Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Identifying disinvestment options to increase the impact of priority setting in health care organizations Dionne, Francois 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_2011_spring_dionne_francois.pdf [ 221.11kB ]
Metadata
JSON: 24-1.0071461.json
JSON-LD: 24-1.0071461-ld.json
RDF/XML (Pretty): 24-1.0071461-rdf.xml
RDF/JSON: 24-1.0071461-rdf.json
Turtle: 24-1.0071461-turtle.txt
N-Triples: 24-1.0071461-rdf-ntriples.txt
Original Record: 24-1.0071461-source.json
Full Text
24-1.0071461-fulltext.txt
Citation
24-1.0071461.ris

Full Text

IDENTIFYING DISINVESTMENT OPTIONS TO INCREASE THE IMPACT OF PRIORITY SETTING IN HEALTH CARE ORGANIZATIONS by FRANCOIS DIONNE B.Bus. Adm. Université de Montréal, 1979 M.Sc. Bus. Adm. University of British Columbia, 1981 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Health Care and Epidemiology) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) November 2010 © Francois Dionne, 2010 ii ABSTRACT Choices about what to fund and what not to fund are necessary in health care because claims on resources always exceed those available. Moreover, the choices faced by decision-makers are often between or amongst a wide range of difficult to compare programs or interventions.  It is no surprise, then, that processes that inform those choices are of considerable interest.  Yet, we know that existing priority setting processes have found limited practical use and when used, are rarely used to their full potential. The objective of the current research was to produce new knowledge that would facilitate the use of formal priority setting processes in decision-making on resource allocation in health care. Based on a detailed review of the literature, a decision was made to focus on one particular aspect of priority setting in health care that has long been recognized as a significant barrier to the successful implementation of priority setting processes: the identification of disinvestment options. Building on initial exploratory research, a proposed procedural change to the Program Budgeting and Marginal Analysis (PBMA) priority setting process was designed to address challenges in identifying disinvestment options. The proposed procedural change was then evaluated in a case study as part of a real-world priority setting exercise. The key finding of this research project was that adding a step -- that focused on the determination and communication of acceptable service reductions, at the outset of process implementation -- to the standard PBMA process, has the potential to assist in ‘disarming’ organizational incentives that have been found to work against the identification of disinvestment options.  This key finding is of critical importance because without practical disinvestment options, priority setting processes are likely to have limited impact on decision-making and therefore limited practical appeal. Further, without formal, structured priority setting processes that actually work in practice, resource allocation decisions will continue to follow historical patterns, leading to iii incremental growth without explicit consideration of return on investment. As such, this study makes a novel contribution to the literature in an area that is highly relevant to the everyday challenges faced by health care decision-makers. iv PREFACE Three published articles resulted from the research work for this dissertation.  Dionne, F., Mitton, C., Smith, N. & Donaldson, C. (2008). Decision maker views on priority setting in the Vancouver Island Health Authority. Cost Effectiveness and Resource Allocation, 6, 13. The proportion of research and writing conducted by the student was 80%. The student did all the interviews on which the study is based, did the thematic analysis and most of the writing. Co-authors contributed to the design and the writing of the paper. This part of the dissertation research work was approved by the UBC Behavioural Research Ethics Board, certificate number H05-90931.  Dionne, F., Mitton, C., Smith, N. & Donaldson, C. (2009). Evaluation of the impact of program budgeting and marginal analysis in Vancouver Island Health Authority. J Health Serv Res Policy, 14:234-242. The proportion of research and writing conducted by the student was 80%. The student did all the interviews on which the study is based, did the thematic analysis and most of the writing. Co-authors contributed to the design and the writing of the paper. This part of the dissertation research work was approved by the UBC Behavioural Research Ethics Board, under the same certificate number as above, H05-90931.  Dionne F., Mitton C., Shoveller J., Peacock S. & Barer M. (2009).  How to control the costs of health care services- an inventory of strategic options. Healthc Manage Forum. 2009 Winter; 22(4):23-30. The proportion of research and writing conducted by the student was 90%. The student conducted the literature search and the abstract review, selected the papers to be retained, decided on the writing plan and did most of the writing. Co- authors contributed to the design and the writing of the paper. This part of the dissertation research work was approved by the UBC Behavioural Research Ethics Board, certificate number H08-03126. vThe research work at Menno Place was approved by the UBC Behavioural Research Ethics Board, certificate number H07-02521 vi TABLE OF CONTENTS Abstract ............................................................................................................................... ii Preface ............................................................................................................................... iv Table of contents................................................................................................................ vi List of tables.........................................................................................................................x List of figures..................................................................................................................... xi Acknowledgements........................................................................................................... xii Chapter One: Introduction ...................................................................................................1            Background...............................................................................................................1            Priority setting overview...........................................................................................1            Evolution of priority setting......................................................................................3            Formal priority-setting processes..............................................................................9            Why the focus on disinvestments in health care? ...................................................15                Definition of disinvestments ..............................................................................15                Why are disinvestments important?...................................................................17             Why the Program Budgeting and Marginal Analysis (PBMA) process? ..............18             Relevance...............................................................................................................19             Study purposes and objectives ...............................................................................20             Overview of dissertation ........................................................................................20  vii Chapter Two: How to control the costs of health care services- An inventory of strategic options .........................................................................................22                Introduction.........................................................................................................22                Methods...............................................................................................................22                Results.................................................................................................................24                             Overview................................................................................................24                             Strategies to reduce or contain health care costs ...................................25                             Classification framework .......................................................................36                             Impact of strategies on services .............................................................37                             Strategies classified by impact on services............................................37                             Strategic options for decision-makers....................................................41               Discussion ............................................................................................................44                              Key messages........................................................................................44                              Potential limitations ..............................................................................44               Conclusion ...........................................................................................................45 Chapter Three:  Exploratory research on the topic of disinvestments in the implementation of the Program Budgeting and Marginal Analysis priority setting process............................................................................46                Introduction.........................................................................................................46                Research questions..............................................................................................47                Methods...............................................................................................................48                             Context and settings...............................................................................48                             Data collection and analysis...................................................................51                Results.................................................................................................................52 Observed results in terms of changes to the resource allocation pattern and challenges in the identification of disinvestment options ....52 Key barriers or facilitators to the identification of disinvestment options.....................................................................................................56 viii Organizational context factors that emerged as potential determinants of the challenges to the identification of disinvestment options.....................................................................................................61                 Discussion ...........................................................................................................64                Conclusion ..........................................................................................................69 Chapter Four: Implementation of a modified version of Program Budgeting and Marginal Analysis at St. Joseph’s Health Centre, Toronto........................71                Introduction.........................................................................................................71                Research question ...............................................................................................73                Methods...............................................................................................................73                             Context and setting ................................................................................73                             PBMA implementation at St. Joseph’s ..................................................74                             Case study research design ....................................................................77                             Data collection and analysis...................................................................77                  Results...............................................................................................................78                             Disinvestment options and resource allocation decisions......................79                             Challenges in the identification of disinvestment options .....................81                Discussion ...........................................................................................................86                Conclusion ..........................................................................................................89 Chapter Five: Conclusion ..................................................................................................90                Why an interest in priority setting and, specifically, in disinvestments .............90                Key finding – what is the main contribution to knowledge................................91                How does this key finding fit with what we knew?............................................94                What are the limitations of this research project?...............................................95                Where do we go from here?................................................................................96 ix Bibliography ......................................................................................................................99 Appendix A  The seven steps of the PBMA process .......................................................115 Appendix B  VIHA’s organizational structure ................................................................116 Appendix C Examples of approved disinvestments at Menno Place .............................118 Appendix D Interview schedule for evaluation interviews at St. Joseph’s Health Centre..........................................................................................................121 xLIST OF TABLES TABLE 1   Steps of PBMA process- 1997 ........................................................................10 