a cost effective surgical procedure for advanced knee OA, the rapid increase in TKA surgeries places significant burdens on healthcare budgets (9, 11, 19). Based on 2007 data from the Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Sample in the US, mean hospital charges per TKA procedure were more than $42,000 suggesting annual total hospitalization charges in excess of $25 billion US (2). Nationwide costs of TKA surgeries in Canada are not available; however, recent data suggests costs associated with TKA surgery and the period up to six months post-op average more than $14,700 per patient leading to an informal estimate of current Canadian expenditures in excess of $557 million dollars annually (1, 20). In the past decade in Canada, hospitalization following TKA has decreased by 50% (1). Data from 2006/07 show that the mean and median acute care length of stay (LOS) to be six and four days respectively for primary TKA and revision surgeries (1). Data suggest a shorter LOS when only primary (2, 8, 9) and non-complex procedures are considered (21). Earlier discharge to home or community settings decreases the time available for physical recovery, acute care rehabilitation, and patient and family education and counseling. A reduced LOS places additional burden and responsibility on the patients, their family and their post-acute health care providers to monitor for and address post-operative complications such as wound infection and deep vein thrombosis (DVT). As well, the short acute care stay puts more emphasis on preadmission education and the role and timing of post-acute physiotherapy interventions. In an editorial examining LOS after THA and TKA, Johanson comments that "it remains unclear just how far this process [of reducing LOS] can be taken without either compromising quality of care or simply shifting costs to a less regulated outpatient environment" (22)(p.1). Currently, no evidence-based practice guidelines exist to inform best practice for post-acute rehabilitation following TKA in North America and specific rehabilitation practices vary greatly among providers (23-28). Current therapeutic interventions are based largely on clinical 95 experience (29), clinician and surgeon preferences (23, 30), facility-based protocols (31), the acute care phase of recovery (32, 33), health insurance funding schemes (13, 34) and outdated approaches (35). Furthermore, little or no guidance exists on recommended activities following discharge from the acute care setting (36). All but two international joint replacement registries (5, 14) monitor purely surgical end-points, such as peri-operative complications and number of revision procedures. Rehabilitation and its contribution to both surgical and patient-centred outcomes are an overlooked aspect of the total episode of care for TKA patients. There are a number of reports to suggest that some patients are not reaching their physical potential following TKA and that post-operative pain, physical impairments and functional limitations are still issues as long as two years after surgery (37-45). Persistent psychological distress and reduced HRQoL below population norms have also been reported 12 months post- surgery (46). Current exercise programs performed during the early phase of rehabilitation are insufficient to restore muscle strength, normal gait and more complex functional activities lost during the pre-operative period (40, 43). Additionally, inadequate quadriceps strength and reduced physical capacity coupled with further reduction of reserve capacity seen with normal aging and progressive OA may lead to an increased fall risk and declining independence in daily living for older adults (44, 47, 48). However, we do not know exactly what kind of physiotherapy performed over what period of time is effective or necessary to enhance short- and long-term surgical outcomes and optimize patient activity, participation, and HRQoL. The National Institutes of Health (NIH) conference on TKA concluded "the use of rehabilitation services is perhaps the most understudied aspect of the peri-operative management of TKA patients"(34)(p. 6). As part of the larger project to develop evidence-based clinical practice guidelines for post-acute rehabilitation after TKA and THA, systematic reviews were undertaken to examine the current 96 literature on various rehabilitation approaches and interventions in this patient population. Previous reviews have addressed continuous passive motion (CPM) following TKA (49), multidisciplinary rehabilitation for primary and revision THA and TKA (50), and therapeutic exercise after discharge from hospital for TKA (51). A Cochrane review currently underway is investigating the effects of surface neuromuscular stimulation to the quadriceps pre- and post- TKA (52). 4.2 Purpose The primary aim of this review is to assess the effects of post-acute physiotherapy following primary TKA for OA on patient-centred outcomes of pain, physical function and HRQoL. 4.2.1 Objectives The specific objectives are to: 1. Document the short-term and long-term effects of post-acute physiotherapy on pain, physical function and HRQoL in persons undergoing primary TKA for OA. 2. Compare outcomes across different clinical settings and treatment approaches, timing and dosages. 3. Make recommendations for clinical practice based on the strength of the evidence. 4. Identify research needs post-acute physiotherapy in the TKA population. 4.3 Methods 4.3.1 Criteria for considering studies for this review We considered both randomized controlled trials (RCTs) and controlled clinical trials (quasi- randomized and controlled before-after designs) for inclusion in this review. Designs not 97 included were cohort, case-control, single case studies, single subject, case series and pre/post studies with no control group. 4.3.2 Types of participants Primary trials of individuals aged 19 years and older who underwent an elective, primary TKA for primary (idiopathic) or secondary OA due to chronic trauma, developmental or congenital problems were considered. Surgical procedures included all forms of fixation (cemented, hybrid or cementless), surgical approaches (medial, lateral, parapatellar, minimally invasive) and types of prostheses (constrained, semi-constrained, mobile platform). Studies were excluded if participants had:  undergone a revision, uni-compartmental or bilateral procedure;  an inflammatory arthritis diagnosis (e.g. rheumatoid arthritis);  an acute fracture;  a tumour;  significant pre- or peri-operative complications (e.g., nerve palsy, deep wound infection, severe anaemia, DVT)  an extended acute hospital stay beyond the standard LOS for that hospital facility. Studies with mixed pre-surgical diagnoses were included if 90% or more of the participants had an OA diagnosis. 4.3.3 Types of interventions 4.3.3.1 Experimental Post-acute physiotherapy includes physiotherapy provided in any setting (in-patient, outpatient or home/community). The 'post-acute' phase was determined to extend from immediately following discharge from the acute hospital ward or setting and up to 12 months post-surgery. The time 98 frame selected reflects patients' views on the duration of rehabilitation services after TKA surgery (53). Physiotherapy includes 1 to1 treatments and group interventions supervised by the physiotherapist (PT) and/or a trained physiotherapy assistant under the direct supervision of a PT. Interventions could include alone or in combination:  therapeutic exercise;  hydrotherapy;  use of thermal or electrical modalities (i.e. cryotherapy, heat, transcutaneous electrical nerve stimulation and neuromuscular electrical stimulation (NMES));  postural, proprioceptive and balance training;  gait training;  functional exercises;  cardiovascular training;  manual therapy techniques (i.e. soft tissue massage, passive range of motion);  patient education and self-management techniques. Interventions not included in this review were CPM, splinting or bracing, and chiropractic or osteopathic manipulation. We characterized interventions by:  timing of the onset of treatment (immediate = < 2 weeks post-surgery, intermediate = between 2 weeks and 3 months post-surgery or delayed = > 3 months post-surgery)  setting (outpatient, inpatient or home-based)  dosage (lower = < 6 visits or 3 weeks of intervention and higher = ≥ 6 visits or 3 weeks of intervention) 99 4.3.3.2 Control Supervised physiotherapy was compared to control situations in which participants received 'routine care' (standard or traditional physiotherapy of a given institution or surgeon), „attention only‟ (no active intervention), unsupervised or self-directed home exercises, or other operationalized comparators (which may include comparing one physiotherapy program or therapeutic setting to another). 4.3.4 Types of outcome measures The primary outcomes of interest were:  Post-operative pain (at rest or with activity)  Physical function (as measured by validated functional assessment self-report tools, surgeon-rated tools or performance-based measures)  Health-related quality of life (as measured by validated generic, condition-specific, individualized or preference-based instruments)  Adverse events due to rehabilitation (e.g. increased pain, fall)  Secondary outcomes were categorized according to the International Classification of Functioning (ICF) (54) but were not included in the analysis:  Body structure and function (e.g. knee ROM, lower extremity muscle strength and endurance, soft tissue flexibility, proprioception, cardiovascular fitness)  Activity (i.e. those not captured under physical function outcomes)  Participation (e.g. paid/unpaid work, care giving, leisure activities, social activities)  Personal contextual factors (e.g. age, gender, body mass index, patient satisfaction, self- efficacy, anxiety, motivation)  Environmental factors (e.g. health insurance coverage, waiting time for surgery) 100 Process measures:  Timing of treatment intervention  Timing of baseline assessment (pre-surgery or post-surgery/pre-intervention)  Treatment setting (inpatient, outpatient, home)  Treatment dosage (e.g. frequency, duration, number of sessions)  Adherence to treatment 4.3.5 Search methods for identification of studies We searched the following electronic databases to identify relevant studies published between January 1990 and August 31, 2008. Alerts were set up in several databases to continue to identify potentially relevant articles up until May 31, 2009. The period of time covered was recommended by the expert panel and covers the period of time when inpatient LOS fell and diagnostic related groups (DRGs) became widespread basis for reimbursement in the US by this year. Articles in any language were reviewed.  Cochrane Musculoskeletal Group Trials Register (September 17, 2008)  Cochrane Central Register of Controlled Trials (CENTRAL)(Issue 3, 2008)  MEDLINE (OVID) (1950 to August 31, 2008)  CINAHL (EBSCO) (1981 to August 31, 2008)  EMBASE (OVID) (1980 to 2008 Week 39)  Cochrane Database of Systematic Reviews (Issue 3, 2008), ACP Journal Club and DARE (OVID)  PEDro (Physiotherapy Evidence Database) (August 31, 2008)  Web of Science (August 31, 2008)  SPORTDiscus (EBSCO) (August 31, 2008) 101 See Appendix E for an example of the search strategy for MEDLINE using a combination of MeSH and key words. The quality filter recommended by the Cochrane Musculoskeletal Review Group was applied. We scanned articles, review papers and textbooks for additional papers. Known experts in the field and authors of existing, high quality studies were contacted for further references. A grey literature search was conducted to identify published and unpublished papers, reports and other documents. We searched papers and proceedings from congresses and symposiums through the databases PapersFirst and ProceedingsFirst. Additionally, we hand searched conference proceedings of key professional organizations, highly relevant orthopaedic and rehabilitation journals, professional organizations' websites, and national joint arthroplasty registries. 4.3.6 Data collection and analysis A team of four primary reviewers (MW, DK, AJ and DJ) and one adjudicator (CB) was assembled. Two of the four reviewers were trained in the Cochrane review methodology. The lead reviewer (MW) and one other primary reviewer (AJ) independently screened all identified papers by scanning titles and abstracts for appropriateness (first level screening). Pairs of reviewers then evaluated potentially relevant studies to see if they met our inclusion criteria (second level screening). The same pair then independently extracted data from accepted papers using a standardized form. Any disagreement was resolved through discussion and consensus. There was no situation in which a tiebreaker was needed. If necessary, study authors were contacted for additional information to complete the screening and abstraction process. 102 4.3.7 Assessment of methodological quality We assessed the risk of bias and overall methodological quality of the studies using the scale by van Tulder et al. (55) for RCTs and controlled clinical trials (CCTs). This 11-item quality assessment scale developed by the Cochrane Back Review Group was chosen for its relevance (Table 4.1). While designed as a checklist for aggregating items and generating a single quality score, this process is no longer favoured (56) and we present the individual quality criteria results for each trial. Table 4.1 Quality Assessment and Risk of Bias Checklist Criteria Scoring 1. Was the method of randomization adequate? yes no don't know 2. Was treatment allocation concealed? yes no unclear 3. Were groups similar at baseline regarding most important prognostic factors? yes no don't know 4. Was the patient blinded to the intervention? yes no don't know 5. Was the therapist/care provider blinded to the intervention? yes no don't know 6. Was the outcome assessor blinded to the intervention? yes no don't know 7. Were co-interventions avoided or similar? yes no don't know 8. Was compliance acceptable in all groups? yes no don't know 9. Was the drop-out rate described and acceptable (20% or less for post- test, 30% or less for long-term follow up) yes no don't know 10. Was the timing of the outcome assessment similar in both groups? yes no don't know 11. Was data for at least one key outcome analyzed by intention-to-treat (ITT)? yes no don't know   103 Initial inter-observer reliability of both the screening and quality assessment was determined on two papers prior to the actual review process. These related studies were drawn from a pool of articles that did not meet the inclusion criteria due to publication date or difference in intervention or patient population (55). Poor agreement on some items during this pilot test phase led to revisions to the screening and data extraction forms. After further discussion, agreement was reached on the screening and data abstraction process. Comprehensive guidelines for data abstraction were prepared and used by all reviewers. 4.3.8 Data analysis and synthesis Continuous data from individual trials were recorded as means and standard deviations. We contacted several authors for missing data. Evidence tables were created and a qualitative analysis performed using the GRADE approach as recommended by the Cochrane Musculoskeletal Review Group (57) (Table 4.2). Detailed methods for upgrading or downgrading studies are described in the Chapter 12.2.1 of the Cochrane Handbook for Systematic Reviews (56). Table 4.2 GRADE approach to assessing overall levels of quality of a body of evidence Underlying methodology Quality rating Randomized trials; or double-upgraded observational studies. High Downgraded randomized trials; or upgraded observational studies. Moderate Double-downgraded randomized trials; or observational studies. Low Triple-downgraded randomized trials; or downgraded observational studies; or case series/case reports. Very low   All authors independently examined the evidence tables for clinical and methodological heterogeneity taking into account the participants, interventions, controls (comparators), outcomes, timing of assessments, measurement tools and methodological quality of the original 104 studies (55). Due to lack of both clinical and methodological homogeneity among trials, we were unable to pool the data and perform a meta-analysis. This is a common problem among Cochrane reviews of allied health interventions (58). We do, however, present the individual studies' treatment effect sizes using weighted mean differences (WMD) from mean post-test values where possible. In addition, for the papers with small sample sizes (less than 50 per group) (59) and differing baseline values, we also calculated WMDs using mean change scores to better reflect the true effect. Since original data sets were not available, we used the formula suggested in the Cochrane Handbook, Chapter 16.1.3.2 (60) and a conservative imputed value of 0.8 as the correlation between pre- and post-test scores. Smaller correlation coefficients are known to overestimate effect sizes and thus the more conservative value was chosen (61). Forest plots were created to display effect estimates with 95% confidence intervals and samples are provided for select outcomes within individual trials. 4.4 Results 4.4.1 Description of studies A total of 2,534 English and non-English papers were identified. Of these, 47 were retained for second level screening and seven for full review. All of the articles were RCTs published in English and four of the trials were conducted outside of North America. The included trials were published between 2003 and 2009. (See Figure 4.1) 105  Figure 4.1 PRISMA flow chart for TKA review     Potentially relevant RCTs identified and screened for retrieval (n = 2, 534) RCTs excluded with reasons (n = 2, 460) Main reasons:  Not TKA-related  Not an intervention study  Not a rehabilitation intervention  No control group RCTs retrieved for more detailed evaluation (n = 74) RCTs excluded with reasons (n = 67) Main reasons:  Multidisciplinary intervention  Significant pre-op component  Acute care intervention  No TKA subgroup data  Mixed diagnoses (no OA subgroup data) Potentially appropriate RCTs to be included in the meta-analysis (n = 7) RCTs excluded  from meta-analysis with reasons (n = 7) (Due to clinical & methodological heterogeneity among trials) RCTs included in meta-analysis  (n = 0)   106 Several studies initially screened as being appropriate and subsequently excluded from this review are listed in Table 4.3. Table 4.3 Excluded studies and main reason for exclusion Study Main reason for exclusion Bulthuis Y et al. 2007 (62) No TKA subgroup data available Johnson AW 2007 (63) Less than 10 participants per group Kumar PJ et al. 1996 (64) Acute care interventions Mahomed NN et al. 2008 (65) No TKA subgroup data available Oehlert  K et al. 2004 (66) Could not confirm key aspects of trial methodology; no response from author Rajan RA et al. 2004 (67) Only ROM data reported Tal-Akabi A et al. 2007 (68) No TKA subgroup data available   4.4.2 Methodological quality The range in methodological quality of the included studies was low to high. Adequate randomization procedures were reported in five trials, not clear in one and satisfactory in one after discussion with the author. Concealed allocation, where neither the researcher nor participant is aware of group assignment in advance, was reported in four studies. Contacting authors subsequently confirmed appropriate concealment procedures in two further trials. Six of the seven trials reported blinding the outcome assessor. Blinding of the provider and patients is not possible in most cases in physiotherapy intervention studies. Short-term drop out rates were satisfactory in five studies, unclear in one and exceeded the acceptable short-term rate of 20% (55) in one. Only four studies included a long- term follow-up phase (12 months or more) with long-term drop out rates satisfactory in two and greater than the acceptable long-term rate of 30% in two. (See Table 4.4) 107 Table 4.4 Methodological quality and risk of bias of included trials Criterion Avramadis 2003 Codine 2004 Harmer 2009 Kramer 2003 Mockford 2008 Moffet 2004 Petterson 2009 1. Randomization Y ? Y Y* Y Y Y* 2. Concealment Y* ? Y Y* Y Y Y* 3. Groups similar ? N Y Y Y Y Y 4. Patient blinded N N N N N N N 5. Therapist blinded N N N ? N N ? 6. Assessor blinded Y* Y* Y Y ? Y Y 7. Co-interventions/ contamination avoided Y ? ? N N Y ? 8. Acceptable compliance Y Y* Y* Y Y Y Y 9. Acceptable drop- out rate (short term) Y ? Y ? Y Y N 10. Timing of outcome assessment similar Y Y Y Y Y Y Y 11. ITT analysis N ? Y † † † † Legend: Y = criterion was met; * = positive response was obtained by contacting the author and not from published article; ? = unclear, incomplete or no data provided; N = criterion was not met; NA = not applicable as there were no drop outs reported; † = performed but results not reported or only partially reported.  4.4.3 Participants A total of 779 participants were included in this review. Individual study sample sizes ranged from 30 to 200 participants. Mean patient ages ranged from 65 to 75 years and a majority was female in all but one trial. Race was not identified in any studies. General health, presence of co- 108 morbidities and details of the surgery were poorly described or not mentioned in most. (See Table 4.5) 109 Table 4.5 Summary of study characteristics  Study Design/ quality Setting/ Country Participants (1° TKA for OA) Exclusion criteria Intervention Comparison Outcomes Avramidis 2003    RCT (n=30)  Low Not clear; UK district hospital TG = mean age 68 (11) yrs, 80% ♀. CG = mean age 71 (8) yrs, 66% ♀ Symptomatic hip or ankle OA; history of epilepsy; use of pacemaker; pathologic skin condition in area of stimulation Start: 2 days post-op NMES to quads muscles, 2 hrs, 2x/day for 6 wks + conventional physio program „Conventional physio‟ program for 6 wks Baseline: pre-op HSS Knee Score, 3MWT, Physiologic Cost Index  Adverse events not discussed No LT follow-up Codine 2004    RCT (n=60)  Low IRF; France TG = mean age 75 (13) yrs, 70% ♀. CG = mean age 71 (15), 53% ♀ Significant muscular deficit, neurological disease or inflammatory reaction to isokinetic testing Start: 10 days post-op Isokinetic, eccentric hamstring training for 15 mins + standard rehab, 2 hrs, 5x/wk for 3 wks „Standard rehab‟ program of 2 hrs, 5x/wk for 3 wks Baseline: post-op KSS (post-test only), active knee ROM, isometric quads  Adverse events reported No LT follow-up Harmer 2009    RCT with 2 treatment arms (n=102)  High Outpatient dept, public hospital; Australia TG 1 = mean age 68 (6) yrs, 57% ♀. TG 2 = mean age 69 (9), 57% ♀ Deep joint infection, bilateral surgery or surgery on other joint within 6 months, dementia or other neurologic problem Start: 2 wks post-op TG 1: PT-led group exercise program in outpt dept, 2x/week for 6 weeks TG 2: PT-led group exercise program in community pool, 2x/wk for 6 wks Baseline: post-op 6MWT, WOMAC, pain (VAS), stair climbing power, passive knee ROM  Adverse events reported 6 month follow-up Kramer 2003   RCT with 2 treatment arms (n=160)  Moderate Outpatient (multi- site); Ontario, Canada At least 90° knee flex‟n & „functional‟ hip on surgical side. TG 1 = mean age 68 (7) yrs, 59% ♀; TG 2 = mean age 69 (8) yrs, 55% ♀ RA, major neurologic conditions Start: 2 wks post-op TG 1: Outpt physio for 1 hr, 2x/wk for 10 wks + daily home exercises TG 2: Home-based, 2-staged exercise program 3x/day, daily for 10 wks with periodic telephone consultation with PT Baseline: pre-op KSS, WOMAC, SF-36, 6MWT, stair climbing, active knee ROM  Adverse events partially reported 12 month follow-up  110 Table 4.5 Summary of study characteristics (continued)     Legend: TG = treatment or experimental group; CG = control or comparator group; IRF = inpatient rehabilitation facility; NMES = neuromuscular electrical stimulation; HSS = Hospital for Special Surgery; 3/6MWT = 3 or 6 minute walk test; KSS = Knee Society clinical rating system; ROM = range of motion; WOMAC = Western Ontario & McMaster Universities OA Index; VAS = visual analogue scale; RA = rheumatoid arthritis; SF-12/36 = MOS Short Form 12/36; TUG = timed up & go test, KOS ADL = Knee Outcome Survey Activity of Daily Living scale. Study/ language Design/ quality Setting/ country Participants (1° THA for OA) Exclusion criteria Intervention Comparison Outcomes Mockford 2008 RCT (n=150)  High  Outpatient dept; Ireland 4% had RA. TG = mean age 69, 65% ♀. CG = mean age 71, 58% ♀. Not described Start: mean 25 days post-op (range 6-54 days) Supervised outpt physio for 30 mins, 1-2x/wk for 6 wks Unsupervised home-based exercise program for 6 wks Baseline: pre-op Knee ROM, Oxford Knee Score, SF-12, Bartlett patellar score  Adverse events reported 12 month follow-up Moffett 2004  RCT (n=77)  High  Outpatient dept of rehab hospital; Quebec, Canada TG = mean age 67 (9) yrs, 63% ♀. CG = mean age 69 (8) yrs, 56% ♀ Previous leg surgery affecting gait, planned revision surgery on other leg joint, neuromuscular or degenerative disease, comorbidities limiting locomotion, major post- TKA complication Start: 2 months post-op Individual, PT-led progressive, functional task-oriented exercises & aerobic conditioning for 60-90 mins, 2x/wk for 6-8 wks. „Standard care‟ which included PT home visits for 26% of participants. Home program not described. Baseline: post-op 6MWT, WOMAC, SF-36  No adverse events reported 12 month follow-up Petterson 2009 RCT with 2 treatment arms + non-rand standard of care cohort (n=200)  Moderate Outpatient ; not-for- profit clinic; USA TG 1 = mean age 65 (9) yrs, 45% ♀. TG 2 = 65 (8) yrs, 47% ♀.  (Note: Only the 2 randomized groups considered.) Uncontrolled hypertension or diabetes, symptomatic OA in other knee, leg orthopaedic problem limiting function, neurologic impairment Start: 3-4 wks post-op TG 1: NMES to quads (10 reps) + progressive strength training program 2-3x/wk for 6 wks. TG 2: Progressive strength training program 2-3x/wk for 6 wks. Baseline: post-op Quads strength & voluntary activation, TUG, stair climbing test, 6MWT, SF-36, KOS ADL scale, knee ROM  Adverse events reported 12 month follow-up   111 4.4.4 Interventions No two trials evaluated similar post-acute physiotherapy interventions following primary TKA with regards to timing, duration, dosage, setting and type of rehabilitation. The timing of the start of the intervention ranged from immediately after surgery (two days post-op) to four months following surgery. Most studies fell into the intermediate range delivering the intervention between two and eight weeks post-operatively. The treatment duration ranged from three weeks to 10 weeks. Overall dosage was difficult to calculate in two trials and ranged from two times a week for six weeks (12 sessions in total) to two times a week for 10 weeks (20 sessions in total). All trials were classified as 'higher dosage' according to our categories established a priori. The setting for physiotherapy included an inpatient rehabilitation hospital in one trial, outpatient departments/clinics in five trials and was not clear in one trial (69). Adherence to the physiotherapy protocol or intervention was monitored in six trials and specific values were provided in four. One study included a subgroup analysis of high versus lower adherers with regards to treatment effects and reported no differences in outcomes for those patients who attended the prescribed nine outpatient sessions compared to those who completed eight or fewer (70). 4.4.5 Controls Control or comparator situations were not consistent among trials and included standard or usual physiotherapy (n=3), home exercises with reduced physiotherapy supervision (n=1), home exercise with no further supervision (n=1), a different exercise type/setting (pool-based) (n=1), and different intervention (electrical muscle stimulation) (n=1). Baseline characteristics of the treatment and control groups were similar in all trials; however, minimal baseline information was provided in most studies.  112 4.4.6 Outcomes More than 25 different outcome tools were used in the seven trials again making comparison across studies difficult. Post-operative pain (ICF level: body function) was evaluated in four trials. Physical functioning (ICF level: activity) was assessed in all seven studies and included performance measures (e.g. walking speed, stair climbing), self-report tools (e.g. WOMAC OA Index, Oxford Knee Score) and surgeon-rated tools (e.g. American Knee Society Score (KSS), Hospital for Special Surgery (HSS) scale). Four trials reported on HRQoL using generic tools (MOS Short Form 36 (SF-36) and MOS Short Form12 (SF-12)). For the secondary outcomes of interest, knee ROM data were most frequently reported; however, methods for assessing and reporting ROM varied and included active range measured by electronic goniometry and passive range using different testing positions. Additional secondary outcomes included quadriceps and hamstrings strength using varied assessment methods, knee edema, and a physiologic cost index related to gait. Patient satisfaction with the physiotherapy intervention or outcomes was not assessed in any trials. We do not include any further discussion on secondary outcomes due to their inconsistent use and reporting methods. The baseline assessment took place immediately prior to TKA surgery in three studies and post- operatively, prior to the intervention in four. The timing of subsequent assessments was similar for treatment and control groups in all trials. Four trials included a long-term follow-up assessment at one year post-op. Adverse events during the intervention study period were adequately documented in four studies:  Six cases of pain and/or swelling equally in both groups not requiring change in study protocol (71)  113  Peri-operative complications experienced in 18 participants after discharge from acute care including thromboemboli, superficial and deep wound infection, hemarthrosis and one death not attributed to interventions (72)  DVT, pulmonary embolis, superficial and deep wound infection and hematoma occurred equally in both groups and not attributed to intervention (70)  One case of dizziness following neuromuscular stimulation (73) Authors of three trials did not feel the adverse events were related to the treatment intervention while the other reported it was not clear (73). One study reported no adverse events (74), another provided incomplete information (75) and the remaining trial did not provide any information (69). 4.4.7 Risk of bias in included studies All trials had design or methodological issues putting them at risk for bias. Allocation (selection) bias may have been present in the two studies that did not describe adequate randomization with concealed allocation. However, as noted earlier, correspondence with the author of one of these studies suggested appropriate randomization techniques had been used. Measurement bias is of concern in the one trial that did not report using a blinded outcome assessor. Performance bias is attributable to unblinded participants and treating therapists and may have been an issue in all of the trials. One study reported withholding detailed study information from the treating therapists (73). Attrition bias due to higher than anticipated dropouts was evident in one study (73) and unclear in another (75). Another issue potentially contributing to performance bias is the lack of data on adherence to study protocols and exercise regimes. While six trials reported monitoring participants' adherence to the intervention and/or comparator in a prospective fashion (e.g. attendances,  114 exercise logs), few provided specific adherence data. This makes it difficult to determine the true effects of the intervention on the outcomes of interest or if there was a dose-response effect. That is, were there greater treatment effects in the subgroup of participants that adhered to the intervention compared to those who only partially completed the program? Similarly, the issue of therapist adherence to the intervention protocol (procedural reliability) was only assessed in one trial (73). While there is always the need for some individualization of therapeutic interventions to address the needs and preferences of patients, deviating from the study protocol complicates matters and introduces elements of co-intervention and even contamination that cannot be controlled for in the analysis if not monitored. Small sample sizes (less than the 50 participants per group standard) (59) in three of the trials may have contributed to underpowered studies and the risk of false conclusions (Type II error) about lack of treatment effect. Only three studies reported performing a power calculation using the primary outcome of interest to determine an appropriate sample size a priori (70, 72, 74). A fourth study stated a power calculation was undertaken to establish sample size; however, no details were provided (73). 4.4.8 Effects of interventions As noted earlier, we were unable to pool the data and perform a meta-analysis due to methodological and clinical heterogeneity among trials. Therefore, we performed a qualitative analysis and best evidence synthesis for the primary outcomes using the GRADE approach (57). For trials with adequate data, WMDs and 95% CI for the primary outcomes and are summarized below and sample forest plots are provided in Figures 4.2 to 4.7. Where appropriate, effect sizes based on mean change scores are compared to those based on post-test values for select trials.  115 4.4.8.1 Pain Four studies reported on post-operative pain following a variety of physiotherapy interventions with mixed results. Harmer et al. found no significant between group difference in WOMAC pain scores post-treatment (WMD 0.14, [-1.21, 0.93]) and six months post-op (WMD 0.82, [- 0.59, 2.23]) after six weeks of supervised out-patient exercises compared to PT-led pool exercises both initiated two weeks post-op (72). Kramer and colleagues reported no long term difference (12 months post-op) in pain assessed with the WOMAC (WMD 1.00, [-2.09, 4.09]) and Knee Society (WMD -3.00, [-6.77, 0.77]) rating scales after a 10-week out-patient clinic- based physiotherapy program compared to a home-based exercise program with telephone support (75). There were no between group differences in function-related pain after six weeks of progressive strengthening compared to progressive strengthening plus neuromuscular electrical stimulation (NMES) in an out-patient setting at three (WMD 0.03 [-0.26, 0.32]) and 12 months (WMD -0.07 [-0.41, 0.27]) post-op (73). Significant short term (WMD -7.60 [-14.15, -1.05]) and intermediate term (WMD -7.10 [-13.68, -0.52]) reductions in pain were reported by Moffet et al. following an intensive functional exercise program (8 weeks) compared to standard rehabilitation at home. These between group differences were not maintained at twelve months post-op (8 month follow-up) based on both mean follow-up values (WMD -2.40 [-8.44, 3.64]); however, were significantly better using mean change scores (WMD 8.20 [2.45, 13.95] (74). (See Figures 4.2 and 4.3) There is high quality evidence that:  Intensive outpatient functional physiotherapy for six to eight weeks results in less short- term pain which is maintained at two months follow-up compared to a standard home program  Six weeks of PT-led gym-based group exercises is no better than pool-based group exercises on short- and intermediate-term pain  116 There is moderate quality evidence that:  Immediate outpatient physiotherapy of six weeks duration results in no differences in long-term pain at rest or with activity when compared to home-based exercise with telephone support  The addition of NMES to a six-week out-patient exercise program results in no differences in short- and long-term pain compared to out-patient exercises alone Figure 4.2 Forest plot for WOMAC pain comparing mean 8-month follow-up values  Figure 4.3 Forest plot for WOMAC pain comparing mean 8-month change scores  4.4.8.2 Function All seven trials included assessment of function using a self-report tool, surgeon-rated tool or performance-based measure alone or in combination for varied interventions and with mixed results. Self-reported function Five trials included self-report tools with the WOMAC OA Index being the most common. Harmer et al. reported no significant improvements in function following a 6-week land-based physiotherapy program compared to a 6-week pool-based program immediately following  117 treatment (WMD 3.34, [-1.02, 7.70]) and six months post-op (WMD -1.50, [-4.60, 1.60]) (72). There were no long-term differences in WOMAC function scores when a 10-week supervised out-patient physiotherapy program was compared to a home-based program with PT telephone support (WMD 4.00, [-3.86, 11.86]) (75). Six weeks of outpatient physiotherapy resulted in a significant short-term difference in the Oxford Knee Score (WMD -3.20, [-6.04, -0.36]) compared to unsupervised home exercises; however, this was not maintained at 12 months follow-up (WMD 1.20, [-1.96, 4.36]) (70). There was no significant difference in WOMAC function following intensive outpatient physiotherapy (WMD -5.30, [-12.68, 2.08]) nor at two (WMD -6.20, [-13.77, 1.37]) and eight months (WMD -3.80, [-11.58, 3.98]) follow-up compared to standard home exercises (74). Significant between-group differences favouring the intervention group were found at all measurement points when mean change scores were used (Figures 4.4 and 4.5) (74). Petterson et al. reported no significant differences in the Knee Outcome Survey – Activity of Daily Living Scale (KOS-ADLS) immediately following (WMD - 0.01, [-0.05, 0.02]) and 12 months after (WMD -0.01, [-0.05, 0.03]) a 6-week progressive strengthening program plus NMES compared to a strengthening program alone (73). There is consistent high level evidence that:  Varied forms of outpatient physiotherapy including the use of NMES results in no long- term differences in self-reported function when compared to „standard‟ care or pool- based exercises. There is conflicting high quality evidence that:  Short-term improvements in self-reported function can be achieved with supervised out- patient physiotherapy compared to „standard‟ home-based care  118 Figure 4.4 Forest plot for WOMAC function comparing mean post-test values  Figure 4.5 Forest plot for WOMAC function comparing mean post-test change scores  Surgeon-rated function Three trials used a surgeon-rated tool to assess function after varied interventions. There was no significant difference in the Hospital for Special Surgery (HSS) Knee Score following six weeks of daily NMES at home plus conventional physiotherapy compared to physiotherapy alone immediately following and at 12-weeks follow-up; however, effect sizes could not be calculated due to incomplete data (69). Using the Knee Society Clinical Rating System, Kramer et al. found no differences 12 months post-operatively (WMD -7.00, [-16.29, 2.29]) when outpatient physiotherapy was compared to home-based therapy with PT telephone support (75). Effect sizes could not be calculated for the third trial as only post-test data were available (71). There is low quality evidence that:  Conventional physiotherapy plus daily NMES results in no short- and intermediate-term differences in surgeon-rated function compared to physiotherapy alone (effect sizes not calculated)  119 There is moderate quality evidence that:  Supervised outpatient physiotherapy results in no long-term differences in surgeon-rated function when compared to home-based exercises with PT telephone support Performance-based function Five trials reported on function using five different performance-based tools and following varied interventions.  Walking speed Walking speed was assessed in five studies. Avramidis et al. reported significant differences in 3-minute walk test (3-MWT) distances immediately following six weeks of daily NMES plus conventional physiotherapy and at 12-week follow-up compared to conventional physiotherapy alone; however, effect sizes could not be calculated due to incomplete data (69). There were no significant differences in walking speed as assessed by the 6-minute walk test (6-MWT) immediately following (WMD -12.20, [-54.76, 30.36]) and at six months (WMD -1.12, [-40.68, 38.44]) after a 6-week gym-based physiotherapy program compared to pool-based program (72). Supervised outpatient physiotherapy resulted in no differences in 6-MWT at 12 months (WMD - 25.00, [-64.69, 14.49]) when compared to home-based exercises with PT telephone support (75). There were no significant differences in 6-MWT immediately post-treatment (WMD 31.00, [- 7.46, 69.46]), at two months (WMD 31.80, [-6.69, 70.29]) and eight months (WMD 30.00, [- 11.14, 71.14]) following intensive outpatient physiotherapy when compared to standard home- based exercises (74). These differences remained non-significant using mean change scores while the authors reported significant between-group differences favouring the treatment group at post-test and 2-month follow-up using an analysis of covariance method to calculate effect sizes (74). Petterson et al. reported no short-term (WMD -5.00, [-40.36, 30.36]) or long-term  120 (WMD -9.00, [-47.60, 29.60]) differences in 6-MWT with the addition of NMES to a standard outpatient exercise program (73). Timed stair climb Timed stair climbing was used in two trials and no significant differences were found after supervised outpatient physiotherapy (WMD -4.00, [-9.57, 1.57]) compared to home-based exercises with PT telephone support (75) nor immediately following NMES as an adjunct to standard outpatient physiotherapy (WMD 1.50, [-0.07, 3.07])(73). There was a significant between-group difference favouring the exercise only group at 12-month follow-up (WMD 1.87, [0.21, 3.53]) (73). A variation on this test, stair climbing power, was used in a third trial and no significant differences were found immediately following (WMD -2.77, [-28.27, 22.73]) and at six months (WMD 17.17, [-9.81, 44.15]) after six weeks of gym-based group physiotherapy compared to pool-based group exercises (72). Timed Up and Go (TUG) The TUG test was used in one trial and no significant short-term (WMD 0.27, -0.31, 0.85]) and long-term (WMD 0.39, [-0.20, 0.98]) differences were found when comparing an outpatient progressive strength training program with and without NMES (73). For performance-based function, there is moderate to high quality evidence that:  Outpatient progressive strength training with and without NMES results in no differences in 6-MWT distances, short-term stair climbing speed or TUG scores  Supervised outpatient physiotherapy does not result in longer 6-MWT distances or faster stair climbing speed compared to home-based exercises with PT telephone support  An outpatient intensive functional exercise program does not result in longer 6-MWT distances compared to a standard home-based exercise program  121  PT-led group pool- and gym-based exercise programs result in similar improvements in 6-MWT distances and stair climbing power There is low quality evidence that:  Daily NMES as an adjunct to conventional physiotherapy results in greater 3-MWT distances compared to conventional physiotherapy alone (effect sizes not calculated) 4.4.8.3 HRQoL Four trials reported on HRQoL using the SF-36 or SF-12 generic tools. Kramer et al. reported no significant long-term differences in the SF-36 total score (WMD -1.00, [-6.66, 4.66]) with 10 weeks of outpatient physiotherapy compared to home-based exercises with telephone support (75). Significant between-group differences in the SF-36 MCS favouring the standard care group were found at 2-month follow-up (WMD -4.00, [-6.84, -1.16]) compared to the intensive functional rehabilitation program (74). This difference was no longer evident at the 8-month follow-up (WMD -3.40, [-7.44, 0.64). No differences were noted in the SF-36 PCS at any measurement point (Figure 4.6, 2-month follow-up). Using mean change scores, significant group differences were found for the SF-36 PCS post-treatment (WMD 3.00, [0.35, 5.65]), and at two months (WMD 6.00, [3.26, 8.74]) (Figure 4.7) and eight months (WMD 4.40, [1.58, 7.22]) follow-up (74). Similarly, between group differences for SF-36 MCS were present at all measurement points using mean change scores. Again a difference in analytical approach to calculating effect size allowed the authors to conclude differently with significant between-group differences in favour of intensive functional rehabilitation only reported at 2-month follow-up for both the MCS and PCS of the SF-36 (74). Petterson et al. found no significant differences in the SF-36 PCS (WMD 0.19, [-2.63, 2.25]) and MCS (WMD 0.40, [-1.81, 2.61]) immediately following and in the PCS (WMD -0.69, [-3.72, 2.34] and MCS (WMD -0.53, [-2.54, 1.48]) 12  122 months after a progressive outpatient strengthening program with NMES compared to strength training alone (73). Using the SF-12, Mockford et al. reported no significant short-term differences in the PCS (WMD 2.30, [-0.98, 5.58]) and MCS (WMD 2.10, [-1.13, 5.33]) after standard outpatient physiotherapy compared to unsupervised home exercises (70). These differences remained non-significant at 12-month follow-up. There is conflicting high quality evidence that:  Intensive outpatient functional physiotherapy results in greater short- and long-term improvements in both the mental and physical components of HRQoL compared to standard home exercises (using mean change scores)  Supervised outpatient physiotherapy of six and 10 weeks duration results in no short- and long-term differences in HRQoL compared to a home exercise program with no or reduced PT supervision  Progressive strength training coupled with NMES does not result in better short- and long-term HRQoL when compared to strength training alone Figure 4.6 Forest plot for SF-36 PCS comparing mean 2-month follow-up values   123 Figure 4.7 Forest plot for SF-36 PCS comparing mean 2-month follow-up change scores  4.5 Discussion 4.5.1 Summary of main results This review synthesized the research published between 1990 and May 2009 that examined the effects of post-acute physiotherapy after primary TKA on pain, function and HRQoL. There was limited evidence of low to high quality to support the use of various forms of post-acute physiotherapy for pain, function and HRQoL. We found favourable between-group results for: short- and intermediate-term pain and HRQoL following an intensive outpatient functional rehabilitation program compared to a standard home physiotherapy program; short-term self- reported function after supervised outpatient physiotherapy versus „standard‟ home-based care; and walking speed after daily NMES as an adjunct to conventional physiotherapy. We determined there were no significant between-group differences for surgeon-rated function, TUG score and timed stair climb, and inconsistent differences for HRQoL based on mean post-test and follow-up values. However, in several cases, differences were evident when mean change scores were used to calculate effect size and in most cases, the findings favoured the experimental group. Due to the heterogeneity among studies, it was not possible to perform a meta-analysis and draw conclusions across studies as to whether one form of physiotherapy was superior to another following TKA surgery. The limited between-group differences are largely explained by the fact that no trials compared a physiotherapy intervention to a 'no treatment' or placebo control group.  124 Clinically important within-group benefits were evident in most groups that received the targeted or experimental intervention (72, 73, 75), but one cannot draw conclusions about the magnitude of the added benefit of physiotherapy compared to the natural course of recovery after TKA. This lack of accumulated evidence is not evidence of a lack of effectiveness, but speaks to the need for high quality trials with carefully specified interventions and uniform outcome measures. 4.5.2 Quality of the evidence The overall quality of trials included in this review was low to high with some trials being downgraded due to lack of concealed allocation and one for using unblinded assessors (57). As noted earlier, it is almost impossible to blind patients and providers due to the nature of the interventions. Three trials were likely under powered to detect meaningful between group differences. Reporting issues were also evident in the included trials with few studies identifying their primary outcome and only three authors reporting they performed a power calculation using this outcome to determine necessary sample size. Participant demographics, particularly presence and severity of co-morbidities, were poorly described in most trials. A growing body of research demonstrates that co-existing health problems are prevalent in individuals undergoing TKA (76, 77) and influence the post-operative course of recovery and TKA outcomes (78). Therefore, co- morbidities should be reported and considered in the analyses of outcomes after TKA rehabilitation. 4.5.3 Strengths of this review To our knowledge, this is the most comprehensive systematic review to date examining the varied forms of physiotherapy commonly delivered during the first year after TKA surgery. Our extensive search strategy resulted in a large body of published and unpublished studies and our  125 inclusion of non-English trials reduced the risk of language bias. We made every effort to contact authors for missing data and methodological clarification. Six different countries and respective health care systems are represented in our review. However, it is possible that some relevant trials are missing and those published since May 2009 are not included. There remain large gaps in the evidence and a paucity of high quality trials on physiotherapy interventions after TKA. 4.5.4 Weaknesses and potential biases of this review Due to trial heterogeneity, we were unable to perform the planned meta-analysis and make definitive statements about the strength of the evidence and magnitude of treatment effects for physiotherapy interventions after TKA. Limiting our review to RCTs and CCTs precluded other study designs that may have contributed additional evidence on the topic of post-acute physiotherapy after TKA. We made every effort to avoid various forms of bias by pursuing unpublished reports in our search strategy, permitting non-English studies, using two authors to independently screen and review papers, and performing a rigorous and standardized quality assessment process. We did not create funnel plots to examine the possibility of publication bias due to the small number of included trials. 4.5.5 Agreements and disagreements with other studies or reviews A recently published Cochrane review compared the effect of multidisciplinary rehabilitation interventions to single discipline care following primary and revision THA and TKA for OA and rheumatoid arthritis (RA) (50). This review, however, focused on the acute care phase of rehabilitation and included use of clinical pathways, formal patient education programs, post- operative pain management and dietary interventions in its range of outcomes. Studies examining physiotherapy interventions alone were excluded. Due to clinical heterogeneity and low study quality, a meta-analysis was not performed. The authors concluded that there was silver level (low to moderate) evidence to support early multidisciplinary rehabilitation in improving activity  126 and participation outcomes; however, optimal intensity, frequency and effects of rehabilitation warranted further study. In order to support their development of evidence-based guidelines for ambulatory physiotherapy following TKA, the French Society of Physical and Rehabilitation Medicine (SOFMER) conducted a systematic review of French and English language articles from an unspecified time frame (79). A meta-analysis was not possible due to trial heterogeneity. Based on their analysis of 16 studies of mixed methodologies and inclusion criteria, the authors concluded that there were some advantages of ambulatory physiotherapy on muscle strength and function. From their limited economic analysis, no differences were found comparing home therapy to other approaches. In 2005, two separate reviews were conducted to inform rehabilitation practice and education in Ontario, Canada. In a review by the Medical Advisory Secretariat of the Ministry of Health and Long-Term Care, it was concluded that there was limited high-quality evidence (from one published abstract) to support the use of home-based physiotherapy instead of inpatient treatment after THA and TKA (80). The English-only literature was reviewed up to 2004 to inform the Greater Toronto Area Rehabilitation Network's development of an online discussion forum to promote collaboration and education among patients and health professionals (81). Based on this review of experimental and observational studies spanning pre-operative, acute and post- operative care, rehabilitation was found to be an effective component in TJA management at various stages of recovery. Best practice recommendations were not clearly differentiated for hip versus knee replacement. Soever and MacKay also concluded that comparison of outpatient care to other rehabilitation processes warrants further study to determine the optimal processes of care following TJA (81).  