TABLE 2   Steps of PBMA process- 2004/6.....................................................................11 TABLE 3 Summary of cost reduction or containment strategies ....................................25 TABLE 4   Potential impact on the make up of services of each strategy ........................41 TABLE 5   Steps of PBMA process at St. Joseph’s ..........................................................75 TABLE 6   Organizational structure of PBMA at St. Joseph’s .........................................76 TABLE 7   Ratings of disinvestment proposals at St. Joseph’s.........................................80 xi LIST OF FIGURES FIGURE 1  Framework for success factors analysis .........................................................14 FIGURE 2  Strategic options available to decision-makers ..............................................43 xii ACKNOWLEDGEMENTS I want to thank the members of my supervisory group, Dr. Craig Mitton, Dr. Morris Barer, Dr. Stuart Peacock and Dr. Jeannie Shoveller. They have been supportive throughout the entire thesis project, consistently showed patience and provided invaluable guidance. I want to especially thank Dr. Craig Mitton, the thesis supervisor, without whom this thesis would never have been completed. He was always there to answer questions and, just as importantly, to provide moral support at the key frustrating moments. 1CHAPTER 1 INTRODUCTION Background Health care priority setting is about making decisions regarding how scarce resources will be used. Over a decade ago, Ham (1997) remarked that “priority setting in health care is not new but it is an issue of growing importance”. This issue continues to be relevant, as noted by Sabik & Lie (2008): “priority setting or rationing in health care continues to be a politically charged topic, but recently its necessity has gained wider recognition”. However, it is also the case that priority setting in health care can be viewed as “a complex task” (Hauck, Smith & Goddard, 2004). Priority setting has been a subject of health services research for more than thirty years, yet it has been widely acknowledged that resource allocation decisions are still made mostly on the basis of historical allocation patterns, adjusted for population growth and political influences (Mitton & Donaldson, 2004b; Bate, Donaldson & Murtagh, 2007). This means that patterns of health care spending change little over time despite changing technology, knowledge and diseases.  Changes are driven primarily by political considerations, keeping up with inflation, and new programming.  If greater use of formal priority setting processes can be achieved, the health care system could become more responsive to its changing environment without relying as much on programmed incrementalism. The main objective of this research is to produce new knowledge that would facilitate the use of formal, structured priority setting processes in real life decision-making about resource allocation in health care. Priority setting overview Choices on what to fund and what not to fund are necessary in health care because claims on resources typically exceed those available. As Wildavsky (1977) explained more than 30 years ago, people will always desire more medical services than any system 2is financially able to provide because of the uncertainty often present around illness (with respect to both diagnosis and treatment). In other words, while most people would prefer to avoid health care interventions, they still wish to receive treatment that may address the condition that may cause their symptoms1. This explanation for never-ending cost escalation in health care still holds and was recently used to support the statement that “[in health care] no amount of money is ever enough” (Evans 2007). When there is not enough money to meet all the claims, choices must be made. This is the very essence of priority setting. The magnitude of the challenges faced in making choices between competing claims for resources depends on: 1) the pressure to limit cost escalation (i.e. the pressure to limit the growth in budgets and to stay within set envelopes), and 2) the growth in claims on resources which is related to a myriad of factors including the development of new technologies that are desired by the population because they may help with, or prevent, some condition. Looking forward, the combination of the worldwide economic crisis that erupted in 2008 and its impact on government balance sheets, and the continued development of new technologies,2 points to an intensifying spotlight on priority setting in health care and, as a result, to many contentious decisions. In this, the key question that needs to be asked is, ‘how significant to health system performance are priority setting decisions?’ In the opinion of some experts, the quality of decisions regarding resource allocation is so important that it can put at risk the viability of health care programs. Rachlis (2005), for example, suggests the fate of Medicare in Canada depends on the system’s ability to improve services with the resources available: “To save Medicare, we must tackle its problems head on. We must roll up our sleeves and restructure our health care system to improve the quality it provides” (p. 25). Evans (2007) is even more specific in stating that the future of Canadian health care depends on  1 This does not imply that cost escalation is demand-driven but that, because of the uncertainty often present around illness, patients and the public tend to respond positively to any offer of service that may potentially help. 2 For example, the body charged with assessing new technologies in Australia has been described as being “perpetually stretched with an almost overwhelming stream of submissions for new and emerging technologies” (Elshaug, Hiller & Moss., 2008). 3its ability to focus on, “the considerable inefficiencies that lurk in the organization and delivery of care” (p.25).3 The reasoning behind these pronouncements is that the performance of health care programs is measured, at least in part4, by comparing costs in relation to benefits achieved (and not just measuring benefits). Therefore, activities such as priority setting that aim to improve the return on investment have a direct and significant role to play in determining health system performance. Evolution of priority setting Whenever resource allocation decisions are made without a marketplace, some mechanism emerges, formally or informally, to provide guidance to decision-makers. Historically, a number of mechanisms have been employed in health care to this end. For our purposes, two mechanisms will be addressed in brief -- needs assessment and economic evaluation. Needs assessment is an approach that has been used since the 19th century (Wright, Williams & Wilkinson, 1998), and became the seminal approach in the 1970s (Gugushvili, 2007). Needs assessment addresses resource allocation questions by guiding the decision-maker towards health care interventions for which there is the greatest ‘need’.5 The outcome of a needs assessment resource allocation process is a matching of interventions or programs to needs. As described by Stevens & Gillam (1998), “The essence of needs assessment is an understanding of what is effective and for whom”. Central to this process is of course the definition of health care ‘needs’. However, as reaching consensus on a definition of need is very difficult (Gugushvili, 2007, p 3), needs assessment was refined over time with the introduction of the concept of ‘capacity to benefit’. With this modification, resources would go to interventions addressing those  3 Of course, for those who are opposed to Medicare on principle, no improvement in efficiency will ever be sufficient to justify the limitations imposed on individual purchasing of services. 