127 In 2008, a systematic review on the effects of therapeutic exercise after discharge from hospital for primary TKA surgery was published (51). This review of trials up to April 2007 included interventions delivered beyond one year post-op and excluded those that involved inpatient rehabilitation and electrical modalities as a treatment adjunct. While there is obvious overlap in the trials included in this and our review, our review is more comprehensive in that it includes inpatient treatment, which is routine rehabilitation practice in many countries (26, 82). The authors performed a meta-analysis for function, walking speed, ROM and HRQoL outcomes despite clinical heterogeneity among the interventions. They found small to moderate pooled effects for short-term function, ROM and HRQoL. Pooled studies were of varied exercise interventions at different time frames, of mixed durations and used different outcome tools. Regardless, Minns Lowe et al. similarly concluded that trial availability, quality and diversity prevented definitive answers regarding effectiveness of post-discharge physiotherapy exercise for primary TKA and recommended further research on the topic (51). 4.6 Conclusions 4.6.1 Implications for practice There is insufficient evidence from the studies included in this review, to suggest that any one type, setting or timing for post-acute physiotherapy for TKA is superior to another. The tremendous variety in the timing, duration, intensity and content of the interventions make developing clinical recommendations problematic. Additional high quality trials with sufficient power and standardized outcome assessment methods are needed to establish specific treatment recommendations for post-acute PT after TKA. With the longstanding impairments, functional limitations and participation restrictions present in individuals awaiting primary TKA (47, 76, 83-87) and still evident as much as two years post-  128 op (37-40), it may be inappropriate and unethical to withhold post-operative physiotherapy all together. The physiotherapy setting may be less important than the type, quality, dosage and adherence to the intervention; however, dose-response calculations were not possible and limited adherence data were available in the trials selected, and further evidence is required to support this conclusion.  Despite the limitations in the availability and quality of evidence, there are a number of implications for clinical practice from the high quality trials:  There is limited evidence from one high quality trial that a more intensive, functional exercise program results in better pain relief, self-reported function and HRQoL when compared to a standard home-based physiotherapy program (74)  There is limited, high quality evidence to support the use of electrical muscle stimulation as an adjunct to progressive strengthening exercises in improving stair climbing ability 12 months after surgery (73)  There is limited evidence from one high quality trial suggesting that similar improvements in pain and function can be achieved with a PT-led group pool-based exercise program as with a gym-based program (72)  There is limited evidence from one high quality trial to support a 6-week supervised outpatient physiotherapy program for improving short-term self-reported function compared to unsupervised home exercises (70).  129 4.6.2 Implications for research The heterogeneity among trials regarding the timing and types of interventions, control situations, outcomes and measures highlights the need for greater standardization in the conducting, evaluating and reporting of future research trials in TKA rehabilitation. Agreement on core outcomes and measures would facilitate comparison among different intervention approaches, settings, timing and costs. Additional suggestions for future research in this field include:  Conduct trials with strong study designs, concealed allocation to groups, blinded assessors and large sample sizes to ensure adequate power  Prospectively monitor participants' adherence to the study protocol (intervention) with subgroup analysis of treatment effects for high and low adherers (dose-response)  Prospectively monitor providers' adherence to the study protocol  Provide more detailed descriptions of the therapeutic interventions including timing, duration, frequency, intensity and progressions  Determine what baseline and post-operative contextual factors (personal and environmental) are most strongly related to rehabilitation outcomes and patient satisfaction, and perform appropriate subgroup analyses  Assess how patient preference for type, timing and/or setting for rehabilitation affects outcomes  Identify a core set of outcome measures for each of the ICF domains to allow for consistent, standardized reporting of outcomes and comparison across interventions  130  Include measures that capture participation, HRQoL and patient satisfaction  Include a follow-up phase of at least 12 months and monitor and report on physical activity levels and use of health care services or rehabilitative therapies during this period  Conduct multi-site trials to increase sample sizes and generalizability of findings  Include an economic analysis including both direct and indirect costs, and patient out-of- pocket expenses during the intervention and follow-up phases  131 4.7 References 1. Canadian Institute for Health Information (CIHI). Hip and total knee replacements in Canada - Canadian Joint Replacement Registry (CJRR). 2008-2009 Annual Report. [cited 2009 Sep 17]. Available from: URL: www.cihi.ca/cjrr. 2. Agency for Healthcare Research and Quality. Facts and figures 2007. Healthcare Cost and Utilization Project (HCUP). [cited 15 Jan 2010]. Available from: URL: http://www.hcup- us.ahrq.gov/reports/factsandfigures/2007/exhibit4_1.jsp. 3. American Academy of Orthopedic Surgeons. Facts on hip replacements and Facts on knee replacements. [cited 2009 Dec 9]. Available from: URL: http://www.aaos.org/research/stats/patientstats.asp. 4. The Norwegian Arthroplasty Register. 2007 Annual Report [homepage on the Internet]. . 2007 October 2007. Available from: http://www.haukeland.no/nrl/eng/Report2007.pdf. 5. New Zealand Orthopaedic Association. The New Zealand Joint Registry, Nine year report. [cited 2009 Jan 15]. Available from: URL: http://www.cdhb.govt.nz/NJR/reports/A2D65CA3.pdf. 6. Australian Orthopaedic Association National Joint Replacement Registry. Annual report 2009. [cited 15 Jan 2010]. Available from: URL: http://www.dmac.adelaide.edu.au/aoanjrr/documents/aoanjrrreport_2009.pdf. 7. Scottish Arthroplasty Project (SAP). Scottish Arthroplasty Project Annual report 2009. [cited 2009 Dec 3]. Available from: URL: http://www.arthro.scot.nhs.uk/Reports/Scottish_Arthroplasty_Project_Report_2009.pdf. 8. DeFrances CJ, Lucas CA, Verita CB, Golosinskiy A. 2006 National hospital discharge survey. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Center for Health Statistics; 2008 Jul [cited 2009 Nov 21]. Report No.: 5. Available from URL: http://www.cdc.gov/nchs/data/nhsr/nhsr005.pdf. 9. March LM, Bagga H. Epidemiology of osteoarthritis in Australia. Med J Aust 2004;180(5 Suppl):S6-S10. 10. Badley EM, Crotty M. An international comparison of the estimated effect of the aging of the population on the major cause of disablement in musculoskeletal disorders. J Rheumatol 1995;22(10):1934-40. 11. Kim S. Changes in surgical loads and economic burden of hip and knee replacements in the US: 1997 - 2004. Arthritis Rheum 2008;59(4):481-8. 12. Jones D, Westby M, Greidanus N, Johanson NA, Krebs DE, Robbins L, et al. Update on hip and knee arthroplasties: Current state of evidence. Arthritis Rheum 2005;53(5):772-80.  132 13. Tian W, DeJong G, Brown M, Hsieh CH, Zamfirov ZP, Horn SD. Looking upstream: Factors shaping the demand for postacute joint replacement rehabilitation. Arch Phys Med Rehabil 2009;90(8):1260-8. 14. Lidgren L, Robertsson O, W-Dahl A. The Swedish Knee Arthroplasty Register - Annual report 2009. [cited 15 Jan 2010]. Available from: URL: http://www.knee.nko.se/english/online/uploadedFiles/113_SVK2009ENGL1.0.pdf. 15. National Joint Registry for England and Wales. 6th Annual report. 2009 [cited 2009 Dec 8]. Available from: URL: http://www.njrcentre.org.uk. 16. Brander VA, Stulberg SD. Rehabilitation after hip- and knee-joint replacement: An experience- and evidence-based approach to care. Am J Phys Med Rehabil 2006;85(Suppl):S98- S118. 17. Ethgen O, Bruyere O, Richy F, Dardennes C, Reginster JY. Health-related quality of life in total hip and total knee arthroplasty. A qualitative and systematic review of the literature. J Bone Joint Surg Am 2004;86-A(5):963-74. 18. Salmon P, Hall GM, Peerbhoy D, Shenkin A, Parker C. Recovery from hip and knee arthroplasty: Patients' perspective on pain, function, quality of life, and well-being up to 6 months postoperatively. Arch Phys Med Rehabil 2001;82(3):360-6. 19. Maravic M, Landais P. Usefulness of a national database to evaluate the burden of primary joint replacement for coxarthrosis and gonarthrosis in patients aged over 40 years. Osteoarthritis Cartilage 2006;14(6):612-5. 20. Hawker GA, Badley EM, Croxford R, Coyte PC, Glazier RH, Guan J, et al. A population- based nested case-control study of the costs of hip and knee replacement surgery. Med Care 2009;47(97):732-41. 21. Nigam M. 2006_13 UBC_CSI completed cases by period additional cases 1. Centre for Surgical Innovation, Vancouver Coastal Health Authority; 2007 Apr. 22. Johanson NA. The hospital course in total hip and knee arthroplasty: Which length of stay is right? J Rheumatol 1997;24(1):1-2. 23. Naylor J, Harmer A, Fransen M, Crosbie J, Innes L. Status of physiotherapy rehabilitation after total knee replacement in Australia. Physiother Res Int 2006;11(1):35-47. 24. Oatis CA, Brennan JL, Kushman S, Werner A, Franklin PD. Describing "usual" physical therapy practice following total knee replacement in Montgomery County, Pennsylvania. Arthritis Rheum 2009;60(10 Suppl):S425. 25. Lingard EA, Berven S, Katz JN, Kinemax Outcomes Group. Management and care of patients undergoing total knee arthroplasty: Variations across different health care settings. Arthritis Care Res 2000;13(3):129-36.  133 26. Roos EM. Effectiveness and practice variation of rehabilitation after joint replacement. Curr Opin Rheumatol 2003;15(2):160-2. 27. Troughton D, Pearce J. Total hip and knee replacement survey: Rehabilitation practices through the continuum of care. Rehabilitation Services, Vancouver Island Health Authority; 2007. 28. Oldmeadow LB, McBurney H, Robertson VJ. Hospital stay and discharge outcomes after knee arthroplasty: Implications for physiotherapy practice. Aust J Physiother 2002;48(2):117-21. 29. Ganz SB, Viellion G. Pre and post surgical management of the hip and knee. In: Wegener ST, Belza BL, Gall EP, editors. Clinical Care in the Rheumatic Diseases. Atlanta: American College of Rheumatology; 1996. p. 103-6. 30. Kampa K. Orthopedic assessment and treatment of the geriatric patient. Lewis CB and Knortz KA, editors. St. Louis (MO): CV Mosby Co.; 1993. 31. Cahill JB, Koxman LM. Total knee arthroplasty. In: Cioppa-Mosca J, Cahill JB, Cavanaugh JT, Corradi-Scalise D, Rudnick H, Wolff AL, editors. Postsurgical Rehabilitation Guidelines for the Orthopedic Clinician. St. Louis: Hospital for Special Surgery, Mosby Elsevier; 2006. p. 17- 28. 32. Bizzini M, Boldt J, Munzinger U, Drobny T. Rehabilitation guidelines after total knee arthroplasty. Orthopade 2003;32(6):527-34. 33. Enloe LJ, Shields RK, Smith K, Leo K, Miller B. Total hip and knee replacement treatment programs: A report using consensus. J Orthop Sports Phys Ther 1996;23(1):3-11. 34. National Institutes of Health. NIH Consensus Development Conference on total knee replacement. NIH Consens State Sci Statements 2003 Dec 8-10 [cited 2004 Mar 17];20(1):1-34. Available from: URL: http://consensus.hin.gov/2003/2003TotalKneeReplacement117PDF.pdf. 35. Mont MA, Seyler TM. Advanced techniques for rehabilitation after total hip and knee arthroplasty. Clin Orthop Relat Res 2009 [cited 2009 Aug 13];467(6). Available from: URL: http://www.springerlink.com/content/5j7t7h3510r203t5/. 36. Givens-Heiss DL, Krebs DE, Riley PO, Strickland EM, Fares M, Hodge WA, et al. In vivo acetabular contact pressures during rehabilitation, part II: Postacute phase. Phys Ther 1992;72(10):700-10. 37. Boutron I, Poiraudeau S, Ravaud JF, Baron G, Revel M, Nizard R, et al. Disability in adults with hip and knee arthroplasty: A French national community based survey. Ann Rheum Dis 2003 Aug;62(8):748-54. 38. Brander VA, Stulberg SD, Adams AD, Harden RN, Bruehl S, Stanos SP, et al. Predicting total knee replacement pain: A prospective, observational study. Clin Orthop Relat Res 2003;416:27-36.  134 39. McClelland JA, Webster KE, Feller JA. Gait analysis of patients following total knee replacement: A systematic review. Knee 2007;14(4):253-63. 40. Meier W, Mizner RL, Marcus RL, Dibble LE, Peters C, Lastavo PC. Total knee arthroplasty: Muscle impairments, functional limitations, and recommended rehabilitation approaches. J Orthop Sports Phys Ther 2008;38(5):246-56. 41. Mizner RL, Petterson SC, Snyder-Mackler L. Quadriceps strength and the time course of functional recovery after total knee arthroplasty. J Orthop Sports Phys Ther 2005;35(7):424-36. 42. Pap G, Meyer M, Weiler HT, Machner A, Awiszus F. Proprioception after total knee arthroplasty: A comparison with clinical outcome. Acta Orthop Scand 2000;71(2):153-9. 43. Silva M, Shepherd EF, Jackson WO, Pratt JA, McClung CD, Schmalzried TP. Knee strength after total knee arthroplasty. J Arthroplasty 2003;18(5):605-11. 44. Walsh M, Woodhouse LJ, Thomas SG, Finch E. Physical impairments and functional limitations: A comparison of individuals 1 year after total knee arthroplasty with control subjects. Phys Ther 1998;78(3):248-58. 45. Valtonen A, Pöyhönen T, Heinonen A, Sipilä S. Muscle deficits persist after unilateral knee replacement and have implications for rehabilitation. Phys Ther 2009;89(10):1072-9. 46. Ackerman I, Graves S, Bennell K, Osborne R. Persistent psychosocial morbidity after joint replacement surgery: Health-related quality of life and psychological distress do no approach population norms after 12 months. (abstract). Association of Rheumatology Health Professionals Annual Scientific Meeting. 2006. 47. Foley SJ, Lord SR, Srikanth V, Cooley H, Jones G. Falls risk is associated with pain and dysfunction but not radiographic osteoarthritis in older adults: Tasmanian older adult cohort study. Osteoarthritis and Cartilage 2006;14(6):533-9. 48. Pandya NK, Draganich LF, Mauer A, Piotrowski GA, Pottenger L. Osteoarthritis of the knees increases the propensity to trip on an obstacle. Clin Orthop Relat Res 2005;431:150-6. 49. Milne S, Brosseau L, Robinson V, Noel MJ, Davis J, Drouin H, et al. Continuous passive motion following total knee arthroplasty. Cochrane Database Syst Rev 2007(3). 50. Khan F, Ng L, Gonzalez S, Hale T, Turner-Stokes L. Multidisciplinary rehabilitation programmes following joint replacement at the hip and knee in chronic arthropathy. Cochrane Database Syst Rev. 2008(2):CD004957. 51. Minns Lowe CJ, Barker KL, Dewey M, Sackley CM. Effectiveness of physiotherapy exercise after knee arthroplasty for osteoarthritis: Systematic review and meta-analysis of randomised controlled trials. Br Med J 2007;335(7624):812. 52. Monaghan B, Caulfield B, O'Mathúna DP. Surface neuromuscular electrical stimulation for quadriceps strengthening pre and post total knee replacement (protocol). Cochrane Database Syst Rev 2008(2):CD007177.  135 53. Westby MD, Backman CL. Patient and health professional views on rehabilitation practices and outcomes following total hip and knee arthroplasty: A focus group study. Submitted 2009 Sep 3. 54. World Health Organization. International classification of functioning, disability and health (ICF). . 2001 [cited 2004 Mar 3]. Available from: URL: http://www.who.int/classifications/icf/. 55. van Tulder M, Furlan A, Bombardier C, Bouter L. Updated method guidelines for systematic reviews in the Cochrane Collaboration Back Review Group. Spine 2003;28(12):1290-9. 56. Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions 5.0.2 [updated 2009 sep]. The Cochrane Collaboration [cited 2009 Dec 15]. Available from: URL: http://www.cochrane-handbook.org. 57. Schunemann HJ, Oxman AD, Vist GE, Higgins JPT, Deeks JJ, Glasziou P, et al. Interpreting results and drawing conclusions. In: Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions 5.0.2. The Cochrane Collaboration; 2009. 58. van den Ende CHM, Steultjens EMJ, Bouter LM, Dekker J. Clinical heterogeneity was a common problem in cochrane reviews of physiotherapy and occupational therapy. J Clin Epidemiol 2006;59(9):914-9. 59. Tugwell P, Shea B, Boers M, Simons L, Strand V, Wells G. Evidence-based rheumatology. London (ENG): BMJ Books; 2004. 60. Higgins JPT, Deeks JJ, Altman DG. Special topics in statistics. In: Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions 5.0.2. The Cochrane Collaboration; 2009. 61. Ray JW, Shadish WR. How interchangeable are different estimators of effect size? J Consult Clin Psychol 1996;64(6):1316-25. 62. Bulthuis Y, Drossaers-Bakker KW, Taal E, Rasker J, Oostveen J, van't Pad Bosch P, et al. Arthritis patients show long-term benefits from 3 weeks intensive exercise training directly following hospital discharge. Rheumatology 2007;46(11):1712-7. 63. Johnson AW. Whole-body vibration compared to traditional physical therapy in individuals with total knee arthroplasty [dissertation]. Brigham Young University, 2007. 64. Kumar PJ, McPherson EJ, Dorr LD, Wan Z. In: Laskin RS, editor. Rehabilitation after total knee arthroplasty: A comparison of 2 rehabilitation techniques. Annual meeting -1996; Lippincott; 1996;p. 93-101. 65. Mahomed NN, Davis AM, Hawker G, Badley E, Davey JR, Syed KA, et al. Inpatient compared with home-based rehabilitation following primary unilateral total hip or knee replacement: A randomized controlled trial. J Bone Joint Surg 2008;90(8):1673-80. 66. Oehlert K, Hassenpflug J. Coordinative abilities of arthroplasty patients - efficacy of a training program. Z Orthop Ihre 2004;142(6):679-84.  136 67. Rajan RA, Pack Y, Jackson H, Gillies C, Asirvatham R. No need for outpatient physiotherapy following total knee arthroplasty: A randomized trial of 120 patients. Acta Orthop Scand 2004;75(1):71-3. 68. Tal-Akabi A, Steiger U, Villiger PM. Neuromuscular adaptation to early post-operative, highintensity, short resistance training of non-operated lower extremity in elderly patients: A randomized controlled trial. J Rehabil Med 2007;39(9):724-9. 69. Avramidis K, Strike PW, Taylor PN, Swain ID. Effectiveness of electric stimulation of the vastus medialis muscle in the rehabilitation of patients after total knee arthroplasty. Arch Phys Med Rehabil 2003;84(12):1850-3. 70. Mockford BJ, Thompson NW, Humphreys P, Beverland DE. Does a standard outpatient physiotherapy regime improve the range of knee motion after primary total knee arthroplasty? J Arthroplasty 2008;23(8):1110-4. 71. Codine P, Dellemme Y, Denis-Laroque F, Herisson C. The use of low velocity submaximal eccentric contractions of the hamstring for recovery of full extension after total knee replacement: A randomized controlled study. Isokinetics Exerc Sci 2004;12(3):215-8. 72. Harmer AR, Naylor JM, Crosbie J, Russell T. Land-based versus water-based rehabilitation following total knee replacement: A randomized, single-blind trial. Arthritis Rheum 2009;61(2):184-91. 73. Petterson SC, Mizner RL, Stevens JE, Raisis L, Bodenstab A, Newcomb W, et al. Improved function from progressive strengthening interventions after total knee arthroplasty: A randomized clinical trial with an imbedded prospective cohort. Arthritis Rheum 2009;61(2):174- 83. 74. Moffet H, Collet JP, Shapiro SH, Paradis G, Marquis F, Roy L. Effectiveness of intensive rehabilitation on functional ability and quality of life after first total knee arthroplasty: A single- blind randomized controlled trial. Arch Phys Med Rehabil 2004;85(4):546-56. 75. Kramer JF, Speechley M, Bourne R, Rorabeck C, Vaz M. Comparison of clinic- and home- based rehabilitation programs after total knee arthroplasty. Clin Orthop 2003;(410):225-34. 76. Philbin EF, Groff GD, Ries MD, Miller TE. Cardiovascular fitness and health in patients with end-stage osteoarthritis. Arthritis Rheum 1995;38(6):799-805. 77. Desmeules F, Dionne CE, Belzile É, Bourbonnais R, Frémont P. Waiting for total knee replacement surgery: Factors associated with pain, stiffness, function and quality of life. BMC Musculoskelet Disord [cited 2010 Jan 21];10:52. Available from: URL: http://www.biomedcentral.com/1471-2474/10/52. 78. Tuominen U, Blom M, Hirvonen J, Seitsalo S, Lehto M, Paavolainen P, et al. The effect of co-morbidities on health-related quality of life in patients placed on the waiting list for total joint replacement. Health Qual Life Outcomes 2007;15:5-16.  137 79. Genêt F, Mascard E, Coudeyre E, Revel M, Rannou F. The benefits of ambulatory physiotherapy after total knee replacement: Clinical practice recommendations. Ann Readapt Med Phys 2007;50(9):793-801. 80. Medical Advisory Secretariat, Ontario Ministry of Health and Long-Term Care. Physiotherapy rehabilitation after total knee or hip replacement. Health technology literature review. 2005 May [cited 2005 Dec 6]. Available from: URL: http://www.health.gov.on.ca/english/providers/program/mas/tech/ohtas/tech_rehabtkr_061705.ht ml. 81. Soever L, MacKay C. Best practices across the continuum of care for total joint replacement. Greater Toronto Area Rehabilitation Network; 2005 July [cited 2005 Dec 6]. Available from: URL: http://www.gtarehabnetwork.ca/downloads/report-totaljointreplacement-2005.pdf. 82. Jaglal SB, MacKay C, Corrigan L. Rehabilitation for total joint replacement. In: Badley E, Glazier R, editors. Arthritis and related conditions in Ontario. ICES research atlas. 2nd ed. Institute for Clinical Evaluative Sciences; 2004. p. 133,- 146. [cited 2008 Oct 1]. Available from: URL: http://www.ices.on.ca/file/ACF13B.pdf. 83. Merk J, Winkler C, Best R, Horstmann T. Joint function and strength in patients before and after knee replacement surgery. Deutsche Zeitschrift Fur Sportmedizin 2008;59(6):136-40. 84. Ouellet D, Moffet H. Locomotor deficits before and two months after knee arthroplasty. Arthritis Rheum 2002;47(5):484-93. 85. Salaffi F, Carotti M, Stancati A, Grassi W. Health-related quality of life in older adults with symptomatic hip and knee osteoarthritis: A comparison with matched healthy controls. Aging Clin Exp Res 2005;17(4):255-63. 86. Stevens JE, Mizner RL, Snyder-Mackler L. Quadriceps strength and volitional activation before and after total knee arthroplasty for osteoarthritis. J Orthop Res 2003;21(5):775-9. 87. van Dijk GM, Veenhof C, Schellevis F, Hulsmans H, Bakker JPJ, Arwert H, et al. Comorbidity, limitations in activities and pain in patients with osteoarthritis of the hip or knee. BMC Musculoskelet Disord 2008;9:95.    138 Chapter 5: Developing consensus on best practice recommendations for total hip arthroplasty rehabilitation: A Delphi study5 5.1 Background The North American population is becoming older, more sedentary and more overweight. As a result, the number of individuals developing hip and knee osteoarthritis (OA) is on the rise (1-3) and elective total joint arthroplasty (TJA) surgery is in increasingly greater demand (4). More than 22,000 primary total hip arthroplasty (THA) surgeries were performed in Canada in 2006/07 (not including Quebec) (5) and 231,000 in the United States (US) in 2006 (6). By far the majority of THAs are carried out for end-stage OA when conservative treatments have failed (7). While significant attention and resources have been directed toward managing TJA waiting times in Canada and optimizing surgical techniques, prosthetic materials and in-hospital care pathways, little attention has been directed at post-acute rehabilitation and its impact on long-term outcomes. Questions about appropriate rehabilitation interventions, settings, timing and dosage remain unanswered as evidenced by significant rehabilitation practice variation both locally (8) and internationally (9, 10), and by the lack of clinical practice guidelines (CPG) (10). An apparent lack of agreement on routine or standard THA rehabilitation against which new interventions are compared is also evident from the published trials of THA physiotherapy interventions (11). More than 15 years ago, the National Institutes of Health (NIH) in the US reported on the need for an “organized, in-depth study to determine optimum regimen, treatment duration, and expected outcomes as clinical data suggest that potential capabilities of THA  5  A version of this chapter will be submitted for publication. Westby MD, Brittain A, Liang M, Raglin Block M, Backman CL. Best practices for post-acute rehabilitation following primary total hip arthroplasty for osteoarthritis: A Delphi study.   139 patients are not being fully developed” (12)(p.8) and more recently acknowledged that “the use of rehabilitation services is perhaps the most understudied aspect of the peri-operative management of TKA patients” (13)(p.6). To date, little progress has been made towards this end for either type of surgery. It is suggested that current rehabilitation approaches following THA are insufficient as evidenced by the numerous reports of prolonged physical impairment and activity limitations as much as two years post-op (14-17). Our Cochrane systematic review of post-acute physiotherapy interventions following primary THA revealed marked heterogeneity in rehabilitation practices and a lack of available evidence to suggest any one approach is clearly superior to another regarding pain, function and health related quality of life (HRQoL) (11). 