4 As with any form of public spending, politics undeniably play a role in the evaluation of performance. 5 A variation on the needs assessment approach is the definition of ‘core services’, with those services then becoming the top priority in resource allocation (Norheim, Ekeberg, Evensen, Halvorsen & Kvernebo, 2001). 4needs with the greatest capacity to benefit, circumventing what had often become a paralyzing problem (Stevens & Gillam, 1998). In some versions of needs assessment, the definition of ‘what is effective’ includes a specific reference to cost-effectiveness, where the assessment of benefits combines epidemiological factors and cost measurements (Stevens & Gillam, 1998). However, because cost-effectiveness is not the basis of the decision-making process with needs assessment, its role can often become secondary (Bernfort, 2003). This means that the linkage between resource use and impact (or benefits) is not directly established. Instead, resource allocation is guided primarily by the fit between services and needs, with the allocation pattern being ‘improved’ when resources are removed from services that provide no or low benefit to patients and reallocated to services that provide greater benefits (as measured by the ability to address needs). In this way, for both individual practices and health care managers, needs assessment can provide information to guide decisions on how to use resources more efficiently6 and effectively in order to improve population health (Wright et al., 1998). However, the lack of an explicit link between needs (i.e. ability to benefit) and costs (i.e. resource use) in needs assessment, as well as the lack of a mechanism to guide choices between options that offer different levels of benefits and different costs, has been identified by many as a fundamental limitation of needs assessment as a priority setting tool. In fact, when the goal is to maximize health benefits from available resources, prioritization done on the basis of needs can result in inefficient resource allocation (Cohen, 1994; Bernfort, 2003; Petrou, 1998). This criticism was instrumental in leading to the development of priority setting approaches based upon economic principles. For example, Donaldson & Mooney (1991) proposed a specific method for determining priorities in health care based on economic evaluation rather than needs assessment. Economic evaluation can be viewed as a natural step in the evolution of priority setting in health care as it incorporates the concept of capacity to benefit (from needs assessment) and explicitly adds the cost dimension to this  6 A distinction must be made between optimizing efficiency and improving efficiency. Removing resources from uses where they provide almost no benefits to redeploy them where they provide some benefits will improve efficiency but optimization would only be accidental. 5benefit measurement (to address the fundamental weakness of resource allocation based only on needs assessment). As a result, economic appraisal of health care programs became increasingly popular in the early 1990s, with the value of information relating intervention costs to health effects being recognized by “governments and other health care decision makers faced with difficult choices about the allocation of scarce health care resources” (O’Brien, Drummond, Labelle & Willan, 1994). Economic evaluation in health care is largely based on a logic similar to that of cost- benefit analysis (CBA) which emerged during the mid-1930s in the discipline of public administration (Hanley & Spash, 1993). The intractability of assigning dollar values to improvements in health status resulted in the adoption of a more health-care-friendly variant – cost-effectiveness analysis (CEA) – in which “benefits” are measured not in dollars but in terms of identifiable changes in dimensions relevant to those receiving care (e.g. improvements in blood pressure, reductions in days of hospitalization). In some situations, there was an interest in an overall measure of impact/benefit, rather than in specific sub-dimensions of impact, which led to the further evolution of CBA in health care based on the development of tools to measure changes in overall health status. The most common of these tools is the Quality-Adjusted-Life-Year (QALY), which provides “a common currency to assess the extent of the benefits gained from a variety of interventions in terms of health-related quality of life and survival for the patient” (Phillips & Thompson, 2009)7. When benefits are measured in QALYs, CEA becomes Cost-Utility-Analysis (CUA) and the incremental cost effectiveness ratio (or ICER) measures the cost of obtaining an additional QALY (Gerard, 1992). A well-known example of the use of economic evaluation in resource allocation decision-making is what became known as the ‘Oregon Plan’. In 1989, the state of Oregon reassessed what services should be covered by its Medicaid program. The goal was to provide the most health care benefits to the client population for the money available. Economic evaluation was the instrumental cornerstone of the plan, with significant allowance for public input and other considerations (with the role of these  7 For a historical review of the development of the QALY see Weinstein, Torrance, McGuire (2009). 6other inputs becoming more important after public backlash over the results of a narrow application of economic evaluation results). The entire process took five years to complete and was implemented in 1994 after adjustments to the economic evaluation list (to allow for other objectives). These adjustments were done on an ad-hoc basis without a formal framework (Sabik & Lie, 2008), leaving the process wide open to political and other influences.  