5.1.1 Clinical practice guidelines Clinical practice guidelines (CPG) are systematically developed evidence-based statements to assist practitioners and patients make decisions about appropriate health care for specific clinical circumstances (18). Guidelines aim to close the gap between evidence and practice, to reduce variation in clinical practice, and to improve the quality, efficiency and, effectiveness of health care by using the best available scientific evidence and expert opinion to make clinical recommendations (19, 20). Without sufficient high-quality and consistent evidence from rigorously conducted controlled studies, guidelines need to be based to some extent on the opinions of clinicians and others with experience in the field (21). Ideally, such expert opinion and consensus comes from a “…multidisciplinary panel of relevant stakeholders representing a wide range of interests and perspectives related to the topic area” (21)(p.1).  140 5.1.2 Rationale for Delphi survey Historically, many clinical guidelines have been based on expert opinion with little scientific credibility, methodological transparency or formal process for achieving consensus (22). Three main consensus approaches are used in health care: the Delphi group process; the nominal group technique; and, the consensus development conference (21). The latter two approaches involve face-to-face meetings with structured interaction among participants. The Nominal group technique has participants share their ideas in a non-judgemental, round-robin, brainstorming format until all suggestions have been listed (21). This is followed by private voting on the options and statistical methods to derive a group decision. The consensus development conference was introduced by the NIH and as mentioned earlier, was used to examine various aspects of THA surgery (12). With this consensus method, a relatively small group of experts meets over the course of a few days in an open meeting format. Experts and representatives of various interest groups present evidence that is later used by the decision-making group to reach consensus (23). The conference is chaired by an individual who is responsible for controlling the proceedings and encouraging consensus. The Delphi group process, which pre-dates these other approaches, is a formal and established survey method that makes best use of the available research evidence and the collective experiences and knowledge of the participants (24, 24). The method allows for the refinement of responses or opinions through a series of three or more questionnaires (rounds) using controlled feedback and analysis of group members‟ responses (25). Although participants never interact directly, the method is useful to structure and organize group communication (26), especially with a large number of participants from different geographic locations (25). This group process was chosen for its relative ease, efficiency, ability to create a diverse panel and its lower costs when compared to a face-to-face Nominal group process (21, 23, 26, 27). As well, the  141 anonymous nature of the Delphi survey allows for equal and independent contributions from panel members without feeling pressured psychologically by more influential members (23, 27). Delphi surveys have been used previously in guideline development (21, 23, 28) and more specifically in rheumatology (29, 30) and orthopaedics (31). 5.2 Purpose The aim of this study was to incorporate health professional expertise and patient experience with the available research evidence to achieve consensus on best practice recommendations for rehabilitation following THA. 5.3 Methods 5.3.1 Participants Purposive sampling was used to form a diverse expert panel with representation from across Canada and the US, both urban and more rural locations, and varied practice settings. Panelists were identified through word of mouth, letters of introduction sent to professional (Appendix F), consumer (Appendix G) and seniors organizations, internet searches of key rehabilitation institutions, and select members of Canadian and American arthritis health professional organizations. Invited panelists included individuals who had undergone THA surgery (patient experts); orthopaedic surgeons and rehabilitation health care professionals (clinician experts); researchers in the field (research experts); and representatives from key rehabilitation and research institutions (clinician and research experts, decision makers). This diversity allowed for a wider variety of perspectives and creative alternatives and ensured areas of uncertainty and controversy would be fully explored (21, 23, 28).  142 Since no guidance exists on the ideal number of panel members for a Delphi survey, we strove for a balance between greater numbers (which would increase the reliability of group judgment) and logistics, costs, and human resources (21, 32). Recent Delphi surveys in the area of health care have reported between 18 and 44 participants (29-31). We followed-up invitations with phone calls or e-mails for those individuals who did not respond within two weeks. One of the two investigators (MW, AB) confirmed their eligibility and ability to complete all Delphi rounds. This process continued until we had a panel of 40 individuals with the desired diversity of experience and professional representation. 5.3.2 The Delphi questionnaires (rounds) While traditionally round one is used to generate ideas in a brainstorming fashion, Keeney et al. suggest there is support for a revised approach in which pre-existing information is provided for rating  thereby reducing the number of rounds to reach consensus (32). Prior to starting the Delphi survey, panelists were sent a glossary of terms and an overview of our systematic reviews to help focus their attention on the task, bring them to a similar level of knowledge, and underscore that the exercise was based on research, not opinion alone (22, 23, 28, 33, 34). The first round questionnaire was organized in 10 sections, each addressing a component of THA rehabilitation (Appendix H). Each section had one or more key statements on which panelists were asked to rate their level of agreement on a five-point Likert scale (1=strongly disagree, 2=disagree, 3=neutral/no opinion, 4=agree, 5=strongly agree) and provide comments and/or justification for their responses. Related to each key statement were a number of questions (items) regarding specific rehabilitation parameters and contextual factors. The personal and external (environmental) factors were derived from our earlier focus group study (35), systematic review (11) and related literature (36-39). Where possible, wording of personal and external  143 factors followed the World Health Organization‟s (WHO) International Classification of Functioning, Disability and Health (ICF) terminology for consistency with the published literature (40). Panelists were invited to suggest additional items and these were included in the subsequent round, where appropriate. As a result of this process, one additional section was added to rounds two and three. (Refer to Appendix I for example of the round three questionnaire.) We pilot tested the first questionnaire with three patients and 13 health professionals for feedback on its organization, time needed to complete, and clarity (27, 41, 42), which led to minor revisions for implementation of round one. After rounds one and two, panelists were provided with descriptive statistics (mean, standard deviation, range) for group responses, their individual ratings, and anonymous comments reflecting the range of views (25, 33, 34). Panelists were asked to review and reflect on this controlled feedback before responding to the subsequent round. This reflective element encourages panelists to reconsider their views in light of new or more relevant information, an important element in moving towards consensus (23, 25, 26, 32, 34). In order to avoid „response exhaustion‟ among busy, time-pressed participants, we decided a priori to stop the survey after three rounds. The Delphi process took six months to complete and is consistent with other reported Delphi studies requiring three to six months to complete (32). 5.3.3 Procedures The questionnaires were developed and administered using the Arthritis Research Centre of Canada‟s Research Survey System (©2008) (https://dq.arthritisresearch.ca) which enabled panelists to log-on and complete the questionnaires in one or more sittings. Panelists were given a choice in method of survey completion (paper or online version); all chose the online format.  144 To maximize response rates through three rounds, e-mail and/or telephone reminders were sent at weekly intervals over a maximum of six, five and four weeks in rounds one, two and three respectively (32). Personalized thank you cards were mailed to panelists after each round and a $100 honorarium (gift card or donation to registered charity of panelist‟s choice) upon completion of round three. 5.3.4 Ethics approval We received ethical approval to conduct this study from the University of British Columbia Behavioural Research Ethics Board (Appendix C) and the Vancouver Coastal Health Research Institute. An informed consent letter was sent to all panelists and completing the first questionnaire was confirmation of their consent. Participants were assured anonymity during the Delphi rounds and only those who gave permission will be acknowledged by name in publications (Appendix J). 5.3.5 Data collection and analysis As there are a variety of ways to define and determine consensus (21, 26), we asked the panelists to determine a priori the level of agreement that would lead them to believe that the panel has reached consensus and was confident about a given recommendation. Of the 49 panelists who agreed to participate on either the THA and/or TKA panel, 43 (88%) responded and the mean value was 79.4% (mode 80%, range = 60% - 95%); thus 80% was assigned as the level needed to reach consensus for each key statement and individual item. This is consistent with the published literature (27). The cut off point for items to go forward to subsequent rounds was ≥ 50%; however, if a majority of patient panelists felt an item was important despite less than 50% of the full panel agreeing, the item was flagged and included in the subsequent round for further consideration. This novel „patient veto‟ approach was used to ensure that professionals remained  145 patient-centered in their deliberations. Panelists were advised of these procedures prior to the first round. After each round, descriptive statistics (mean, standard deviation (SD), range) and level of consensus (percentage of respondents who selected “agree” or “strongly agree” for a given statement) were calculated for rated items (continuous data) and percentage in agreement for check list items (dichotomous data) consistent with previous Delphi studies (29, 30, 41). The mean represents group agreement while the SD represents the amount of disagreement within the group (42). In round three, panelists were asked to further rate the items on rehabilitation interventions, outcomes and measuring outcomes according to their perceived importance or clinical feasibility (Table 5.1).  146 Table 5.1 Sample controlled feedback  Example  1. Indicate your level of agreement with the following statement: Post-acute rehabilitation should include….    Strongly Disagree   Disagree   Neutral   Agree   Strongly Agree     New questions for Round 2 - Stage of Rehabilitation  1. Indicate your level of agreement with the following statement: When developing best practice recommendations for post-acute rehabilitation after primary THR for OA, it is important to specify or distinguish between an early and a late phase.  Strongly Disagree   Disagree   Neutral   Agree   Strongly Agree             The early phase or the only phase should have patients walking with confidence. Designing another phase could be helpful but optional.  Need to specify what underlying concept you are working from in terms of definition of early and late (e.g. related to soft tissue healing?) Agreement = 80.5% Mean = 3.9 (±1.4) Range Agreement level in text box Arrow points to the Mean (±SD in brackets) Your response  Topics on which consensus was not reached, had the greatest range and SDs, or generated the greatest divergence in comments (outlier views) were further explored through subgroup analyses. Three sets of subgroups were identified a priori based on earlier research suggesting  147 unique perspectives (35): 1) Panelist type (PT, surgeon or patient); 2) Professional or primary role (clinician/surgeon or researcher/academic); and 3) Country (Canada or US). It is well documented that panelists tend to advocate treatments consistent with their specialty or procedures they perform (23) and we were interested in exploring this phenomenon. Group mean scores were compared using one-way ANOVA (SPSS Version 17, SPSS Inc., Chicago, IL). Categorical data were analyzed descriptively by the various subgroup variables. A qualitative analysis of panelists‟ comments was performed after each round and the number of new topics/items recorded (41). Two of the authors (MDW, AB) independently reviewed the narrative data and assigned codes to recurring or important topics and grouped these into categories and key themes (33, 41). Positive (supportive), negative and outlier comments related to each questionnaire topic and theme were provided to panelists (23, 34) along with their individual and pooled ratings approximately 10 days prior to rounds two and three. Panelists were instructed to review and refer to this controlled feedback while completing the next round. 5.4 Results 5.4.1 Panelist demographics Panelists represented a broad range of stakeholders in post-acute rehabilitation after THA in Canada and the US (Table 5.2). All panelists indicated English was their first language. Two of the three patient panelists had post-graduate degrees, all were retired and all had participated in some form of structured rehabilitation for an average of 22 (SD=9) weeks following their THA. The professionals included PTs (n=17), other allied health professionals (n=4), surgeons (n=10), and other physicians (n=4) including rheumatology, physiatry and family practice. The greater proportion of PTs is in line with their major role reported by patients in the focus group study and this thesis‟ focus on physical rehabilitation. Just under half of professionals reported between  148 five and 14 years of experience in THA care, another 17% had 15 to 24 years and 32% had 25 or more years of experience. Roughly one-quarter of professionals saw between one to 99 THA patients each year, just under half saw 100 to 199 and the remainder saw ≥200 patients each year. Table 5.2 THA panelists’ demographics  Clinicians/ Surgeons (n=22) Academics/ Researchers (n=10) Other* (n=3) Patients (n=3) Mean age (SD), years 48 (8) 47 (7) 52 (10) 68 (4) Gender (% female) 55% 40% 33% 67% Canadian residents (%) 68% 60% 33% 67% * Manager (n=2) and educator (n=1)  5.4.2 Response rates Fifty-seven professionals and three patients (n=60) were invited to participate on the THA and/or TKA panel; of these, 40 (67%) agreed to serve on the THA panel and were sent the pre-reading package. One patient panelist dropped out prior to the start of round one for health reasons and was replaced. One physician requested to be removed from the panel prior to the start of round one due to other work commitments. Response rates ranged from 94% to 97% in each round (Table 5.3). Main reasons for not completing a round were time constraints and travel during the study period. Only panelists who completed the previous round were included in the subsequent round.  149 Table 5.3 THA panelists’ response rates by rounds  Round 1 Round 2 Round 3 Invited to participate 60 Agreed to participate 40 Dropped out before start of round 2* 0 0 Sent questionnaire 39 38 36 Completed questionnaire  By deadline (10 days) 22 (56%) 14 (37%) 20 (56%)  After 1 reminder 10 (26%) 11 (29%) 11 (30%)  After 2 or more reminders 6 (15%) 11 (29%) 3 (8%) Total completed round 38 (97%) 36 (95%) 34 (94%) *Patient who dropped out was replaced  5.4.3 Response times Panelists were given 10 days to complete each round; however, only 37% to 56% met this deadline (Table 5.3). In total, it took six, five and four weeks to complete rounds one, two and three respectively. On average, it took panelists 49 to 59 minutes to complete the THA questionnaire in each round. 5.4.4 Round one results Consensus was achieved for 17 of 28 key statements (Table 5.4). Consensus was not reached in statements pertaining to preoperative screening, levels of professional supervision, setting, evaluation of outcomes not captured within the ICF and need for long-term follow-up (FU). These were also the areas in which there were the greatest number of comments and new items/options suggested. Of the 89 items that did not reach the 50% cut off point, 21 were  150 selected by two or more patient panelists and therefore brought forward to round two. Seventy- five new items were suggested in round one.  151 Table 5.4 Level of agreement by Delphi round  Round 1 (n=38) Round 2 (n=36) Round 3 (n=34) Section/Statement Percent agreementa Mean (SD) Range Percent agreement Mean (SD) Range Percent agreement Mean (SD) Range New. Important to recognize an early & late phase    81% 3.9 (1.4) 1-5 94% 4.2 (1.0) 1-5 Important to recognize a maintenance phase    67% 3.9 (1.0) 1-5 62% 3.7 (0.9) 2-5 Individual approach    86% 4.3 (1.1) 1-5 94% 4.5 (0.8) 1-5 A. Patients should be offered structured rehab 82% 4.2 (1.2) 1-5 83% 4.3 (1.1) 1-5 91% 4.4 (0.7) 2-5 Importance of screening pre- operatively to asses need for rehab 68% 3.8 (1.2) 1-5 75% 3.9 (1.0) 1-5 82% 4.2 (0.7) 3-5 Influence of personal factors on need 87% 4.2 (1.0) 1-5 83% 4.3 (0.9) 2-5 94% 4.4 (0.6) 3-5 Influence of external factors on need 69% 3.7 (0.9) 1-5 72% 3.9 (0.8) 1-5 85% 4.1 (0.6) 2-5 B. Rehab should be provided by trained professionals with THA knowledge & experience 92% 4.5 (0.9) 1-5 92% 4.6 (1.0) 1-5 97% 4.6 (0.6) 3-5 Standardized training on THA rehabilitation should be available for health professionals b        88% 4.4 (0.7) 3-5 C. Need for direct supervision by health professional 71% 3.8 (1.1) 2-5 72% 4.0 (1.0) 1-5 76% 3.9 (0.9) 2-5 Indirect/reduced supervision by health professional is appropriate 61% 3.3 (1.1) 1-5 53% 3.4 (1.1) 1-5 53% 3.2 (1.1) 1-5 Self-directed (no supervision) is appropriate 16% 2.1 (1.1) 1-4 20% 2.3 (1.1) 1-5 15% 2.0 (1.1) 1-5 a Percent agreement based on sum of „agree‟ and „strongly agree‟ responses b  New themes/topics added in Round 2 or 3 Note: Bolded numbers indicate consensus ≥80% reached   152 Table 5.4 Level of agreement by Delphi round (continued)  Round 1 Round 2 Round 3 Section/Statement Percent agreement Mean (SD) Range Percent agreement Mean (SD) Range Percent agreement Mean (SD) Range D. Timing of rehab is important for outcomes 82% 4.1 (1.0) 1-5 92% 4.3 (0.8) 1-5 88% 4.3 (0.7) 3-5 Influence of personal factors on timing 50% 3.2 (1.0) 1-5 67% 3.7 (1.1) 2-5 62% 3.6 (0.9) 2-5 Influence of external factors on timing 47% 3.0 (1.1) 1-4 72% 3.8 (1.0) 2-5 62% 3.6 (0.9) 1-5 E. Setting for rehab is important for outcomes 53% 3.3 (1.0) 1-5 53% 3.5 (1.1) 1-5 53% 3.5 (0.9) 1-5 Influence of personal factors on setting 84% 3.9 (0.8) 1-5 83% 4.0 (0.9) 1-5 100% 4.1 (0.3) 4-5 Influence of external factors on setting 87% 3.9 (0.8) 2-5 83% 3.9 (0.9) 1-5 85% 3.8 (0.5) 2-4 F. Appropriate rehab interventions are important for outcomes 84% 4.3 (0.9) 2-5 91% 4.4 (0.8) 1-5 88% 4.3 (0.7) 3-5 G. Dosage of rehab is important for outcomes Not asked (technical error) 77% 3.9 (0.8) 2-5 Influence of personal factors on dosage 90% 4.1 (0.9) 1-5 92% 4.2 (0.9) 1-5 97% 4.2 (0.5) 3-5 Influence of external factors on dosage 84% 3.8 (0.9) 1-5 92% 4.1 (0.7) 1-5 91% 3.9 (0.5) 2-5    153 Table 5.4 Level of agreement by Delphi round (continued)  Round 1 Round 2 Round 3 Section/Statement Percent agreement Mean (SD) Range Percent agreement Mean (SD) Range Percent agreement Mean (SD) Range H. Routinely assessing body structure/function outcomes is important 89% 4.3 (0.7) 2-5 94% 4.3 (0.8) 1-5 94% 4.3 (0.6) 3-5 Routinely assessing activity/ participation outcomes is important 90% 4.2 (0.7) 2-5 92% 4.2 (0.8) 1-5 94% 4.2 (0.6) 2-5 Routinely assessing non-ICF outcomes is important 34% 3.3 (0.8) 2-5 47% 3.4 (1.0) 1-5 53% 3.6 (0.8) 2-5 Influence of personal factors on outcomes 92% 4.2 (0.6) 3-5 95% 4.3 (0.8) 1-5 94% 4.3 (0.7) 2-5 Influence of external factors on outcomes 87% 4.1 (0.7) 2-5 97% 4.3 (0.5) 1-5 94% 4.1 (0.6) 2-5 I. Using appropriate tools to measure body structure/function outcomes is important 90% 4.2 (0.6) 3-5 89% 4.2 (0.8) 2-5 97% 4.3 (0.6) 2-5 Using appropriate tools to measure activity/participation outcomes is important 92% 4.2 (0.7) 2-5 94% 4.3 (0.7) 2-5 94% 4.2 (0.6) 2-5 Using appropriate tools to measure non-ICF outcomes is important 61% 3.6 (0.8) 2-5 56% 3.6 (0.8) 2-5 62% 3.6 (1.2) 1-5 J. Short-term patient follow-up is important 82% 4.0 (1.0) 2-5 92% 4.3 (0.8) 1-5 88% 4.3 (0.8) 2-5 Long-term patient follow-up is important 66% 3.6 (1.2) 1-5 67% 3.9 (1.0) 2-5 79% 4.0 (0.9) 2-5 Access to appropriate follow-up services is important 89% 4.4 (0.8) 2-5 89% 4.4 (1.3) 2-5 94% 4.4 (1.3) 1-5   154 5.4.5 Round two results Consensus was achieved for 19 of 31 key statements – the total of 31 reflects three new key statements that were not included in round one (Table 5.4). The same items noted in round one still did not reach consensus while two of the three new topics related to phases of post-acute rehabilitation reached agreement. Twenty-seven new items/options were suggested in this round. Of the 52 items below the 50% cut off point, 21 were selected by a majority of patient panelists and included in the final round. There were 20% fewer comments in this round compared to round one. 5.4.6 Round 3 results Consensus was achieved on a further three key statements for a total of 22 of 33 key statements – the total of 33 reflects two new statements not included in previous rounds (Table 5.4). One statement approached consensus to recommend against in that more than 79% of panelists disagreed or strongly disagreed that „unsupervised‟ rehabilitation (ie. no direct contact with a rehabilitation provider) was appropriate after THA. Consensus was not reached on the importance of format, setting, and dosage on patient outcomes; outcomes and outcome measures not captured by the ICF; and the importance of long-term follow-up. As well, the influence of personal factors on the timing of post-acute rehabilitation did not achieve consensus. These topics continued to have the greatest range in scores, largest standard deviations and greatest number of comments. Personal and external factors were felt to influence several components of rehabilitation and outcomes (Table 5.5).  155  Table 5.5 Influence of personal and external factors by section Personal factors  Need for rehab Timing Setting Dosage Outcomes General health ●  ● ● ● Body weight     ● Other symptomatic joints     ● Fitness level ●  ●  ● Pain status  ●  ● ● Healing/wound status  ● Post-op complications  ● ●  ● Functional status ● ● ●  ● Psychological status ●    ● Mental/cognitive status   ● ● ● Patient expectations    ● ● Patient goals ● Patient attitude    ● ● Patient engagement    ● ● Patient motivation    ● Physical response to rehabilitation*    ● External factors  Need for rehab Timing Setting Dosage Outcomes Support of spouse/family ● ● ● ● Attitude of physician ● Access/availability of rehab professionals  ● ● ● ● Access to rehab programs     ● Access to transportation   ● ● Health insurance    ● Health professional skills     ● Surgeon skills     ● Contextual factors achieving ≥80% agreement in round 3   156 5.4.7 Results by survey sections 5.4.7.1 Rehabilitation phases This section was added after round one in response to panelist comments. While panelists strongly agreed (94%) that it was important to differentiate between an early and late phase of post-acute rehabilitation, there was no consensus on the duration of each phase. Suggestions ranged from three to 16 weeks with six weeks being selected most often. Late phase suggestions ranged from 12 weeks to eight months with 16 weeks receiving the most frequent endorsement. The need for a maintenance phase did not reach consensus. Panelists strongly agreed (94%) that an individualized approach was needed for post-acute rehabilitation after THA. Subgroup analysis revealed that there was no significant between-group differences for patients, PTs and surgeons regarding the need to recognize an early and late phase (p=.63) and maintenance phase (p=.73). 5.4.7.2 Need for post-acute rehabilitation There was strong support for the need for post-acute rehab (91%) and recognition that personal and external factors influenced this need (Table 5.5). Surgeons rated the need lower than did PTs and patients; however, this difference did not reach statistical significance (p=.10). Pre-operative screening to identify patients in need of post-acute rehabilitation also reached consensus (82%) with no significant differences in responses by panelist type, primary role or country. 5.4.7.3 Rehabilitation providers Panelists strongly agreed (97%) that there was a need for trained health professionals to provide post-acute rehabilitation after THA and this did not differ by primary role of professionals. Physical therapists (94%) and occupational therapists (OTs) (85%) were the two health care professionals viewed as appropriate to provide post-acute rehabilitation. Rehabilitation or PT assistants approached agreement (77%) after three rounds while there remained marked  157 uncertainty amongst panelists regarding the appropriateness of community nurses, kinesiologists, fitness professionals and general practitioners as providers of rehabilitation. Subgroup analysis revealed that there was similar levels of support from PTs, surgeons and patients regarding the appropriateness of PTs, OTs and rehabilitation assistants. No patients or surgeons felt community nurses or family practitioners were appropriate and no PTs or patients felt kinesiologists were appropriate providers of post-acute rehabilitation. Fitness professionals were not selected by any subgroup; however, 35% of panelists remained unsure after three rounds. Panelists agreed (88%) that there was a need for standardized, evidence-based training for health professionals to ensure the knowledge and skills necessary to provide safe and effective THA rehabilitation care; an additional statement that was added to round two based on panelist comments. 