There was much opposition to the rationale for the Oregon Plan with, for example, the Swedish Parliamentary Priorities Commission stating that: “The Commission…feels that the cost-efficiency principle should only be applied to comparisons of methods for treating the same illness” (Ham, 1997, p.54). Three types of problems with using CUA as a priority setting tool have emerged over the years: 1) concerns over the accuracy of the QALY measurements (issues of reliability, validity, and relevance) (McGregor, 2003), 2) issues regarding the use of thresholds (i.e. maximum acceptable cost per QALY) in resource allocation decisions, and 3) health care decision-makers pursuing more than the single objective of maximizing health benefits. Problems related to the use of thresholds and the multiple objectives of health care decision-makers directly affected the evolution of priority setting in health care. As mentioned, CUA can yield a cost per QALY gained (the ICER). In theory, decision- makers should allocate their resources to interventions with the lowest cost per additional QALY if their objective is to maximize health benefits to the population (as measured by QALYs gained) from a given pool of resources. However, decision-makers typically do not have locally relevant ICER estimates (and often no ICER estimates at all) for most interventions. When a more costly intervention does not produce any additional benefits, the conclusion is straightforward (this is also true for an intervention that is less expensive and still produces at least similar benefits). However, when there are additional benefits and additional costs, a decision must be made as to whether these additional benefits are worth the additional resources required. Because comparison with all other intervention options is not possible (due to the lack of ICER estimates), the decision as to whether or not this cost (additional resources) is ‘acceptable’ is usually made on the basis 7of a threshold level: proposed interventions are rejected if above a pre-specified cost per QALY gained and accepted if below the cut-off. Generally accepted threshold figures are $50,000 per QALY in the US (Vonlanthen, Slankamenac, & Ernst, 2010, p.535) and £30,000 per QALY in the UK (Neumann, 2005), although these figures have shifted over time and are, at the end of day, arbitrary.  Therefore, while the threshold in CUA is a key determinant of decisions, there is no consensus on the appropriate threshold figure or on the means to determine such a threshold. The issue of the multiple objectives of health care decision-makers compounds the problems posed by the need for a threshold. Theoretically, calculation of QALY gains may account for a broad spectrum of impacts (for example, one could argue that the degree to which one experienced empathy or fairness in his or her treatment episode affects future quality of life), but that is typically not the case (as gains tend to be more narrowly defined in terms of health status changes). Therefore, the decision-maker is left with a cost per QALY estimate that must be weighed against other objectives such as equity, risks, access, impact on research, and organization-specific concerns (such as religious considerations) in addition to being compared to the relevant threshold. Furthermore, CUA provides no guidance, indeed cannot provide guidance, as to how these other objectives can be used in tandem with the ICER to arrive at a decision. As summarized by Richardson (1997), “debate in the peak health economic journals over issues such as the Healthy Year Equivalent (HYE), the QALY and their correspondence with the axioms of economic theory are a long way from the type of analysis needed to clarify real world social objectives”. A specific example of this issue can be found in the challenges faced by NICE (the National Institute for Clinical Excellence) in the UK where, in addition to concerns over the level of the ICER threshold, “there is also a lack of clarity over the way in which factors other than health gain are taken into account – that is, the tradeoffs that are accepted between efficiency and objectives such as equity” (Devlin & Parkin, 2004). Wilson, Peacock and Ruta (2009) are very explicit on the limitations of the use of CUA for resource allocation decisions. They note that “QALY is a helpful but insufficient means of ranking alternative uses for scarce health-care funds at 8the local level”, as the QALYs “do not by themselves capture all criteria decision makers need to take into account” (p. 467). In summary, needs assessment, which was popular in the 1970s and 1980s, ultimately proved insufficient to guide resource allocation in health care. Economic evaluation, which came to prominence in the late 1980s and early 1990s, filled an obvious deficiency with needs assessment by explicitly addressing the issue of resource cost. While this form of analysis provides useful information for the purposes of comparing alternative interventions, the concerns identified above suggest that it is unlikely to meet decision makers’ needs with respect to priority setting across a broad set of services and program areas. In fact, despite advances in these methods, by the mid 1990s, resource allocation decisions in health care were still “influenced more strongly by inherited commitments and political bargaining than analytical methods” (Ham, 1997, p. 54). Analytic tools to assist decision makers with priority setting or resource allocation decisions clearly required further development.  This came in the form of structured, formal priority setting processes. Just as shortcomings with needs assessment led to advances that built on the information provided by this process, formal priority setting processes have built on the foundations provided by economic evaluation. CUA used information similar to that of needs assessment in the calculation of benefits related to specific interventions and then explicitly related resources used to the estimated benefits. Formal priority setting processes, in turn, use the information from CUA, whenever available, as a measure of how a given intervention meets the objective of efficiency in resource use, while also explicitly integrating multiple other criteria and stakeholders to provide the decision-maker with an overall assessment that matches his or her decision context. It is to this type of process, and in particular a framework known as Program Budgeting and Marginal Analysis (PBMA) that we now turn. 9Formal priority setting processes The advances in formal priority setting processes can be grouped into three areas: 1) procedures (development of techniques and tools), 2) process characteristics or qualities (fairness and equity have been the key characteristics), and 3) fit within the organizational decision-making structures. Broadly speaking, work in these three areas has been sequential (i.e. the focus was first on improving procedures, then it shifted to process characteristics and finally the emphasis is now on the fit within organizational structures), although in practice progress on all three areas has overlapped. Procedural development was aimed at developing a set of procedures that would guide decision-makers faced with a resource allocation question. Two basic approaches were followed.8 In the first approach, procedures were designed to create ‘quasi-market conditions’ for the management of certain activities so that decision-makers could use traditional business tools to arrive at decisions. This approach resulted in programs such as the ‘purchaser-provider contracting’ policy of the National Health Service (NHS) in the UK (Goddard, Mannion & Ferguson, 1997). Under this policy, a set of procedures divides the health care system into purchasers and providers and prescribes conditions and rules of engagement for determining the terms of contracts. For example, procedures defined types of contracts (e.g., block contracts or average cost contracts), quality measures to be included in contracts, and duration of contracts (Goddard et al., 1997). Within the organizational structure imposed by the policy’s procedures, market conditions applied and decisions reflected those conditions. As Mackintosh (1993) remarked, “Money will follow patients, aligning resources with need and rewarding the efficient” (p.139). The second approach to procedural development was to design processes that substitute for the marketplace. Under this approach, no attempt was made to recreate market conditions but instead the procedures focused on creating a decision-making  8 It must be noted that the two basic approaches to procedural design are not mutually exclusive and are in fact often complementary, as is the case in the NHS in the UK, for example. 