5.4.7.4 Rehabilitation format Panelists did not reach consensus (76%) on the need for direct supervision by a trained professional for post-acute rehabilitation after THA. Panelists agreed (87%) that 1:1 patient supervision was appropriate but did not reach consensus on the appropriateness of group treatment with all THA patients (77%) or a mixed format (individual and group treatment) (77%). There was no consensus on the appropriateness of reduced or indirect supervision. Of the 23 panelists who responded favorably to this form of supervision, there was strong support for the use of a rehabilitation or PT assistant (100%) and slightly less support for a trained fitness professional (87%), both under the supervision of a PT. Also rated as appropriate were self- directed rehabilitation/exercise programs with health professional support on request (83%), with a PT consulting during scheduled surgical post-operative visits (96%) and with scheduled periodic checks by a PT (87%). Self-directed rehabilitation with no professional supervision approached negative consensus with 79% of panelists disagreeing with this option. Of the seven  158 who responded favorably, there was unanimous support for self-directed rehabilitation using an illustrated exercise sheet provided by the hospital and equal agreement (86%) on the appropriateness of coaching from a family member who received exercise instruction from the inpatient PT, web-based exercise illustrations or videos, and use of an exercise video at home along with a contact number should questions arise. Subgroup analysis revealed no differences on the appropriateness of various levels of professional supervision when comparing responses of Canadian and American panelists. As well, a similar proportion of American (82%) and Canadian (75%) panelists felt group treatment (all THA patients) was appropriate; although agreement fell on either side of the consensus value of 80%. A majority of panelists (87%) indicated that small groups (1 provider to 3-4 patients) were most appropriate. Panelists agreed (85%) that structured post-acute rehabilitation using a multi-phased approach based on stages of tissue healing and recovery of muscle function was appropriate, regardless of setting and available supervision. 5.4.7.5 Timing of post-acute rehabilitation Consensus was reached (88%) on the importance of timing of post-acute rehabilitation following THA. There were varied levels of agreement on the influence of different personal and external factors on such timing (Table 5.5). Panelists did not reach consensus on the optimal start time for rehabilitation with 61% suggesting less than one week and a further 27% and 12% selecting one to three weeks and three to six weeks respectively. Subgroup analysis identified a significant between group difference (p=.01) with patients rating the importance of timing highest followed by PTs, then surgeons. A similar trend was found for optimal start time with all patients suggesting within one week of surgery compared to 59% of PTs and 43% of surgeons. After collapsing response options, 88% of PTs and 86% of surgeons agreed post-acute rehabilitation should start within three weeks of surgery. Additional subgroup analysis revealed that more  159 Canadians than Americans agreed that external factors influenced the timing of rehabilitation; however, the difference was not statistically significant (p=.07). 5.4.7.6 Rehabilitation setting Consensus was not achieved regarding the importance of rehabilitation setting on optimal outcomes after THA. More than one third (38%) of panelists remained undecided or neutral after three rounds of questions and feedback. Only an outpatient setting was selected as being appropriate with both physical therapy outpatient departments (100%) and private clinics (97%) reaching consensus. Inpatient settings including inpatient departments within the acute care hospital (55%), inpatient rehabilitation facilities (IRFs) (65%) and skilled nursing facilities (SNFs) (55%) did not reach consensus. Home-based rehabilitation approached consensus (77%). A number of personal and external factors were felt to have strong influences on the selection of rehabilitation setting Table 5.5. Subgroup analyses revealed that more patients than PTs and surgeons (p=.09) and more clinicians than researchers (p=.06) rated setting as being important for optimal outcomes after THA however, these differences did not reach statistical significance. There was no difference in ratings when Canadian panelists were compared to Americans. An equal percentage of Canadians and US panelists felt that „health care systems and policies‟ influenced where post- acute rehabilitation was carried out and 20% more Americans agreed that „health insurance policies and coverage‟ affected choice of settings. 5.4.7.7 Rehabilitation interventions Panelists agreed (88%) that appropriate rehabilitation interventions were important for optimal outcomes after THA. Consensus was reached on a number of specific interventions and those  160 selected as either „appropriate and somewhat important‟ or „appropriate and very important‟ by 80% or more of panelists in round three are shown in Table 5.6. Table 5.6 Appropriate and important post-acute rehabilitation interventions after THA Therapeutic and functional exercises  Active ROM  Strength training  Stretching  Postural training  Core stability training  Home exercises  Static balance  Dynamic balance  Stair climbing  Rising/lowering to chair  Rising/lowering to floor  Getting in/out of car  Getting in/out of bathtub  Getting on/off toilet  Dressing  Gait training  Correct use/progression of walking aids  Correction of altered gait pattern  Ensuring proper weight bearing status on operated limb  Cardiovascular training  None reached consensus  Electrical/thermal modalities  None reached consensus  Manual therapy  None reached consensus  Patient education topics  Monitoring for complications  Position/movement restrictions  Return to driving  Sexual activity/safe positioning  Safe use of toilet  Safe use of bath/shower  Return to recreational/sporting activities  Ergonomic/work station set up  Use of assistive devices  Appropriate footwear  Use of medications for pain management  Use of non-medication techniques for pain management  Long-term joint protection  Safe exercise intensity/progression  Note: Bolding indicates ≥80% of panelists rated item as “appropriate and very important”  Several interventions were rated as „appropriate but not important‟ including low to moderate intensity cardiovascular (CV) training, use of CV machines (e.g. stationary bike), pool based exercises, cryotherapy, massage for swelling and scar mobility, passive stretching and proprioceptive neuromuscular facilitation techniques.  161 Subgroup analysis revealed that patients rated the overall importance of appropriate interventions higher than PTs and surgeons (p<.01) and a greater proportion of patients selected low to moderate intensity CV training as being very important (67%) compared to surgeons (43%) and PTs (35%). There was no difference in ratings on the importance of appropriate interventions for optimal patient outcomes when comparing Canadian and American panelists‟ responses (p=.65) 5.4.7.8 Dosage of rehabilitation Consensus was not reached regarding the importance of the overall dosage of post-acute rehabilitation for optimal patient outcomes and there were no differences in ratings by primary role or country. A number of personal and external factors were felt to influence dosage (Table 5.5) With further examination of the details of rehabilitation dosage, there was clearly no agreement within or across types of panelists regarding optimal duration with 41% of PTs selecting four to eight weeks (range <4 weeks to 20 – 24 weeks), a majority of surgeons (57%) selecting four to eight weeks (range 4 – 8 weeks to 12 – 16 weeks) and an equal proportion of patients selecting eight to twelve weeks, 12 to 16 weeks and „other‟ (variable based on individual patient). The total number of treatment sessions was similarly diverse with a majority of PTs (59%) responding that this should be individualized (range 5 – 9 sessions to 36 sessions), an equal number of surgeons selecting five to nine and 10 to 14 sessions (range <5 sessions to 15-19 sessions) and one surgeon suggesting that an individualized prescription was most appropriate. All patients indicated that the number of treatment sessions should be based on individual needs. The preferred frequency of rehabilitation sessions was more consistent with 47% of PTs selecting an individualized approach followed next by two to three times per week (24%). Surgeons felt two to three times per week was most appropriate (43%) and the rest were equally  162 supportive of the other options. Two of three patients agreed two to three times per week was appropriate. 5.4.7.9 Rehabilitation outcomes There was strong agreement on the importance of routinely assessing outcomes conceptualized within the ICF including body structure and function (97%) and activity and participation (94%); however, panelists did not reach consensus (62%) on the need to evaluate outcomes not captured by the ICF (e.g., patient satisfaction). Outcomes that were selected by at least 80% of panelists as being both appropriate and important after THA are shown in Table 5.7. Panelists strongly agreed (94%) that both personal and external factors influenced outcomes after THA (Table 5.5).  163 Table 5.7 Important outcomes to routinely assess and/or monitor after THA Body structure and function outcomes Activity and participation outcomes  Pain (at rest)  Pain (with activity)  Pain coping  Sleep functions  ROM (operated joint)  ROM (other lower limb joints)  Leg length discrepancy (LLD)  Posture and alignment  Gait (pattern, use of aids)  Joint proprioception (position sense)  Muscle strength (operated limb)  Muscle strength (non-operated limb)  Muscle strength (upper limbs)  Muscle recruitment/voluntary activation  Muscle atrophy  Core stability (trunk/pelvic deep muscle control)  Soft tissue flexibility (contractures)  Wound/tissue healing  Energy and vigor  Other outcomes  Health related quality of life  Self-efficacy for exercise  Self-efficacy for rehabilitation  Patient satisfaction with rehabilitation outcomes/process  Patient knowledge (e.g. post-operative complications, precautions)  Patient global assessment (self rating of how he/she is doing)  Health professional/surgeon global assessment (of how patient is doing)  Static balance  Dynamic balance  Walking speed  Walking distance  Stair ascent/descent  Carrying/lifting  Ability to use public transportation  Ability to drive a vehicle  Run errands/shop  Ability to do self care (dressing)  Ability to attend participate in religious activities (pray, kneel)  Ability to do light household activities (cooking, dusting)  Ability to participate in sexual activity  Ability to perform care giving activities (to child or spouse)  Ability to participate in low- moderate intensity leisure/sporting activities  Ability to participate in paid employment  Outcomes with a combined rating of “somewhat” or “very important” reaching ≥80% agreement Bolding indicates that ≥80% of panelists rated item as “very important” Subgroup analysis was conducted to further examine the lack of consensus on the importance of outcomes not captured by the ICF and revealed non-significant differences by panelist type (p=.11) and primary role (p=.79). More patients (100%) and PTs (88%) felt access to  164 rehabilitation services influenced patient outcomes than did surgeons (57%) and a similar trend was seen regarding access to health professionals. A similar proportion of Canadian (91%) and American (83%) panelists agreed that access to rehabilitation services influenced outcomes while a much greater number of Canadian panelists (91%) felt access to health professionals was an issue compared to Americans (67%). More Canadian panelists (81%) also felt that a patient‟s financial situation affected outcomes than did Americans (67%) and a similar percentage from both countries saw „health insurance coverage and policies‟ and „health care systems and policies‟ having an influence. 5.4.7.10 Rehabilitation outcome measurement Panelists strongly agreed on the importance of using appropriate tools or methods to evaluate body structure and function (97%) and activity and participation (94%) outcomes after primary THA. Consensus was not reached on the importance of such tools to evaluate outcomes not captured by the ICF. Those outcome methods and tools selected as being both important and clinically feasible by 80% or more of the panelists are shown in Table 5.8. Subgroup analysis showed that ratings on the importance of measuring outcomes outside of the ICF in a standardized fashion using appropriate tools differed by provider type (p<.01) with PTs rating this statement higher than patients and surgeons. Clinicians and researchers did not differ in their ratings for this statement.  165 Table 5.8 Measures and tools considered feasible and important for routine clinical outcome evaluation and/or monitoring after primary THA Body structure and function measures Activity and participation measures  Pain visual analogue scale (VAS)  Numeric pain rating scale (NPRS)  Standard goniometer to assess passive ROM  Standard goniometer to assess active ROM  Visual observation to assess passive ROM  Visual observation to assess active ROM  Tape measure to assess leg lengths  Visual observation to assess leg lengths  Visual observation to assess lower limb alignment  Standard goniometer to assess lower limb alignment  Visual observation of gait  Trendelenburg test  Patient‟s ability to reproduce target angle (joint position sense)  Skin sensation over operated limb  Manual muscle testing (e.g. Grades 0–5)  Palpation/observation to assess voluntary activation/muscle recruitment  Standardized test positions to assess flexibility/muscle lengths (e.g. Thomas Test for hip flexor length)   Timed walk  Timed Up and Go (TUG)  Single leg static balance test  Repeated stands test (sit-to-stand)  Timed stair ascent/descent  WOMAC OA Index   Other measures  Numeric rating of patient‟s satisfaction with functional outcome  Bolding indicates ≥80% of panelists rated item as “very important”  5.4.7.11 Follow-up care Panelists agreed that it was important to monitor patients on a short-term follow-up (FU) basis (88%) after primary THA and neared consensus regarding long-term FU (79%). There was strong support (94%) for the importance of patients having access to appropriate FU services  166 within the first two years following THA. However, there was less agreement on who should provide this FU care with only surgeons being identified (91%) as being appropriate to routinely provide FU services in the short-term. The only other providers receiving more than 50% support were PTs (65%) and advanced practice PTs (62%). Results differed by subgroups with 94% of PTs and 67% of patients agreeing surgeons were appropriate, 12% of PTs, 57% of surgeons and no patients suggesting physician assistants were appropriate and 65% of PTs, 29% of surgeons and all patients agreeing PTs were appropriate to provide short-term FU services. Advanced practice PTs were viewed as appropriate FU providers by PTs (82%) but received less support from surgeons (29%) and patients (33%). Roughly one-third of patients, PTs and surgeons felt family practitioners were appropriate to provide short-term FU services. Views on appropriate FU providers also differed by country with more Canadian than American panelists selecting surgeons (96% versus 83%) and advanced practice PTs (77% versus 33%). More Americans felt physician assistants (33% versus 23%) were appropriate and similar proportions of roughly two- thirds for PTs and one-third for family practitioners agreed these provider types were appropriate for routine FU care. Consensus was not reached on the appropriate schedule of short-term FU visits which ranged from once at six months to „6 weeks, 3 months, 6 months, 1 and 2 years‟. The schedule selected by the largest number of panelists (21%) was „6 weeks, 3 months, 1 and 2 years‟. There was a similar lack of agreement on the duration of long-term FU which ranged from one year to „life time every 3 years‟. The greatest support (32%) was for „indefinitely‟. One form of short-term FU care and services achieved 80% support: scheduled FU clinic visit (88%). The only other FU services that reached 50% were telephone support from a health professional as needed (74%) and community-based THA exercise programs (50%). Subgroup analysis by type of provider revealed differing levels of agreement on forms of FU care.  167 5.4.8 Thematic analysis of comments Based on the panelists‟ comments in round one, two themes emerged. Panelists emphasized the need to distinguish between an early and later phase of post-acute rehabilitation and commented that their selection of various response options was influenced by this overarching theme. (See Table 5.9 for sample comments.) The second theme speaks to the need for an individualized approach regarding several aspects of post-acute rehabilitation. Panelists commented that rehabilitation providers should remain client-centred and focus on each patient‟s specific needs, goals and response to treatment (Table 5.9). These themes were integrated into the process by creating three new statements for round two. Two misconceptions were evident from round one. Several panelists confused post-acute care with inpatient care and others needed clarification around provision of rehabilitation versus health monitoring. These issues were addressed through Delphi moderator‟s comments in the controlled feedback after round one and the instructions at the beginning of round two. Two additional themes emerged from round two comments. Panelists felt strongly that professionals required standardized training and clinical experience to provide safe and effective rehabilitation services after THA surgery (Table 5.9). Many queried whether any post-entry level training programs in THA assessment and rehabilitation existed in Canada or the US. This theme was addressed through an additional question in round three. The other theme related to the need to acknowledge different subgroups of patients undergoing THA surgery such as the frail elderly and the young active patient (Table 5.9). Panelists were encouraged to comment on such subgroups further in round three. Comments were invited in the final round; however, new items were not specifically requested as there were no further rounds planned.  168 Table 5.9 Sample comments from thematic analysis Theme 1: Need to consider different phases of post-acute rehabilitation  “Different interventions are appropriate for different stages of recovery as muscles gain strength.” “Cardiovascular training should be initiated in the late phase…” “In reality a late phase could last through one‟s lifetime. Patients should be educated on the importance of exercise after focused rehabilitation.”  Theme 2: Need to individualize treatment approaches  “In a sense, all treatment should be individualized but the general practice guidelines need to be constructed based on the generalized patient needs.” “The amount of supervision varies with individuals – hence the need for an individualized approach.” “[Rehabilitation duration] is entirely patient specific, could be less than four weeks right through to 6 months.”  Theme 3: Need for standardized rehabilitation provider training  “Knowledge translation and dissemination of latest research findings to the clinicians involved in front line care should be a priority.” “A background in rehabilitation knowledge and skills for THA patients is essential for providing the best quality of care.” “…some sort of standard accreditation or designation would raise the standard on quality of care by clinicians.”  Theme 4: Recognizing different subgroups of patients  “In order for a patient to be a good candidate for „indirect or reduced supervision‟, the patient must be ideal for this type of rehab… there are a number of low needs, highly motivated, educated patients who would fit this model very well.” “Some people are never going to be able to cross country ski nor run marathons and they should not be used as a standard optimum goal. Younger people with less joint damage will likely out perform older, more damaged patients.”    5.5 Discussion This study is the first to assemble a diverse panel of Canadian and American health professionals, researchers and patients and employ the Delphi method to identify best practices  169 for post-acute rehabilitation following primary THA for OA. Further, it couched the recommendations within the „real world‟ by identifying personal and environmental contextual factors (including institutional and systems-level factors) that can‟t be ignored. Such factors may serve to either facilitate or hinder the ability to apply rehabilitation best practices in this rapidly growing population and in two very different health care systems. Consensus was reached in 22 of 33 potential statements, yet a number of key areas of rehabilitation planning, delivery and outcome evaluation remained uncertain after three rounds of discussion. The trend of growing convergence in opinion and greater number of statements reaching consensus over the three rounds is evidence of the Delphi process at work (34, 41, 42). Critics of the Delphi argue this phenomenon is nothing more than the natural tendency of panelists to conform to the majority view (34, 41, 42), while proponents of this method suggest that the quantitative and qualitative feedback works in a constructive way to help panelists who are out of line with the group opinion to refine their judgments (42). Several THA rehabilitation topics did not achieve consensus nor was a trend towards greater agreement evident after three rounds. Whether additional rounds would have led to consensus on these statements is not known; however, examination of their mean ratings and SDs suggested this was unlikely. Disparate opinions regarding rehabilitation format, setting, dosage and specific treatment interventions are strongly entrenched in local practices/customs and widely reported (9, 10, 35, 43). Subgroup analyses revealed significant between group differences when comparing the ratings of PTs, patients and surgeons in regard to appropriateness of differing levels of supervised rehabilitation and the importance of timing, appropriate interventions, dosage and outcomes not captured by the ICF. No significant differences were found comparing clinicians‟ and  170 researchers‟ and Canadian and American panelists‟ ratings on the key statements. Descriptive analysis of individual items, however, revealed some marked differences by provider type, primary role and country. For example, more Canadian (77%) than American panelists (33%) indicated advanced practice PTs were appropriate to provide short-term FU care. This difference may be partially explained by our differing health care systems, and designations of health professionals as the „advanced practice‟ designation in physical therapy is not formally recognized in the US (A. Guccione, US Department of Veterans Affairs, Health Services Research and Development Service, January 7, 2010). Outcome measurement is another area of THA rehabilitation for which there remained marked uncertainty and dissenting views within the Delphi panel. Further, as much as 36% of panelists were unfamiliar with some of the most commonly reported tools, including the Harris Hip Score and Short Form-12 health status instrument. Riddle et al. suggest that the inconsistency in use of outcome measures makes clinical interpretation of treatment effects problematic (44) and are undertaking a consensus process to identify a core set of outcome measures for use in research trials (45). Our Delphi process has provided a baseline for a similar initiative for use of outcome measures in clinical practice. Our study clearly shows the importance of identifying and acknowledging the large and varied number of personal and external factors that play significant roles in the delivery and outcomes of THA rehabilitation. While several panelists commented that health provider attitudes and health care policies and funding should not influence one‟s need for post-acute rehabilitation, more than three quarters acknowledged these and other contextual factors did indeed influence whether a patient was offered structured rehabilitation. Preoperative screening received mixed support yet was suggested as being useful in identifying those personal and provider-level factors that can be modified or addressed.  171 While a number of therapeutic interventions including ROM exercises, strength and gait training were recommended following THA, treatment details such as key muscle groups to address and specifics of exercise prescription and progression, were not captured by this study. There remains a paucity of high quality research supporting any one form of treatment and optimal treatment dosage in this population (11) thus leaving the clinician to rely more heavily on clinical judgment and individual patient results to guide treatment. While clinician experience and patient values and preferences are important elements of evidence-based practice, there remains a need for increased research on effectiveness of rehabilitation interventions and greater dissemination of already published information on the value of specific approaches. 5.5.1 Strengths and limitations There is no “gold standard” for achieving consensus and developing practice guidelines in the absence of a large body of research evidence (21). Therefore, we used a rigorous method to ensure validity, objectivity and transparency while remaining feasible with regard to overall expense and burden on study participants. Every effort was made to ensure representation from different patients, health care disciplines, practice settings and geographical locations to ensure diverse views and expertise. The decision to include consumer experts is somewhat controversial and the dilemma is likely rooted in the varied definitions of the term „expert‟. While recent Delphi processes in the area of rheumatology have not included patients or consumers in their „expert panels‟ (29-31), we felt it was critical to recognize the lived experience of undergoing a THA as a form of expertise. Consumer involvement in all stages of guideline development is an important element that is often overlooked (46). Further, our novel approach of giving patients „veto power‟ over items that would otherwise be omitted from subsequent rounds ensured that patient-centred care was given utmost prioty and resulted in 21 retained items through round three. We are not aware of other published reports using this approach.  172 For the Delphi method to be successful and its results to be both credible and applicable, the participants must complete all rounds of the survey. Our response rates (94% - 95%) in each of the three rounds exceed (47, 48) or are similar to (49) other Delphi studies in related fields. While response rates of 100% are very rare in Delphi surveys (50), the controlled feedback provided after rounds one and two encourages panel members to become involved and to stay motivated. Active involvement of stakeholders early in the process of guideline development helps participants realize and feel that they are partners in the study, whose judgments will ultimately influence the clinical recommendations and best practice statements for THA rehabilitation. This also leads to the perception of ownership by panel members and greater acceptance of the findings. By providing clear controlled feedback after rounds one and two, along with frequent reminders, thank-you cards, timely technical support and an honorarium, we believe we encouraged panelist engagement and motivation. A number of limitations are also worth noting. The higher number of surgeon drop outs (4 of 5) than other professionals may have led to non-response bias in which those who did not complete all three rounds systematically differed in their views on the various components of post-acute rehabilitation from those who completed the survey. A delay in starting the Delphi survey due to recruitment difficulties and subsequent change to the original time frame may have contributed to panelist drop out. As well, with the majority of panelists serving on both the THA and TKA panel, respondent burn out may have led to some panelists leaving before completing the final round. With the researchers‟ extensive professional affiliations and collaborations, there exists the possibility of response bias in that panelists may have answered questions in the way they thought the researchers wanted them to answer rather than according to their true beliefs. To  173 address this potential form of bias, we pilot tested the first questionnaire to check for leading questions and were careful to provide a range of views through the controlled feedback. Using time stamps, we were able to calculate the length of time panelists took to complete each questionnaire (round) and discovered that three panelists took less than 20 minutes (outside the standard deviation) in each of the three rounds. This threatens the validity of the Delphi results and suggests that a thorough review of questionnaire instructions, reference to the controlled feedback and reflection on one‟s previous ratings likely did not occur. Snap judgments undermine the Delphi method and lead to conformist pressure in succeeding rounds (34). To maintain a manageable size for the Delphi survey, we limited the number of panelists in each stakeholder group while ensuring broad representation. One could argue that involving only three patient experts would not likely reflect the range of rehabilitation experiences and outcomes typical following THA surgery. However, we feel giving patients „veto power‟ over items to that would otherwise be dropped helped to ensure that their views remained at the forefront. It is unlikely that the subgroup analyses had sufficient power to detect statistically significant differences between groups; however, in several cases, such differences were observed and suggest real and important differences in opinion regarding some aspects of post- acute rehabilitation. There were a number of challenges to identification and recruitment of panelists from all parts of Canada and the US, which may bring into question the representativeness of the panel (see Appendix K). With differences across provinces and states in the organization and delivery of health care services, it is likely that some important contextual and practice issues were not taken into account. It is suggested that the only way to test the validity of the recommendations  174 developed by a given panel of experts is to replicate the process to confirm the findings with a different sample of similarly experienced panelists (34, 51). 5.6 Conclusion After a three-round Delphi survey, expert panelists reached consensus on 22 statements related to post-acute rehabilitation after primary THA. Panelists agreed on the need for structured rehabilitation, importance of preoperative screening, use of trained health professionals to provide rehabilitation, importance of timing, appropriate interventions and FU services, and routine use of outcome measurement. The importance of setting and specifics of treatment dosage, format and timing did not reach consensus and continued to generate new ideas and comments through the three rounds. A number of personal and external factors were felt to influence various aspects of the delivery and outcomes of care. Subgroup analysis revealed statistically significant differences in several aspects of rehabilitation delivery and outcome assessment when comparing by panelist type, professional role and country. This supports the importance of multidisciplinary input when creating guideline recommendations. 5.6.1 Clinical implications Our findings suggest that:  Early intervention by trained health professionals following THA is important for optimal outcomes. The need for „trained‟ health professionals means that education to develop, maintain and upgrade skills of practicing health professionals is important. The recommendation (88% agreement) for standardized training on THA rehabilitation may prompt universities and colleges across Canada and the US to standardize the relevant entry level curriculum for medical and allied health professional students.  This may also  175 encourage health regions and major rehabilitation providers to standardize treatment approaches within and between their facilities and organizations. Timing of post-acute rehabilitation is important and panelists recommended it be started within three weeks of THA surgery. This information needs to be conveyed to administrators and policy makers to ensure adequate service delivery post-op (e.g., increased funding to acute care hospitals for more THA surgeries needs to translate into „downstream‟increases in resources allocated for post-acute rehabilitation).  The panel recommends a multi-phased approach to post-acute rehabilitation (based on stages of tissue healing and recovery of muscle function) regardless of the setting or format (e.g., level of supervision, group versus 1 to 1 care). Both health professionals and patients need to understand this important concept as failure to do so may lead to the prescription of inappropriate exercises at the wrong time and less favorable outcomes.  The panel strongly recommends inclusion of the following components in the routine assessment of patients following THA: o Numeric pain rating scale to assess pain at rest and with activity o Manual muscle testing to assess muscle strength o Visual observation of gait o Dynamic balance (technique not specified) o Ability to manage stairs o Ability to perform self-care  And the following therapeutic interventions: o Dynamic balance training o Stair climbing o Rising/lowering to chair  176 o Gait training including use/progression of walking aids, correction of altered gait pattern and ensuring proper weight bearing on operated leg o Patient education (e.g., how to monitor for post-op complications, position/movement restrictions, return to driving) o Home exercises  Although consensus was not achieved with respect to the optimum dosage and format for rehabilitation, the panel agreed that type of intervention (88%) and neared agreement (77%) that overall dosage affect outcome after THA. Finally, the panel unanimously agreed that both personal and external factors influence patient outcomes and that health professionals cannot forget the need to consider individual patient needs and preferences when designing and implementing post-acute THA rehabilitation. 5.6.2 Implications for future research A number of questions arise from this study that could be addressed through future research. Most obviously, the specifics on rehabilitation dosage (duration, frequency, number of treatments) warrant investigation through well designed controlled trials in which one or more of these dosage parameters are altered. Calculating a dose-response relationship would provide clinicians and decision makers with important information on the optimal amount of rehabilitation care following THA and contribute to the establishment of standardized care. Another Delphi survey or smaller face-to-face consensus meeting of highly experienced rehabilitation providers is needed to identify the details of the key therapeutic interventions such as strength, gait and balance training. Additionally, more high quality trials need to be undertaken to determine which interventions have the greatest effects on both the rate of recovery and short- and long-term outcomes after THA coupled with the fewest adverse events. A staged approach to post-acute rehabilitation is suggested yet further work will need to be done  177 to identify the key features of each stage and the most appropriate providers, settings, interventions and outcome measures. The influence of patient expectations, attitudes and engagement in the rehabilitation process on the delivery and outcomes of care would benefit from further exploration so that rehabilitation programs and interventions can be tailored to individual preferences and circumstances. Allowing for greater patient choice in rehabilitation format, level of supervision, setting and follow-up care poses methodological challenges in RCTs; however, better adherence to therapeutic regimes and post-operative outcomes may be achieved. Pre-operative screening to identify those individuals with the greatest need for structured rehabilitation, at risk for poor outcomes or likelihood of delayed recovery is an area for more research. Development of a screening tool and decision-making algorithm may help to ensure the appropriate provider, setting, format and types of rehabilitation and follow-up care are available to those patients demonstrating the greatest need. Further, such a tool would help to identify other subgroups of patients who may respond more or less favourably to certain rehabilitation approaches as suggested in our findings. Finally, future research should assess the final practice guideline recommendations using standardized assessment methods and tools as identified in this study and related research (45) to permit the evaluation and comparison of current and emerging rehabilitation practices on patient outcomes and costs in varied health care settings and contexts in Canada and the US. This study used a rigorous consensus method to develop recommendations on important and clinically feasible multidisciplinary rehabilitation practices and patient outcomes after THA. Ultimately, this work will lead to better service delivery and begin to develop a standard for  178 rehabilitation care for the hundreds of thousands of Canadians and Americans undergoing THA surgery each year.  179 5.7 References 1. Badley EM, Crotty M. An international comparison of the estimated effect of the aging of the population on the major cause of disablement in musculoskeletal disorders. J Rheumatol 1995;22(10):1934-40. 2. Birrel F, Johnell O, Silman A. Projecting the need for hip replacement over the next three decades: Influence of changing demography and threshold for surgery. Ann Rheum Dis 1999;58(9):569-72. 3. Kim S. Changes in surgical loads and economic burden of hip and knee replacements in the US: 1997 - 2004. Arthritis Rheum 2008;59(4):481-8. 4. Tian W, DeJong G, Brown M, Hsieh CH, Zamfirov ZP, Horn SD. Looking upstream: Factors shaping the demand for postacute joint replacement rehabilitation. Arch Phys Med Rehabil 2009;90(8):1260-8. 5. Canadian Institute for Health Information (CIHI). Hip and total knee replacements in Canada - Canadian Joint Replacement Registry (CJRR). 2008-2009 Annual Report. [cited 2009 Sep 17]. Available from: URL: www.cihi.ca/cjrr. 6. American Academy of Orthopedic Surgeons. Facts on hip replacements and facts on knee replacements. [cited 2009 Dec 9]. Available from: URL: http://www.aaos.org/research/stats/patientstats.asp. 7. Katz JN. Total joint replacement in osteoarthritis. Best Pract Res Clin Rheumatol 2006;20(1):145-53. 8. Troughton D, Pearce J. Total hip and knee replacement survey: Rehabilitation practices through the continuum of care. Rehabilitation Services, Vancouver Island Health Authority; 2007. 9. Müller E, Blatt K, Gülich M, Jäckel WH. Treatment schemes and guidelines for rehabilitation after total hip replacement (THR): Results of an international survey and a multicenter study in Germany. Int J Rehabil Res 2007;30(1):86-7. 10. Roos EM. Effectiveness and practice variation of rehabilitation after joint replacement. Curr Opin Rheumatol 2003;15(2):160-2. 11. Westby MD, Kennedy DM, Jones DL, Jones A, Doyle-Waters M, Backman CL. Post-acute physiotherapy for primary total knee arthroplasty: A Cochrane systematic review. Arthritis Rheum 2009;60(10 Suppl):S425. 12. National Institutes of Health. Total hip replacement. NIH Consens State Sci Statement 1994 Sep 12-14 [cited 2004 Apr 20];12(5):1-31. Available from: URL: http://consensus.nih.gov/1994/1994HipReplacement098html.htm.  180 13. National Institutes of Health. NIH Consensus Development Conference on total knee replacement. NIH Consens State Sci Statements 2003 Dec 8-10 [cited 2004 Mar 17];20(1):1-34. Available from: URL: http://consensus.hin.gov/2003/2003TotalKneeReplacement117PDF.pdf. 14. Trudelle-Jackson E, Emerson R, Smith S. Outcomes of total hip arthroplasty: A study of patients one year postsurgery. J Orthop Sports Phys Ther 2002;32(6):260-7. 15. Shih CH, Du YK, Lin YH, Wu CC. Muscular recovery around the hip joint after total hip arthroplasty. Clin Orthop Relat Res 1994;302:115-20. 16. McMeeken JM, Galea MP. Impairment of muscle performance before and following total hip replacement. Int J Ther Rehabil 2007;14(2):55-62. 17. Brander VA, Malhotra S, Jet J, Heinemann AW, Stulberg SD. Outcome of hip and knee arthroplasty in persons aged 80 years and older. Clin Orthop Relat Res 1997(345):67-78. 18. Bigos S, Bowyer O, Braen G, Brown K, Deyo R, Haldeman S, et al. Acute low back problems in adults. Rockville (MD): US Department of Health and Human Services; 1994 Dec.: Publication No. 95-0642. Clinical Practice Guideline No. 14. 19. Liang MH. The American College of Rheumatology Practice Guidelines Strategy. Arthritis Care Res 1995;8(3):132-3. 20. Canadian Physiotherapy Association. Clinical practice guidelines: A discussion paper for the Canadian Physiotherapy Association. 1996 May. 21. Murphy MK, Black NA, Lamping DL, McKee CM, Sanderson CF, Askham J, et al. Consensus development methods, and their use in clinical guideline development. Health Technol Assess 1998;2(3). 22. Raine R, Sanderson C, Hutchings A, Carter S, Larkin K, Black N. An experimental study of determinants of group judgments in clinical guideline development. Lancet 2004;364(9432):429- 37. 23. Black N, Murphy M, Lamping D, McKee M, Sanderson C, Askham J, et al. Consensus development methods: A review of best practice in creating clinical guidelines. J Health Serv Res Policy 1999;4(4):236-48. 24. Adler M, Ziglio E. Gazing into the oracle. The Delphi method and its application to social policy and public health. London (ENG): Jessica Kingsley Publishers; 1996. 25. Jones J, Hunter D. Chapter 5: Using the Delphi and nominal group technique in health services research. In: Qualitative Research in Health Care. London: BMJ Books; 2000. 26. Powell C. The Delphi technique: Myths and realities. J Adv Nurs 2003;41(4):376-82. 27.  Keeney S, Hasson F, McKenna HP. A critical review of the Delphi technique as a research methodology for nursing. Int J Nurs Stud 2001;38(2):195-200.  181 28. Rycroft-Malone J. Formal consensus: The development of a national clinical guideline. Qual Health Care 2001;10(4):238-44. 29. Zhang W, Moskowitz RW, Nuki G, Abramson S, Altman RD, Arden N, et al. OARSI recommendations for the management of hip and knee osteoarthritis, Part II: OARSI evidence- based, expert consensus guidelines. Osteoarthritis Cartilage 2008;16(2):137-62. 30. Roddy E, Zhang W, Doherty M, Arden NK, Barlow J, Birrell F, et al. Evidence-based recommendations for the role of exercise in the management of osteoarthritis of the hip and knee - The MOVE consensus. Rheumatology (Oxford) 2005;44(1):67-73. 31. Enloe LJ, Shields RK, Smith K, Leo K, Miller B. Total hip and knee replacement treatment programs: A report using consensus. J Orthop Sports Phys Ther 1996;23(1):3-11. 32. Keeney S, Hasson F, McKenna H. Consulting the oracle: Ten lessons from using the Delphi technique in nursing research. J Adv Nurs 2006;53(2):205-12. 33. Hasson F, Keeney S, McKenna H. Research guidelines for the Delphi survey technique. J Adv Nurs 2000;32(4):1008-15. 34. Sackman H, The Rand Corporation. Delphi critique. Expert opinion, forecasting and group process. Lexington (MA): Lexington Books; 1975. 35. Westby MD, Backman CL. Patient and health professional views on rehabilitation practices and outcomes following total hip and knee arthroplasty: A focus group study. Submitted 2009 Sep 3. 36. Dreinhöfer K, Stucki G, Ewert T, Huber E, Ebenbichler G, Gutenbrunner C, et al. ICF core sets for osteoarthritis. J Rehabil Med 2004;44 Suppl:75-80. 37. Pisoni C, Giardini A, Majani G, Maini M. International classification of functioning, disability and health (ICF) core sets for osteoarthritis. A useful tool in the follow-up of patients after joint arthroplasty. Eur J Phys Rehabil Med 2008;44(4):377-85. 38. Scheuringer M, Stucki G, Huber EO, Brach M, Schwarzkopf SR, Kostanjsek N, et al. ICF core set for patients with musculoskeletal conditions in early post-acute rehabilitation facilities. Disabil Rehabil 2005;27(7-8):405-10. 39. Weigl M, Cieza A, Andersen C, Kollerits B, Amann. Identification of relevant ICF categories in patients with chronic health conditions: A Delphi exercise. J Rehabil Med 2004;44 Suppl:12- 21. 40. World Health Organization. International classification of functioning, disability and health (ICF). 2001 [cited 2004 Mar 3]. Available from: URL: http://www.who.int/classifications/icf/. 41. Holey EA, Feeley JL, Dixon J, Whittaker VJ. An exploration of the use of simple statistics to measure consensus and stability in Delphi studies. BMC Med Res Methodol 2007 [cited 2008 Nov 13];7:52. Available from: URL: http://www.biomedcentral.com/1471-2288/7/52.  182 42. Greatorex J, Dexter T. An accessible analytical approach for investigating what happens between the rounds of a Delphi study. J Adv Nurs 2000;32(4):1016-24. 43. Naylor J, Harmer A, Fransen M, Crosbie J, Innes L. Status of physiotherapy rehabilitation after total knee replacement in Australia. Physiother Res Int 2006;11(1):35-47. 44. Riddle DL, Stratford PW, Bowman DH. Findings of extensive variation in the types of outcome measures used in hip and knee replacement clinical trials: A systematic review. Arthritis Rheum 2008;59(6):876-83. 45. Riddle DL, Stratford PW, Singh JA, Strand CV. Variation in outcome measures in hip and knee arthroplasty clinical trials: A proposed approach to achieving consensus. J Rheumatol 2009;36(9):2050-9. 46. Van Wersch A, Eccles M. Involvement of consumers in the development of evidence based clinical guidelines: Practical experiences from the North of England evidence based guideline development programme. Qual Health Care 2001;10(1):10-6. 47. Gazi H, Pope JE, Clements P, Medsger TA, Martin RW, Merkel PA, et al. Outcome measurements in scleroderma: Results from a delphi exercise. J Rheumatol 2007;34(3):501-9. 48. Deane KHO, Ellis-Hill C, Dekker K, Davies P, Clarke CE. A Delphi survey of best practice occupational therapy for parkinson's disease in the United Kingdom. Br J Occup Ther 2003;66(6):247-54. 49. Brown AK, O'Connor PJ, Roberts TE, Wakefield RJ, Karim Z, Emery P. Recommendations for musculoskeletal utrasonography by rheumatologists: Setting global standards for best practice by expert consensus. Arthritis Rheum 2005;53(1):83-92. 50. Kennedy DM, Stratford PW, Hanna SE, Wessel J, Gollish JD. Modeling early recovery of physical function following hip and knee arthroplasty. BMC Musculoskelet Disord 2006;7:100. 51. Kennedy HP. Enhancing Delphi research: Methods and results. J Adv Nurs 2004;43(5):504- 11.    183 Chapter 6: Developing consensus on best practice recommendations for total knee arthroplasty rehabilitation: A Delphi study6 6.1 Background The North American population is becoming older, more sedentary and more overweight. As a result, the number of individuals developing hip and knee osteoarthritis (OA) is on the rise (1, 2) and elective total joint arthroplasty (TJA) surgery is in increasingly greater demand (3). More than 35,300 primary total knee arthroplasty (TKA) surgeries were performed in Canada in 2006/07 (not including Quebec) (4) and 0.6 million in the United States (US) in 2007 (5). By far the majority of joint replacements are carried out for end-stage OA when conservative treatments have failed. It is estimated that costs for TKA surgery and the six month-period following average $14,700 per patient thus costing the Canadian health care system more than $557 million per year (4, 6).  This figure is disproportionately higher in the US at more than $17.65 billion annually (based on 2005 data) (7). While significant attention and resources have been directed toward managing TJA waiting times in Canada and optimizing surgical techniques, prosthetic materials and in-hospital care pathways, little attention has been directed at post-acute rehabilitation and its impact on long-term outcomes. Questions about optimal rehabilitation interventions, settings, timing and dosage remain unanswered as evidenced by significant rehabilitation practice variation both locally (8, 9) and internationally, (10-12) as well as by the lack of clinical practice guidelines (CPG) (10, 12). An apparent lack of agreement on routine or standard TKA rehabilitation against which new  6  A version of this chapter will be submitted for publication. Westby MD, Brittain A, Liang M, Raglin Block M, Backman CL. Best practices for post-acute rehabilitation following primary total knee arthroplasty for osteoarthritis: A Delphi study. .  184 interventions are compared is also evident from the published trials of TKA physiotherapy interventions (13). A 2003 NIH consensus conference acknowledged that the use of rehabilitation services is perhaps the most understudied aspect of the peri-operative management of TKA patients (14). To date, little progress has been made towards this end. It is suggested that current rehabilitation approaches following TKA are insufficient as evidenced by the numerous reports of prolonged physical impairment and activity limitations as much as two years post-surgery (15-19). Our Cochrane systematic reviews of post-acute physiotherapy interventions following primary TKA revealed marked heterogeneity in rehabilitation practices and a lack of available evidence to suggest any one approach is clearly superior to another regarding pain, function and health related quality of life (HRQoL) (13). 6.1.1 Clinical practice guidelines and the Delphi method Clinical practice guidelines are based on the best available scientific evidence and expert opinion and inform clinical decision-making in areas of health care for which there are noticeable gaps between evidence and practice and unexplained practice variation (20, 21). With the aforementioned limitations in the evidence base for post-acute physiotherapy after TKA, guidelines will need to be based to some extent on expert opinion and consensus from a diverse panel of relevant stakeholders (22). The Delphi group process is a formal and established survey method that makes best use of the available research evidence and the collective experiences and knowledge of the participants (23) and has been used previously for guideline development (22, 24) in the fields of rheumatology (25, 26) and orthopaedics (27). As described in chapter 5.1.2, the Delphi method involves a series of three or more questionnaires (rounds) and uses quantitative and qualitative feedback to help panelists refine  185 responses and move towards consensus on key discussion points (28). Panelists and their responses remain anonymous through all rounds of the Delphi survey thus encouraging equal contributions and minimizing pressure to conform to the views of more influential members (22, 29).  To date, no formal consensus process has not been undertaken to establish guidelines for post-acute TKA rehabilitation. 6.2 Purpose The aim of this study was to incorporate health professional expertise and patient experience with the available research evidence to achieve consensus on best practice recommendations for rehabilitation following TKA. 6.3 Methods 6.3.1 Participants Invited panelists included individuals who had undergone TKA surgery (patient experts); orthopaedic surgeons and rehabilitation health care professionals from a variety of disciplines (clinician experts); leading researchers in the field (research experts); and representatives from key rehabilitation and research institutions (clinician and research experts, decision makers). Group diversity was again important to ensure a wide range of perspectives was incorporated (22). 6.3.2 Sampling, recruitment and panel composition Using the purposive sampling and panelist recruitment methods described in chapter 5.3.2 and 5.3.3, we assembled a diverse panel of approximately 40 experts from across Canada and the US. As stated previously, only those individuals who intended to complete all Delphi rounds were selected to participate.  