10 process specific to the task of allocating health care resources. PBMA, which started to gain some traction in the mid 1990s (Viney, Haas & Mooney, 1995), represents an example of this approach. PBMA had been used by about 70 health care organizations in seven countries by 2003 (Mitton & Donaldson, 2003). Other similar approaches in the Canadian context include SLEEPERS, developed at the London Health Sciences Centre in London, Ontario (Martin, 2007), and EVIDEM, a process currently being developed by a team of health care consultants in Montreal (EVIDEM, 2009). These are all based on the same fundamental economic principles of opportunity cost and the margin. Changes to PBMA exemplify progress in formal priority setting processes. First, the process was refined in terms of the procedures that comprised the approach. For example, PBMA was described in the mid-1990s as a process comprising four steps (Peacock and Edwards, 1997) (see Table 1).  By the early 2000s, PBMA was modified to include a total of seven steps (Mitton & Donaldson, 2004b; Peacock et al., 2006) (see Table 2). TABLE 1   Steps in the PBMA process - 1997 (from Peacock & Edwards, 1997) Step Description 1 The first stage of a PBMA study consists of defining program areas to be studied. The exact choice and nature of the programs to be examined will depend on the organisation, its objectives, and existing budget areas. 2 The second stage is to identify services which may be potential options for contraction or expansion in the future. That is, a range of services are identified which may be considered for change in the future to attempt to improve the overall levels of benefits, or health gains, which individuals and populations receive. 3 The third step is to take the identified options for changes in services and evaluate those services in terms of their effectiveness, and in terms of the costs of providing those services. 4 The fourth, and final, step is to reallocate money according to the cost- effectiveness criteria and equity judgements. 11 TABLE 2  Steps in the PBMA process - 2004/6 (from Peacock et al., 2006) Step Objective Description PREPARATION PHASE 1 Determine the aim and scope of the priority setting exercise Will the analysis examine changes in services within a given programme or between programmes? 2 Compile a programme budget The resources and costs of programmes combined with activity information 3 Form a marginal analysis advisory panel The panel should include key stakeholders (managers, clinicians, consumers, etc) in the priority setting process 4 Determine locally relevant decision making criteria The advisory panel determines local priorities (maximising benefits, improving access and equity, reducing waiting times, etc) with reference to national, regional, and local objectives IMPLEMENTATION PHASE 5 Identify where services could grow and where resources could be released through improved efficiency or scaling back or stopping some services The panel uses the programme budget along with information on decision making objectives, evidence on benefits from service, changes in local health care needs, and policy guidance to highlight options for investment and disinvestment 6 Evaluate investments and disinvestments Evaluate the costs and benefits for each option and make recommendations for change 7 Validate results and reallocate resources Re-examine and validate evidence and judgments used in the process and reallocate resources according to cost-benefit ratios and other decision making criteria Two of the additions to the process as it was in the mid-1990s formalized what became considered standard practice (compiling a program budget that lays out where the resources are currently used, and determining locally relevant rating criteria). The other addition (the formation of an Advisory Panel) addressed an important limitation of the PBMA process of the mid-1990s: the lack of explicit acknowledgement of the role of expert opinion. This suggests a formal recognition that priority setting in health care 12 cannot be solved through a mechanical process because of the inherent uncertainty embedded in health care decisions. Beyond the additional steps, another notable difference between the PBMA process descriptions of 1997 and 2004/6 is the reference in the latter version to ethics and how ethical considerations can be integrated into PBMA (the process formally became more than the series of steps). The explicit inclusion of ethics is an example of the impact of the second phase of formal priority setting processes development, which focused on process characteristics or qualities. After extensive research work on procedures and in the face of ongoing challenges with demonstrating value in terms of impact on decisions, the research focus turned to process characteristics and their measurement, posing questions such as whether the process is fair and equitable.9 Ethics and equity in PBMA were first addressed through the formal evaluation of PBMA using a framework known as ‘Accountability for Reasonableness’ (A4R)10 (Daniels & Sabin, 2002; Gibson, Mitton, Martin, Donaldson & Singer, 2006). This area of development was introduced by Peacock et al. (2006) who presented two ways to address ethics: (1) interview or survey relevant stakeholders after the results have been implemented, or (2) make the conditions of A4R an integral part of the design of a priority setting process so that ethical evaluation is conducted alongside economic appraisal. Urquhart, Mitton and Peacock (2008) report on a PBMA pilot project in which A4R was integrated from the outset into PBMA. Their objective was specifically to implement a priority setting process incorporating both ethical considerations and economic principles11. The third phase of development for priority setting processes is addressing issues of integration into organizational decision-making processes or, in other words, the practical aspects of process implementation. One important component of  research on the practical aspects of process implementation has been the work on the identification of  9 The underlying thinking of the second phase was that decisions made with fair and equitable processes would demonstrably be fair and equitable, which in itself would be an improvement. 10 A4R has been described as an ethical approach to priority setting and has frequently been used in the health care context (Gibson et al., 2006; Peacock et al., 2006). 11 This pattern of evolution also applies to processes under the ‘quasi-market conditions’ approach mentioned above. Changes to these processes first focused mostly on the refinements to procedures and then increasingly incorporated ethical considerations. 