186 6.3.3 The Delphi questionnaires (rounds) The identical methods described in chapter 5.3.2 were undertaken to prepare the panelists prior to the first round and reinforce the evidence-based nature of the Delphi exercise (22, 30-32). The first questionnaire was pilot tested with the same group of patients and health professionals and any concerns related to clarity and leading questions were addressed (29, 33, 34). Each of the 10 sections of the first questionnaire addressed different components of TKA rehabilitation (Appendix H). Panelists rated their level of agreement on individual statements using a five-point Likert scale (1=strongly disagree, 2=disagree, 3=neutral/no opinion, 4=agree, 5=strongly agree) and were instructed to provide comments and/or justification for their responses at the end of each section. Specific rehabilitation parameters and contextual factors derived from our earlier research (13, 35) and the related literature (36-39) were included in several sections. One additional section was added to round two based on panelist comments. (See Appendix I for example of round three questionnaire). Descriptive group statistics (mean, standard deviation, range), individual responses and comments were fed back to panelists after rounds one and two (28, 40). Panelists were asked to review and reflect on this controlled feedback before responding to the subsequent round (28, 41, 42). As with the THA survey, this iterative process was completed over the same six-month period and discontinued after three rounds to avoid „response exhaustion‟ among participants. The questionnaires were created and administered using The Arthritis Research Centre of Canada‟s Research Survey System (©2008) (https://dq.arthritisresearch.ca) which enabled panelists to log-on and access both the TKA and THA questionnaires at any time (24/7). The on- line format was selected by all panelists.  187 As described earlier, efforts were made to maximize response rates through three rounds including use of regular reminders, personalized thank you cards and a $100 honorarium (42). 6.3.4 Ethics The University of British Columbia Behavioural Research Ethics Board (Appendix C) and the Vancouver Coastal Health Research Institute provided approval for panelists to participate in one or both Delphi surveys as appropriate. Completion of the first round questionnaire was confirmation of panelists‟ consent. Participants were assured anonymity during the Delphi rounds and only those who gave permission will be acknowledged by name in publications (Appendix L). 6.3.5 Data collection and analysis The details of the data collection, criteria for retaining or eliminating Delphi items for subsequent rounds and the novel „patient veto‟ approach are described in chapter 5.3.5 (24, 41). As per the THA Delphi study, 80% was set as the needed level of agreement to achieve consensus on key statements and individual items. Panelists were advised of these procedures prior to the first round. After each round, the mean, standard deviation (SD), range and level of consensus (percentage of respondents who selected “agree” or “strongly agree” for a given statement) were calculated for rated items (interval data). Percentage of panelists who selected “yes” (e.g., percentage of respondents who selected an item as being important) was determined for dichotomous items (33). The mean represents group agreement while the SD represents the amount of disagreement within the group (34). In round three, panelists were asked to further rate items related to rehabilitation interventions, outcomes and outcome measurement according to their perceived importance or clinical feasibility (Appendix I).  188 We applied the same rationale and approach described in Chapter 5.3.5 to analyze select findings by subgroups established a priori: 1) Panelist type (PT, surgeon or patient); 2) Professional or primary role (clinician/surgeon or researcher/academic); and 3) Country (Canada or US). Group mean scores were compared using one-way ANOVA (SPSS Version 17, SPSS Inc., Chicago, IL). Categorical data were analyzed descriptively by the various subgroup variables. Panelists‟ comments and suggestions for new items were carefully reviewed after each round and the number of new topics/items recorded (33). A thematic analysis was performed to identify recurring and important topics and group them into key themes (40). Positive (supportive), negative (unsupportive) and outlier comments along with the key themes were fedback to panelists approximately 10 days before the next round (22, 32). Panelists were instructed to review and refer to this controlled feedback while completing the next round. New items were added to rounds two and three based on panelists‟ suggestions in the previous round. 6.4 Results 6.4.1 Panelist demographics Panelists represented a broad range of stakeholders in post-acute rehabilitation after TKA in Canada and the US (Table 6.1). All panelists reported English was their first language, they had completed undergraduate degrees or higher and were retired. Patients had all participated in some form of structured rehabilitation, including individual and group treatment, for an average of eight (SD=0) weeks following their TKA. The professionals included PTs (n=21), other allied health professionals (AHPs) (n=4), surgeons (n=9), and other physicians (n=5). As with the THA panel, the greater proportion of PTs was planned. Just over half of professional panelists reported 5 – 14 years of experience in TKA care, another 13% had 15 - 24 years and 28% had 25 or more  189 years of experience. Roughly one third of professionals‟ saw between 1 - 100, 100 – 199, and ≥200 patients each year. Table 6.1 TKA panelists’ demographics   Clinicians/ Surgeons (n=22) Academics/ Researchers (n=14) Other* (n=3) Patients (n=3) Mean age (SD) 47 (9) 46 (8) 52 (10) 71 (8) Gender (% female) 55% 50% 33% 33% Canadian residents (%) 68% 43% 33% 33% Manager (n=2) and educator (n=1)  6.4.2 Response rates Fifty-seven professionals and three patients (n=60) were invited to participate on the TKA and/or THA panel; of these, 47 (78%) agreed to serve on the TKA panel and were sent the pre-reading package. Thirty four of the professionals participated on both panels. One patient was removed from the panel after admitting he did not meet the specific inclusion criteria and was replaced. Two physicians requested to be removed from the panel prior to the start of round one due to other commitments. The response rates ranged from 93% to 95% in each round (Table 6.2). Main reasons for not completing a round were time constraints and travel during the study period. Only panelists who completed the previous round were included in the subsequent round.  190 Table 6.2 TKA panelists’ response rates by rounds  Round 1 Round 2 Round 3 Invited to participate 60 Agreed to participate 47 Dropped out before start of round 2* 0 0 Sent questionnaire 45 42 40 Completed questionnaire  By deadline (10 days) 21 (47%) 17 (40%) 16 (40%)  After 1 reminder 12 (27%) 12 (29%) 17 (43%)  After 2 or more reminders 9 (20%) 11 (26%) 5 (12%) Total completed round 42 (93%) 40 (95%) 38 (95%) *Patient who dropped out was replaced  6.4.3 Response times Panelists were given 10 days to complete each round; however, less than half met this deadline in each of the rounds (Table 6.2). In total, it took six, five and four weeks to complete rounds one, two and three respectively. On average, it tool 50 to 54 minutes to complete the TKA questionnaire in each round. 6.4.4 Round one results Consensus was achieved for 19 of 28 key statements (Table 6.3). Consensus was not reached in statements pertaining to indirect and reduced professional supervision, rehabilitation setting, long-term patient follow-up, and assessment and tools related to outcomes not captured within the ICF framework. These were also the areas in which there were the greatest number of comments and new items/options suggested. Of the 94 items that did not reach the 50% cut off  191 point, 22 were selected by two or more patient panelists and therefore brought forward to round two. Fifty-seven new options were suggested in round one. 6.4.5 Round two results Consensus was achieved for 22 of 32 key statements – this total of 32 reflects four new key statements that were not included in round one (Table 6.3). The same items noted in round one, as well as topics related to pre-operative screening and a rehabilitation maintenance phase still did not reach consensus in round two.  Forty-three new options were suggested in this round and 38 were below the 50% cut off point. The patient veto process resulted in 17 items being carried forward to the final round. There were 32% fewer comments in this round compared to round one.  192 Table 6.3 Level of agreement by Delphi round Section/Statement Round 1 (n=42) Round 2 (n=40) Round 3 (n=38)  Percent agreement Mean (SD) Range Percent agreementa Mean (SD) Range Percent agreement Mean (SD) Range New. Important to recognize an early & late phase b    100% 4.5 (0.5) 4-5 97% 4.5 (0.6) 3-5 Important to recognize a maintenance phase b    73% 4.0 (0.8) 3-5 76% 4.0 (0.8) 2-5 Need to acknowledge individual needs b    95% 4.6 (0.6) 3-5 97% 4.6 (0.6) 3-5 A. Patients should be offered structured rehabilitation 95% 4.6 (0.9) 1-5 98% 4.6 (0.6) 3-5 95% 4.57 (0.7) 2-5 Importance of screening pre- operatively to assess need for rehabilitation 81% 4.1 (1.0) 2-5 73% 3.9 (0.9) 2-5 76% 3.9 (0.9) 2-5 Influence of personal factors on need 88% 4.3 (0.9) 2-5 95% 4.4 (0.6) 3-5 97% 4.3 (0.5) 3-5 Influence of external factors on need 83% 4.0 (0.9) 2-5 88% 4.2 (0.6) 3-5 90% 4.0 (0.8) 2-5 B. Rehabilitation should be provided by professionals with TKA knowledge & experience 100% 4.7 (0.5) 4-5 100% 4.8 (0.4) 4-5 97% 4.6 (0.6) 2-5 Standardized training on TKA rehabilitation should be available for health professionalsb       95% 4.5 (0.7) 2-5 a Percent agreement based on combined number of „agree‟ and „strongly agree‟ responses b New theme/topic added in Round 2 or 3 Note: Bolded numbers indicate 80% consensus reached   193 Table 6.3 Level of agreement by Delphi round (continued) Section/Statement Round 1 Round 2 Round 3  Percent agreement Mean (SD) Range Percent agreement Mean (SD) Range Percent agreement Mean (SD) Range C. Need for direct supervision by health professional 86% 4.3 (0.9) 2-5 88% 4.3 (0.8) 2-5 87% 4.3 (0.8) 2-5 Indirect/reduced supervision by health professional is appropriate 43% 2.9 (1.2) 1-5 43% 3.0 (1.2) 1-5 39% 2.8 (1.1) 1-5 Self-directed (no supervision) is appropriate 10% 1.7 (1.0) 1-5 15% 1.7 (0.9) 1-4 8% 1.8 (0.9) 1-4 D. Timing of rehabilitation is important for outcomes 100% 4.7 (0.5) 4-5 100% 4.6 (0.5) 4-5 97% 4.5 (0.6) 3-5 Influence of personal factors on timing 52% 3.4 (1.3) 1-5 63% 3.5 (1.3) 1-5 74% 3.8 (1.1) 2-5 Influence of external factors on timing 64% 3.4 (1.2) 1-5 65% 3.4 (1.2) 1-5 74% 3.8 (0.9) 2-5 E. Setting for rehabilitation is important for outcomes 69% 3.8 (1.1) 2-5 70% 3.8 (1.0) 2-5 79% 3.9 (0.8) 2-5 Influence of personal factors on setting 81% 4.0 (0.7) 2-5 85% 4.0 (0.6) 2-5 95% 4.1 (0.5) 2-5 Influence of external factors on setting 86% 4.0 (0.9) 1-5 95% 4.2 (0.5) 3-5 90% 4.0 (0.6) 2-5 F. Appropriate rehabilitation interventions are important for outcomes 95% 4.7 (0.6) 2-5 98% 4.5 (0.8) 1-5 92% 4.5 (0.8) 1-5 G. Dosage of rehabilitation is important for outcomes Not asked (technical error) 88% 4.2 (0.7) 2-5 84% 4.1 (0.7) 2-5 Influence of personal factors on dosage 95% 4.3 (0.6) 2-5 98% 4.3 (0.5) 3-5 92% 4.2 (0.8) 1-5 Influence of external factors on dosage 79% 3.9 (0.9) 2-5 90% 4.1 (0.6) 2-5 95% 4.1 (0.7) 2-5   194 Table 6.3 Level of agreement by Delphi round continued  Round 1 Round 2 Round 3 Section/Statement Percent agreement Mean (SD) Range Percent agreement Mean (SD) Range Percent agreement Mean (SD) Range H. Routinely assessing body structure/function outcomes is important 98% 4.5 (0.6) 3-5 100% 4.4 (0.5) 4-5 95% 4.3 (0.7) 2-5 Routinely assessing activity/participation outcomes is important 98% 4.5 (0.6) 3-5 98% 4.3 (0.5) 3-5 97% 4.3 (0.6) 2-5 Routinely assessing other outcomes is important 62% 3.7 (0.9) 1-5 60% 3.6 (0.8) 2-5 79% 3.8 (0.6) 2-5 Influence of personal factors on outcomes 95% 4.3 (0.6) 3-5 100% 4.4 (0.5) 4-5 100% 4.3 (0.5) 4-5 Influence of external factors on outcomes 86% 4.0 (0.9) 1-5 98% 4.1 (0.5) 2-5 97% 4.1 (0.5) 2-5 I. Using appropriate tools to measure body structure/function outcomes is important 98% 4.4 (0.6) 2-5 100% 4.4 (0.5) 4-5 97% 4.3 (0.6) 2-5 Using appropriate tools to measure activity/participation outcomes is important 98% 4.3 (0.6) 2-5 98% 4.3 (0.6) 2-5 97% 4.3 (0.6) 2-5 Using appropriate tools to measure other outcomes is important 57% 3.7 (0.8) 2-5 68% 3.8 (0.7) 2-5 74% 3.8 (0.6) 2-5 J. Short-term patient follow-up is important 93% 4.4 (0.7) 2-5 90% 4.3 (0.8) 1-5 95% 4.3 (0.6) 2-5 Long-term patient follow-up is important 71% 3.9 (1.1) 2-5 78% 3.9 (0.9) 2-5 84% 4.0 (0.8) 1-5 Access to appropriate follow-up services is important 95% 4.4 (0.7) 2-5 95% 4.3 (0.7) 2-5 97% 4.5 (0.6) 3-5   195 6.4.6 Round three results Consensus was reached on a further two statements for a total of 24 key statements – the total of 33 reflects one new statement not included in previous rounds (Table 6.3). One item reached „negative consensus‟ in that more than 80% of panelists disagreed or strongly disagreed that „unsupervised‟ rehabilitation (i.e. no direct contact with a rehabilitation provider) was appropriate after TKA. Consensus was not reached on the importance of setting on patient outcomes, the specifics of dosage (frequency, duration, number of visits), format (including level of supervision) and optimal timing (start of rehabilitation). Additionally, there was no agreement on the importance of outcomes and outcome measurement not captured within the ICF framework. Personal and external factors were felt to influence several components of rehabilitation and outcomes (Table 6.4). Statements not reaching consensus continued to have the greatest range of scores, largest standard deviations and greatest numbers of comments.  196  Table 6.4 Influence of personal and external factors by section Personal factors  Need for rehab Timing Setting Dosage Outcomes General health ●  ● ● ● Other symptomatic joints     ● Pain status  ●  ● ● Healing/wound status  ● Post-op complications  ●  ● Functional status ● ● ● ● ● Fitness/physical activity level ●  ●  ● Psychological status     ● Mental/cognitive status ● ● ●  ● Patient expectations ●   ● ● Patient attitude     ● Patient engagement    ● ● Physical response to rehabilitation    ● Patient adherence     ● External factors  Need for rehab Timing Setting Dosage Outcomes Support of spouse/family ●  ● ● ● Access to rehabilitation professionals  ● ● ● ● Waiting list for rehabilitation services  ● Access to rehabilitation programs     ● Access to transportation  ● ●  ● Health insurance policies/coverage   ● Health care system/policies  ●  ● Health professional skills     ● Surgeon skills     ● Contextual factors achieving ≥80% agreement in round 3   197 6.4.7 Results by sections 6.4.7.1 Rehabilitation phases This section was added after round one in response to panelists‟ comments. While panelists strongly agreed (97%) that it was important to differentiate between an early and late phase of post-acute rehabilitation, there was no consensus on the duration of each phase. Suggestions for early phase ranged from three to 16 weeks with six weeks being selected most often. Late phase suggestions ranged from 12 weeks to12 months with both extremes of range receiving the most frequent endorsement. Subgroup analysis revealed that patients, PTs and surgeons provided similar ratings on the need to acknowledge early and late phases. There was no consensus on the need for a maintenance phase with surgeons rating this lower than that of patients and PTs (p=.07). Panelists strongly agreed (97%) on the need for an individualized approach to post-acute rehabilitation. 6.4.7.2 Need for post-acute rehabilitation There was strong support for the need for post-acute rehabilitation (95%) and this was consistent with subgroup analysis by panelist type and country. Pre-operative screening to identify patients in need of post-acute rehabilitation was viewed as important; however, it did not reach consensus (76%) and did not differ significantly by panelist type, primary role or country. Panelists recognized that personal and external factors influenced the need for rehabilitation (Table 6.4). 6.4.7.3 Rehabilitation providers There was strong support (97%) for trained health professionals providing post-acute rehabilitation after TKA; however, this differed by subgroup with more clinicians than researchers supporting this recommendation (p=.04). Physical therapists were the only health professionals identified as being appropriate to provide post-acute rehabilitation. Advanced practice PTs (70%), rehabilitation or PT assistants (70%), and occupational therapists (OTs)  198 (65%) were the other providers achieving more than 50% support. After three rounds, there remained marked uncertainty amongst panelists regarding the appropriateness of nurse practitioners, kinesiologists and advanced practice OTs as providers of rehabilitation. Panelists agreed (95%) that there was a need for standardized, evidence-based training for health professionals to ensure the knowledge and skills necessary to provide safe and effective TKA rehabilitation care; an additional statement that was added to round two based on panelist comments. Subgroup analysis by panelist type revealed variable support for different types of rehabilitation providers with all surgeons and patients agreeing advanced practice PTs were appropriate more so than PTs themselves (67%); this trend held true for OTs and advanced practice OTs. Fewer patients (33%) and PTs (24%) felt nurse practitioners were appropriate providers of rehabilitation than did surgeons (60%). All patients (n=3) agreed that rehabilitation assistants or PT assistants were appropriate rehabilitation providers compared to 67% of PTs and 60% of surgeons. Similarly, all patients felt family physicians were appropriate to provide TKA rehabilitation compared to 14% of PTs and 20% of surgeons. There were no subgroup differences for kinesiologists with one patient, two PTs and one surgeon agreeing these exercise professionals were appropriate to provide post-acute TKA rehabilitation. 6.4.7.4 Rehabilitation format Panelists agreed (87%) on the importance of supervised rehabilitation after TKA; however, support for specific forms of supervision and formats for rehabilitation was inconsistent. Panelists agreed (87%) that 1:1 patient supervision was an appropriate format for rehabilitation care; however, failed to reach consensus on whether group treatment with all TKA patients (70%) or a mixed format (individual and group treatment) (70%) was appropriate. There was no consensus on the appropriateness of reduced or indirect supervision. Of those panelists who  199 responded favourably, use of trained fitness professionals under PT supervision (83%) and self- directed home exercises with periodic clinic-based checks with a PT (100%) were felt to be appropriate. Self-directed rehabilitation with no professional supervision achieved negative consensus with 82% of panelists agreeing this was not appropriate. Of the seven panelists who responded favourably, self-directed exercise with internet-based video instruction (86%) and use of an iPod with downloaded exercises (86%) were suggested. Surgeons rated the need for supervised rehabilitation as slightly less important and reduced supervision as more appropriate for optimal patient outcomes than did patients and PTs; however, differences were not statistically significant (p=.33, p=.14). Subgroup analysis further revealed a similar number of Canadian (71%) and American (65%) panelists felt group treatment (all TKA) was appropriate. Group size was also discussed and a majority of panelists (70%) indicated that small groups (1 provider to 3-4 patients) were most appropriate. Panelists agreed (90%) that structured post-acute rehabilitation using a multi-phased approach based on stages of tissue healing and recovery of muscle function was appropriate, regardless of setting and available supervision. 6.4.7.5 Timing of rehabilitation There was strong agreement (97%) on the importance of timing of post-acute rehabilitation following TKA. Consensus was not reached for the overall influence of personal and external factors on the timing of rehabilitation yet a number of specific factors were identified as being important (Table 6.4). Half of PTs (48%) and surgeons (50%) and two-thirds of patients felt post-acute rehabilitation should start within 72 hours of acute hospital discharge. By far the majority of PTs (95%) and all patients agreed treatment should be initiated within 1 week compared to 67 % of surgeons. Ratings did not differ by country.  200 6.4.7.6 Rehabilitation setting The importance of setting on optimal outcomes after TKA approached consensus (79%) after three rounds with 16% of panelists remaining undecided or neutral regarding this matter. There was unanimous agreement on the appropriateness of hospital outpatient clinics or departments for TKA rehabilitation and slightly less support for private physical therapy clinics (94%) and outpatient orthopaedic centers (89%). Rehabilitation in inpatient settings including inpatient departments within the acute care hospital (61%), inpatient rehabilitation facilities (IRFs) (67%) and skilled nursing facilities (SNFs) (53%) did not reach consensus. Home-based rehabilitation approached consensus (78%). Personal and external factors were felt to have strong influences on the selection of rehabilitation setting (Table 6.4). Subgroup analyses revealed that more patients than PTs and surgeons (p=.03) rated setting as being important for optimal outcomes after TKA. More clinicians than researchers felt the rehabilitation setting was important; however, the difference was not statistically significant (p=.11). There was no difference in ratings when Canadian panelists were compared to Americans. 6.4.7.7 Rehabilitation interventions Panelists strongly agreed (92%) that appropriate rehabilitation interventions were important for optimal outcomes after TKA and further consensus was reached on a number of specific interventions. Those selected as either „appropriate and somewhat important‟ or „appropriate and very important‟ by 80% or more of panelists in the final round are shown in Table 6.5.  201 Table 6.5 Appropriate and important post-acute rehabilitation interventions after TKA Therapeutic and functional exercises  Active ROM  Active assisted ROM  Passive ROM  Strength training  Stretching  Postural training  Home exercises  Neuromuscular re-education  Static balance  Dynamic balance  Stair climbing  Rising/lowering to chair  Rising/lowering to floor  Transfers  Gait training  Correct use/progression of walking aids  Correction of altered gait pattern  Indoor/outdoor training  Variable surface training  Cardiovascular training  Low-moderate intensity training  Use of appropriate CV machines (e.g. stationary bike, treadmill)  Electrical/thermal modalities  Ice  Manual therapy  Massage for swelling  Massage for scar mobility  Passive stretching techniques  Proprioceptive neuromuscular facilitation (PNF)  Joint mobilizations (e.g. glides)  Patient education topics  Monitoring for complications  Position/movement restrictions  Return to driving  Sexual activity/safe positioning  Safe use of toilet  Safe use of bath/shower  Return to recreational/sporting activities  Ergonomic/work station set up  Use of assistive devices  Appropriate footwear  Use of medications for pain management  Use of non-medication techniques for pain management  Long-term joint protection  Safe exercise intensity/progression  Interventions with a combined rating of “somewhat” or “very important” reaching ≥80% agreement; bolding indicates ≥80% of panelists rated item as “very important”  Subgroup analyses revealed that patients rated the overall importance of appropriate rehabilitation interventions on TKA outcomes lower than did PTs and surgeons (p=.02). There was no difference in ratings when comparing Canadian and American panelists‟ responses (p=.19). One physical therapy intervention of interest and controversy was the use of neuromuscular electrical stimulation (NMES) after TKA. A comparison of response frequencies showed that 76% of PTs compared to only 33% of surgeons and 50% of patients agreed that  202 NMES was somewhat or very important after TKA while there was little difference in this rating when clinicians (64%) were compared to researchers (60%). 6.4.7.8 Dosage of rehabilitation Panelists agreed (84%) that the overall dosage of rehabilitation following TKA was important in achieving optimal outcomes; however, there was strong concordance that both personal and environmental factors influenced the amount of rehabilitation care patients receive. Specific contextual factors are shown in Table 6.4. After three rounds, there was no consensus on the optimal duration (greatest support for 4-8 weeks (40%) and 8-12 weeks (40%), range <4 weeks to 20-24 weeks), frequency (greatest support for 2-3 times/week (42%)) or number of treatment sessions (greatest support for 15-19 sessions (29%), range 5-9 to 24-36 sessions) and no apparent trend towards agreement was emerging. Surgeons rated the importance of rehabilitation dosage lower than patients and PTs (p=.02) while there was no statistically significant difference when Canadian and American ratings were compared. By examining the details of rehabilitation dosage, it was evident that there was no agreement within each subgroup regarding duration with 48% of PTs selecting four to eight weeks (range <4 weeks to 20-24 weeks), an equal number of surgeons (33%) selecting four to eight weeks and eight to 12 weeks (range 4-8 weeks to 12-16 weeks) and most patients (67%) indicating eight to 12 weeks (range 4-8 weeks to 8-12 weeks) as the optimal length of a post- acute rehabilitation program. The overall number of treatment sessions was equally diverse with one quarter of PTs selecting 15 to 19 sessions (range 5-9 sessions to 24-36 sessions), half of surgeons selecting 10 to 14 sessions (range 10-14 sessions to 15-19 sessions) and two thirds of the patients choosing 15 to 19 sessions. Frequency of rehabilitation sessions was more consistent among groups with 43% of PTs, 67% of surgeons and 67% of patients indicating two to three times a week as optimal for TKA rehabilitation. Several panelists within each subgroup felt that  203 the elements that comprise the overall dosage of rehabilitation (e.g., frequency, duration) should be variable or individualized taking into account each patient‟s specific needs and goals. For this reason, one third of PTs did not specify an optimal duration, number or frequency of rehabilitation sessions. 