13 ‘success factors’ for priority setting (e.g., Carter et al., 2000; Mitton & Donaldson, 2003; Mitton & Donaldson, 2004; Gibson et al., 2006; Peacock, Richardson, Carter & Edwards, 2007; Teng, Mitton & MacKenzie, 2007 ). These ‘success factors’, or facilitators and barriers, are organizational features that are thought to affect the level of impact that the implementation of a priority setting process has on resource allocation. Key success factors identified include: strong leadership, organizational stability (low staff turnover, especially in strategic planning divisions), a high-level champion for the process, an organizational culture to learn, having resource allocation decisions that must be made (i.e. where the status quo is not an option), and a strong relationship between physicians and managers (Donaldson, 2001; Mitton & Donaldson, 2003; Peacock et al., 2009; Teng, et al., 2007). Despite approximately fifteen years of research and improvements to procedures and process characteristics and a better understanding of organizational success factors, issues remain with formal priority setting processes. In a study of commissioning practice12 in the NHS in the UK, Bate et al. (2007) found that resource allocation decisions still tended to be largely driven by political, historical and clinical considerations (p.258). And this despite the facts that the commissioning strategy mandated by the NHS involves the use of economic principles and that decision-makers were found to largely understand these economic principles and show a good grasp of how they work. These findings led these authors to conclude that “if economic methods are to have an impact in ‘real-life’ priority setting decisions, then they must also be pragmatically applicable” (p.259, emphasis added). Peacock, Mitton, Bate, McCoy and Donaldson (2009) summarize the situation with regards to priority setting process implementation in a similar way: “In recent literature, economic approaches to priority setting have been criticised on the grounds that they have only had limited success in practice. This has been, at least in part, due to their failure to adequately capture the complex and multifaceted nature of both objectives and constraints in health service decision-making” (p.897).  It seems that priority setting processes continue to face  12 In the NHS, commissioning is making decisions about resource allocation. In the context of the NHS, studying commissioning practice is therefore studying how resource allocation decisions are made. 14 practical challenges to their implementation. In the words of Bate et al. (2007), the task ahead is clear: “In our view, acknowledging the importance of contextual factors (e.g. organizational behavior, decision-makers’ requirements) and the pragmatics of priority setting, allows health economists to identify barriers faced by decision-makers in adopting their methods in practice which can then be addressed in adapting them for future use” (p.260). Set within this overarching context, this research project continues in the tradition of investigating the practical challenges to priority setting implementation. More specifically, this project continues with the success factors approach to the investigation of those practical challenges. The research on success factors will be set within a framework developed by Teng et al (2007) who classified success factors as either facilitators or barriers and as affecting either the way inputs are integrated in the process or the way the process recommendations are handled. This framework is illustrated in Figure 1. FIGURE 1 Framework for success factors analysis   While the decision was made to pursue the Explicit priority setting process Outputs Barriers - no genuine buy-in - lack of skills in priority setting - lack of shared vision* - competing priorities* Barriers - vertical budget silos - politics trumps evidence-based decisions - no (real or perceived) authority to change - vested interest* - misalignment of incentives Facilitators - high level champion; strong leadership - culture of learning - earmarked resources for process - commitment to process* Facilitators - culture open to change - earmarked resources for follow-up - consistent application* - demonstrated results* - data-driven culture* Inputs Ideal Process communication of vision - Integrate strategic goals - Time-sensitive - Evidence-based 15 While this project continues with the current approach to the study of practical challenges to implementation, the research strategy selected was different than what was previously done. The focus of research was narrowed to  one specific practical challenge and one specific priority setting process. The challenge selected was the identification of disinvestment options (specific potential disinvestments offered for consideration) within the PBMA process. The goal, in line with the findings of Bate et al. (2007), was to increase the ‘pragmatism’ of the PBMA procedures around the identification of disinvestment options so that the priority setting process can become more effective.  The reasoning behind the choices of the issue of disinvestments and of the PBMA process is explained in the next two sections. Why the focus on disinvestments in health care? Before answering this question it is necessary to have a clear understanding of what a disinvestment in health care is. Definition of disinvestment There are two types or forms of disinvestments in the health services literature. The first form of disinvestments “relates to the processes of (partially or completely) withdrawing health resources from existing health care practices, procedures, technologies, or pharmaceuticals that are deemed to deliver little or no health gain for their cost, and thus do not represent efficient health resource allocation” (Elshaug, Hiller & Moss, 2008, p. 2). Procedures to identify these types of disinvestment options focus on identifying interventions suspected of providing no or minimal benefits and measuring their actual impact. The logic is simple, yet three important challenges have been identified: the establishment of institutions dedicated to the identification of these disinvestments, raising the profile of the potential benefits of such disinvestments in health care to ensure political support, and the development of analytical tools specific to this task (Elshaug et al., 2008). The decision criterion for this form of disinvestment is very similar to that for new technologies, and typically revolves around a threshold 16 level13 (this is the same concept of threshold as was explained in reference to Cost Utility Analysis) above which an intervention should be discontinued. The second form of disinvestments refers to “services which are providing positive benefits but where it is considered that there may yet be greater benefits to be had by shifting resources elsewhere” (Mitton & Donaldson, 2004b, p.118). With disinvestments of this form, the procedures to identify disinvestment options, as part of a priority setting implementation, are of a different nature and typically revolve around a requirement to put forward disinvestment proposals totaling a preset value (typically broken down by organizational sub-units such as