6.4.7.9 Rehabilitation outcomes There was strong agreement on the importance of routinely assessing outcomes conceptualized within the ICF including body structure and function (95%) and activity and participation (97%) domains; however, panelists were just shy of consensus (79%) for outcomes not captured by the ICF such as health related quality of life (HRQoL) and patient satisfaction. Outcomes that were selected by at least 80% of panelists as being both appropriate and important after TKA are shown in Table 6.6. The panel unanimously agreed that personal factors influence patient outcomes and was of a similar view (97%) for external factors. Specific contextual factors are shown in Table 6.4.  204 Table 6.6 Important outcomes to routinely assess and/or monitor after TKA Body structure and function outcomes Activity and participation outcomes  Pain (at rest)  Pain (with activity)  Pain coping  Knee effusion  Lower limb edema  Sleep functions  ROM (operated joint)  ROM (other lower limb joints)  Posture and alignment  Gait (pattern, use of aids)  Joint (ligamentous) stability  Joint proprioception (position sense)  Coordination  Muscle strength (operated limb)  Muscle strength (non-operated limb)  Muscle strength (upper limbs)  Muscle recruitment/voluntary activation  Muscle atrophy  Core stability (trunk/pelvic deep muscle control)  Soft tissue flexibility (contractures)  Wound/tissue healing  Energy and vigor  Emotional functioning (stress, coping)  Other outcomes  Self-efficacy for exercise  Self-efficacy for rehabilitation  Patient satisfaction with rehabilitation outcomes/process  Patient knowledge (e.g. post- operative complications, precautions)  Patient global assessment (self rating of how he/she is doing)  Health professional/surgeon global assessment (of how patient is doing)  Static balance  Dynamic balance  Walking speed  Walking distance  Stair ascent/descent  Carrying/lifting  Ability to use public transportation  Ability to drive a vehicle  Run errands/shop  Ability to do self care (dressing)  Ability to attend social functions  Ability to attend participate in religious activities (pray, kneel)  Ability to travel (air travel, bus tour)  Ability to do light household activities (cooking, dusting)  Ability to do moderate/heavy household activities (laundry, vacuuming)  Ability to do light outdoor work  Ability to do moderate/heavy outdoor work (rake leaves, shovel snow)  Ability to participate in sexual activity  Ability to perform care giving activities (to child or spouse)  Ability to participate in low- moderate intensity leisure/sporting activities  Ability to participate in paid employment  Ability to participate in unpaid/volunteer employment Outcomes with a combined rating of “somewhat” or “very important” reaching ≥80% agreement; bolding indicates ≥80% of panelists rated item as “very important”   205 Subgroup analysis was conducted to further examine the lack of concordance on outcomes not captured within the ICF framework and revealed a significant difference (p=.02) among provider types with patients rating this category of outcomes highest and surgeons the lowest. Health related quality of life, for example, was rated „very important‟ by 48% of PTs, 33% of patients and no surgeons. Clinicians and researchers were in agreement on the importance of evaluating outcomes not captured by the ICF and similar percentages (27% versus 30%) rated HRQoL as being „very important‟. Further analyses demonstrated between group differences regarding the influence of external factors on outcomes with more Canadian than American panelists selecting access to health professionals (95% versus 77%), and access to rehabilitation services (95% versus 88%) as issues. Although not reaching consensus, patients‟ financial situation was selected by more American panelists (77% versus 65%) while health insurance and health policy were similarly rated. 6.4.7.10 Rehabilitation outcome measurement Panelists reached consensus on the importance of using appropriate tools in a consistent manner to evaluate or monitor body structure and function (97%) and activity and participation (97%) outcomes; however, again did not achieve consensus on „other‟ tools (74%). Those outcome methods and tools selected as being both important and clinically feasible by 80% or more of the panelists are shown in Table 6.7. Subgroup analysis showed that ratings on the importance of measuring outcomes outside of the ICF in a standardized fashion did not differ by provider type or primary role.  206 Table 6.7 Measures and tools considered feasible and important for routine clinical outcome evaluation and/or monitoring after primary TKA Body structure and function measures Activity and participation measures  Pain visual analogue scale (VAS)  Numeric pain rating scale (NPRS)  Standard goniometer to assess passive ROM  Standard goniometer to assess active ROM  Visual observation to asses lower limb alignment  Standard goniometer to assess lower limb alignment  Visual observation of gait  Trendelenburg test  Ligamentous stress testing (for joint stability)  Patient‟s ability to reproduce target angle (joint position sense)  Skin sensation over operated limb  Manual muscle testing (e.g. Grades 0–5)  Palpation/observation to assess voluntary activation/muscle recruitment  Standardized test positions to assess flexibility/muscle lengths (e.g. Thomas Test for hip flexor length)  Visual observation of patellar alignment and tracking  Timed walk  Timed Up and Go (TUG)  Single leg static balance test  Functional Reach Test for balance  Repeated stands test (sit-to-stand)  Timed stair climbing  6-minute walk test  WOMAC OA Index  Lower Extremity Functional Scale (LEFS)  Performance battery (e.g. timed walk, stair climb and TUG)  Other measures  Numeric rating of patient‟s satisfaction with functional outcome  Numeric rating of patient‟s satisfaction with rehabilitation process  Self-efficacy for self-management Bolding indicates ≥80% of panelists rated item as “very important”  6.4.7.11 Follow-up care Panelists agreed strongly that it was important to monitor patients on both a short-term (95%) and to a lesser extent, long-term (84%) follow-up (FU) basis after primary TKA. Further, there was strong support (97%) for the importance of patients having access to appropriate follow-up services to address their needs in the initial two year-period after TKA; however, marked differences in opinion surfaced when panelists were asked to suggest who should provide such  207 care. Orthopaedic surgeons were the only health professionals reaching consensus (92%) with all surgeons and patients in support and 91% of PTs. Physician assistants (PAs), nurse practitioners, PTs, advanced practice PTs and dieticians were under consideration in round three but did not reach consensus when considering the total panel‟s views. Subgroup analysis showed that surgeons rated PAs and nurse practitioners higher than did PTs and patients. Conversely, PTs rated their fellow PTs and advanced practice PTs higher than did surgeons and both designations had 100% support from patients on the panel. Views on appropriate providers of short-term FU care differed by country with more American panelists selecting PAs (59% versus 38%) and nurse practitioners (47% versus 24%) as appropriate than Canadians. Slightly fewer American panelists felt surgeons (88% versus 95%) and markedly less agreed advanced practice PTs (53% versus 91%) were appropriate to provide FU care. Support for PT follow-up care did not differ by country with approximately two-thirds of both subgroups selecting this provider. Consensus was not reached on the appropriate schedule of short-term FU visits which ranged from FU at „2 months and 1 year post-op‟ to „6 weeks, 6 months, 1 year and 2 years post-op‟. The schedule selected by the largest number of panelists (51%) was „6 weeks, 3 months, 1 year and 2 years post-op‟. Similarly, when asked to select the appropriate time frame and schedule for long-term FU there was no agreement and the greatest support (27%) was given to „indefinitely‟. Forms of short-term FU care and services that reached consensus were telephone support from a health professional as needed (87%) and scheduled FU clinic visit with the surgeon (82%). Support for the different forms of FU care differed by panelist type with more PTs (95%) than surgeons (50%) and patients (67%) agreeing telephone support from a health professional was appropriate. Scheduled FU clinic visit with surgeon and exercise booster sessions with a PT  208 received more support from patients and PTs than from surgeons. Surgeons unanimously agreed with scheduled FU clinic visits with radiographic evaluation compared to two-thirds of PTs and patients. Community-based TKA exercise programs were deemed appropriate by 57%, 50% and 33% of PTs, surgeons and patients respectively. 6.4.8 Thematic analysis of comments Based on panelists‟ comments in round one, two themes emerged. The need to differentiate stages of post-acute rehabilitation was emphasized across panelist groups and influenced how they responded to items in several sections of the questionnaire. (See Table 6.8 for sample comments.) A second theme reinforces the need for an individualized approach to some aspects of post-acute rehabilitation. Panelists felt it was important to remember that „best practice‟ guidelines are not a substitute for clinical judgment and individual patient needs and preferences (Table 6.8). These themes were integrated in the process by creating new statements for rating in round two. A recurring topic related to several panelists‟ misconceptions on delivery of post- acute rehabilitation care (e.g., post-acute rehabilitation meant inpatient care) and was addressed by providing clarification in the controlled feedback after round one and the instructions at the beginning of round two. We identified one additional theme from round two comments. Need for standardized rehabilitation provider training captures panelists‟ views that all professionals providing rehabilitation services after TKA surgery should have appropriate training and clinical experience (Table 6.8). It was acknowledged that no standardized, post-entry level training program currently exists in Canada and the US. This theme was addressed by the adding a related statement to the third questionnaire. Comments were invited in round three; however, new items were not specifically requested as there were no further rounds planned. Thematic analysis revealed no new concepts.  209 Table 6.8 Sample comments from thematic analysis Theme 1: Need to consider different phases of post-acute rehabilitation  “Early and late phases should be defined by milestones relating to body and function outcomes, not time related.” “Maintenance to me indicates no further improvement or gains, which in my experience with TKA patients rarely occurs within the first year post op.” “While I think that maintenance and health/wellness are important, I don‟t believe that long-term outcome of a TKR will be affected by a maintenance program.” “Clearly data support the continued need for and ability to achieve improvement 12 months and beyond.” “Setting will vary depending on early and late post-acute rehab.”  Theme 2: Need to individualize treatment approaches  “Dosage should be individualized based on patient status, progress and outcomes.” “Best [practice] guidelines apply to the majority of patients but individual needs must always be considered and may indicate intensity, duration, timing etc. of rehab interventions.” “Having a choice of how to structure rehab may allow therapists to allocate less resources to those patients who don‟t need it and give more resources to those who do. This is more complex and less „one size fits all‟ and requires good judgment from the therapist.”  Theme 3: Need for standardized rehabilitation provider training  Re: Follow-up care: “Need special training in evaluation and management of joint replacement patients” “…quality training experiences are rare.” “[Providers] should have clinical experience that is comprehensive with sound knowledge of ortho-biomechanics, gait abnormalities, looking beyond the knee to factors that may contribute to potential aberrant problems that may affect the short and long-term effects following TKA.”    6.5 Discussion This study is the first to assemble a diverse panel of Canadian and American health professionals, researchers and patients and employ the Delphi method to identify best practices for post-acute rehabilitation following primary TKA for OA. Further, it couched the  210 recommendations within the „real world‟ by identifying personal and environmental contextual factors (including institutional and systems-level factors) that can‟t be ignored. Such factors may serve to either facilitate or hinder the ability to apply rehabilitation best practices in this rapidly growing population and in two very different health care systems. While panelists were not instructed to differentiate whether these factors were facilitators or barriers to optimal patient outcomes, most are intuitive such as access to rehabilitation professionals, rehabilitation programs and transportation. Consensus was reached in 24 of 33 potential statements; however, similar to our THA Delphi findings, a number of key areas of rehabilitation planning, delivery and outcome evaluation remained uncertain after three rounds of discussion. Subgroup analyses revealed significant between group differences when comparing the ratings of PTs, patients and surgeons in regard to appropriateness of unsupervised rehabilitation and the importance of setting, rehabilitation and FU provider-type, specific interventions, treatment dosage, and outcomes not captured by the ICF. Further differences were found in the perceived need for trained professionals in providing TKA rehabilitation when clinicians‟ views were compared to researchers‟, and in the appropriateness of reduced or indirect health professional supervision comparing Canadian and American responses. Some of these latter differences again may be explained by our differing health care systems, and designations of health professionals. For example, only 38% of Canadian panelists felt that PAs were appropriate to provide FU care compared to 59% of American panelists. In Canada, there is one accredited PA training program compared to 149 such programs in the US (I. Jones, Canadian Association of Physician Assistants, December 23, 2009). Similar to the THA study, we saw a trend towards convergence in opinion and a greater number of statements reaching consensus over the three rounds - evidence of the Delphi process at work  211 (33, 34). As mentioned in the previous chapter, however, the Delphi method is not without its critics (32). Several TKA rehabilitation topics including rehabilitation setting, dosage and specific treatment interventions did not achieve consensus and may have benefited from further discussion through another Delphi round. Yet evidence of similar lack of agreement and resultant inconsistencies in rehabilitation practices is widely reported (11, 12, 35, 43) and unlikely to change without significant efforts at all levels of health care. Again, similar to our THA Delphi findings, outcome measurement in TKA rehabilitation remains problematic. No activity and participation measures were identified as „feasible and should do‟ and only numeric ratings of patient satisfaction and self-efficacy were selected as appropriate and feasible despite several outcomes not captured within the ICF recognized as being important across stakeholders. Inconsistent and underuse of outcome measures eliminates valuable clinical information that would assist in clinical decision making. To address the inconsistent use of outcome tools in research (44), Riddle et al. are undertaking a consensus process to identify a core set of outcome measures for TKA trials (45). The importance of identifying and acknowledging the numerous personal and external factors that play significant roles in the delivery and outcomes of TKA rehabilitation was evident from our findings. As suggested earlier, preoperative screening would be valuable for identifying personal and provider-level factors that could be modified or addressed to better prepare the patient for TKA surgery and flag those individuals at risk for delayed or incomplete recovery. A number of therapeutic interventions including active ROM, strength training and dynamic were strongly recommended following TKA, but treatment specifics were not part of this study and still need to be teased out. With the limitations in the current evidence and no one form of  212 physical therapy and treatment dosage clearly superior to another (13), it poses a challenge to the clinician wanting to provide evidence-based care. Instead greater emphasis must be placed on clinical judgment and individual patient preferences to guide treatment. While the latter are important elements of evidence-based practice, further research needs to be done on the effects of rehabilitation interventions and greater efforts made to disseminate and integrate available evidence on specific approaches. For example, there is a growing body of research supporting the use of NMES after TKA yet less than 60% of panelists rated it as appropriate and two-thirds of surgeons felt it was neither appropriate nor important. 6.5.1 Strengths and limitations of research Strengths and weakness of the Delphi approach and its application to this study are discussed in chapter 5.5.1. Our rigorous methodology and efforts to ensure anonymous and broad representation from key stakeholders strengthens the overall validity and generalizability of our findings. Inclusion of patient experts in the design and implementation of this study helped to ensure it remained client centred (46). In addition, our novel „patient veto‟ approach resulted in 17 items being retained for further discussion in the final round – a technique that we have not found described elsewhere. Response rates of 93% to 95% per round are considered excellent and exceed those of other Delphi studies in related health care fields (42, 47, 48). Varied strategies to engage and retain panelists were successful and through these efforts, may have led to their appreciation that they were partners in the study, whose judgments will ultimately influence the clinical recommendations and best practice statements for TKA rehabilitation. A number of limitations are worth noting and have been addressed in chapter 5.5.1. The higher number of surgeon drop outs (four of seven drop outs) than other professionals may have led to  213 non-response bias in which those who did not complete all three rounds systematically differed in their views on the various components of post-acute rehabilitation from those who completed the survey. The delayed Delphi start date and possible response exhaustion for those professionals serving on both panels may have contributed to this drop out. As noted earlier, the authors‟ extensive professional affiliations increased the risk of response bias and therefore we pilot tested the first questionnaire to check for leading questions and provided a range of perspectives in the controlled feedback. Three panelists took less than 20 minutes to complete each questionnaire (outside the standard deviation) suggesting snap judgements and threatening the validity of the results (32). It was a challenge to form a diverse and representative panel of experts while maintaining a manageable size. While we were only able to include three patient experts, we feel their „veto power‟ over items not otherwise carried forward helped to ensure that their views remained at the forefront. An additional challenge to forming a representative panel was identifying and recruiting experts from all parts of Canada and the US (see Appendix K). Some important health care delivery issues may not have been taken into account as a result of not having panelists from all regions of Canada and the US. Again, the validity of the expert panels‟ recommendations can only be confirmed by replicating the process with a new yet similarly experienced panel (32, 49). 6.6 Conclusion After a three-round Delphi survey, expert panelists reached consensus on 24 statements related to post-acute rehabilitation after primary TKA. There was strong agreement on the need for structured rehabilitation, use of trained health professionals to provide rehabilitation, importance of timing, appropriate interventions and follow-up services, and routine use of outcome measurement.  214 The importance of setting and specifics of treatment dosage, format and timing did not reach consensus and continued to generate new ideas and comments through the three rounds. A number of personal and external factors were felt to influence various aspects of the delivery and outcomes of care. Subgroup analysis revealed statistically significant differences in several aspects of rehabilitation delivery and outcome assessment when comparing panelist type, professional role and country. This supports the importance of multidisciplinary input when creating guideline recommendations. 6.6.1 Clinical implications Our findings suggest that:  Early intervention by trained health professionals following TKA is important for optimal outcomes. The need for „trained‟ health professionals means that education to develop, maintain and upgrade skills of practicing health professionals is important. The recommendation (95% agreement) for standardized training on TKA rehabilitation may prompt universities and colleges across Canada and the US to standardize the relevant entry level curriculum for medical and allied health professional students.  There was strong agreement (97%) on the importance of early intervention (within 1 week) following TKA. This information needs to be conveyed to administrators and policy makers to ensure adequate service delivery post-op (e.g., increased funding to acute care hospitals for more TKA surgeries needs to translate into increased resources allocated for post-acute rehabilitation).  The panel recommends that a multi-phased approach (based on stages of tissue healing and recovery of muscle function) be adopted when providing intervention, regardless of  215 the setting or format (e.g., level of supervision) to avoid inappropriate exercise prescription and less favorable outcomes.  The panel strongly recommends inclusion of the following components in the routine assessment of patients following TKA: o Numeric pain rating scale to assess pain at rest and with activity o Goniometry to measure active and passive ROM o Manual muscle testing to assess muscle strength o Soft tissue flexibility (technique not specified) o Visual observation of gait o Dynamic balance (technique not specified) o Ability to manage stairs o Ability to perform self-care  And the following therapeutic interventions: o Active ROM o Strengthening exercises o Dynamic balance training o Stair climbing o Rising/lowering to chair o Gait training including use/progression of walking aids, correction of altered gait o Patient education (e.g., how to monitor for post-op complications) o Home exercises  Although consensus was not achieved with respect to the optimum dosage and format for rehabilitation, the panel agreed that dosage (84%) and type of intervention (92%) affect outcome after TKA.  216 Finally, the panel unanimously agreed that both personal and external factors influence patient outcomes and individual patient needs and preferences need to take precedence in the planning and delivery of post-acute rehabilitation. 6.6.2 Implications for future research Recommendations for future research are similar to those outlined in chapter 5.6.3. Specifics of rehabilitation including timing, dosage, format and interventions require further investigation before consensus is reached on these key elements of rehabilitation after TKA. Another Delphi survey or smaller face-to-face consensus meeting is needed to identify details of the key therapeutic interventions such as strength, gait and balance training. Additionally, more high quality trials are warranted to determine which interventions have the greatest effects on both the rate of recovery and short- and long-term outcomes after TKA. Identification of essential elements of key treatment parameters will advance the field of TKA rehabilitation research and contribute to improved and more consistent standards of care. Empirical data will need to be coupled with our findings and related literature that shows the substantial influence of patient expectations, attitudes and engagement in the rehabilitation process on the delivery and outcomes of care. As noted previously, attention to these factors may lead to better adherence to therapeutic regimes and more favourable outcomes. Research on pre-operative screening and use of decision-making tools may further our understanding of risk factors contributing to delayed recovery or poor outcomes after TKA. Lastly, the final practice guideline recommendations will need to be rigorously evaluated in varied rehabilitation settings and using standardized assessment tools identified in this study and related research (45) to enable comparison of their effects and costs against current and emerging rehabilitation practices in Canada and the US.  217 6.7 References 1. Badley EM, Crotty M. An international comparison of the estimated effect of the aging of the population on the major cause of disablement in musculoskeletal disorders. J Rheumatol 1995;22(10):1934-40. 2. Birrel F, Johnell O, Silman A. Projecting the need for hip replacement over the next three decades: Influence of changing demography and threshold for surgery. Ann Rheum Dis 1999;58(9):569-72. 3. Tian W, DeJong G, Brown M, Hsieh CH, Zamfirov ZP, Horn SD. Looking upstream: Factors shaping the demand for postacute joint replacement rehabilitation. Arch Phys Med Rehabil 2009;90(8):1260-8. 4. Canadian Institute for Health Information (CIHI). Hip and total knee replacements in Canada - Canadian Joint Replacement Registry (CJRR). 2008-2009 Annual Report. [cited 2009 Sep 17]. Available from: URL: www.cihi.ca/cjrr. 5. Agency for Healthcare Research and Quality. Facts and figures 2007. Healthcare Cost and Utilization Project (HCUP). [cited 15 Jan 2010]. Available from: URL: http://www.hcup- us.ahrq.gov/reports/factsandfigu