departments or portfolios)14. Such requirements are necessary because no absolute rule (e.g. cost per QALY above a certain threshold, or no evidence of benefits) determines whether an intervention should be reduced or eliminated (i.e. decision-makers do not have a precise description of what to look for). The decision as to whether or not a given intervention will be subject to review for possible disinvestment depends upon its relative value, which depends on the value of alternative spending (the opportunity cost). While both forms of disinvestments are necessary to ensure good management of health care resources, priority setting is concerned with the second form because its application requires comparison, or an assessment of relative value vis-à-vis alternative uses of resources. Therefore this research project is concerned primarily with the second form of disinvestment and the use of the term ‘disinvestment’ will, henceforth, refer to this form. In theory, the two forms of disinvestments are mutually exclusive: interventions suspected of producing minimal or no benefits should be assessed on their own while priority setting processes deal with ‘productive’ interventions. However, in reality,  13 Interestingly, while the threshold or maximum acceptable cost per additional QALY for a new technology should conceptually be the same as that for existing technologies or practices, studies show people are willing to pay more per QALY for existing interventions. In other words, the higher threshold for disinvestments reflects a reluctance to discontinue current practices (Drummond, Sculpher, Torrance, O'Brien & Stoddart, 2005). 14 These proposals should involve the services that decision-makers believe provide the least benefits. 17 decision-makers do not know, a priori, in which category a particular program might fit (Elshaug et al., 2008). This means some of the first form of disinvestment options may be included in priority setting processes. The only problem that this may pose is the risk that such disinvestment options may displace disinvestments options of the second type in the process (given that a preset amount of disinvestment options is typically required to be brought forward).  This might be thought of as the ‘low-hanging fruit’ approach to disinvestment, and helps decision-makers avoid tougher decisions in which competing productive options must be chosen for disinvestment. Since disinvestment options of the first type should be acted on anyways, their inclusion in a priority setting process simply reduces the potential impact of the process. While the two forms of disinvestment are different in nature, and require different process of identification, one key common characteristic is that they both represent an administrative decision to withdraw some or all of the resources from a program. This formed the basis for the definition of disinvestment that was used in this research project: A disinvestment is an explicit decision to reduce the volume and/or the quality of an existing service to reduce or contain costs15. Why are disinvestments important? Disinvestments are important to priority setting processes because, typically, very little uncommitted new money (i.e. money not tied to a specific activity) flows to any health care organization. Without uncommitted new money, resource allocation decisions are limited to the re-allocation of resources saved in one of only two possible ways: 1) through efficiency gains or 2) through disinvestments. As the potential for further efficiency gains is constantly reduced through ongoing efforts to improve efficiency, the impact of priority setting processes becomes increasingly determined by the extent to which disinvestment options are identified and acted upon.  15 Therefore, the second form of disinvestment would be an explicit decision to reduce the volume and/or the quality of a ‘productive’ service to reduce or contain costs. 18 Disinvestments have long been a challenge for decision-makers. As early as 1994, decision-makers had displayed a “natural reluctance” to participate in the identification of disinvestment options within formal priority setting processes (Cohen, 1994). More recently, the identification of disinvestment options was still regarded as a “major hurdle” in the implementation of priority setting activities (Mitton & Donaldson, 2004b, p.118), despite all the changes made to these processes. Identifying and agreeing on possible disinvestments is therefore a long-standing and significant problem in priority setting, and yet these steps represent necessary elements if the process is to have a meaningful impact on resource allocation patterns in the absence of significant amounts of new, uncommitted, funding. In summary, disinvestments in productive health interventions are a necessary part of priority setting when there is limited new, uncommitted, funding available. Yet, decision-makers have long showed a reluctance to put forward proposals for such disinvestments as part of priority setting implementation. The identification of disinvestment options, or proposals, is therefore an important practical challenge to successful priority setting implementation. Why the Program Budgeting and Marginal Analysis (PBMA) process? A decision was made to focus on one specific priority setting process so as to ensure consistency throughout the research project and avoid the potential confounding effect of using different processes in different parts of the project. PBMA was selected because it is a well-developed, well-known and widely-used process, it has been a significant subject of research for at least 15 years, and it is fundamentally sound, as it is supported by basic economic principles. One particularly relevant outcome of all the previous research on PBMA is very clear guidelines around the issue of disinvestment options. A literature review identified 103 published papers on PBMA as of 2003 (Mitton & Donaldson, 2004b). Its use has been documented in more than 70 health care organizations in seven countries (Mitton & Donaldson, 2004b), including five of the six health authorities in British Columbia. For these reasons, it was believed that using 19 PBMA throughout the project would provide the desired consistency while minimizing the risk that the practical challenges to the identification of disinvestment options that would be observed would be related to overall process weaknesses or misunderstandings. Relevance The evidence suggests that resource allocation decisions in health care are becoming an increasingly important determinant of performance but that problems exist with priority setting processes. The main indicator of the existence of such problems is that historical resource allocation patterns and political influences are still major determinants of resource allocation decisions. The specific practical challenge that is being addressed in this research project is one aspect of priority setting that has long been difficult for decision-makers to embrace. Yet, without disinvestment options, priority setting becomes an assessment, and ranking, of possible investments if new money becomes available. This is a useful exercise, but one that does not address how an organization is to get from where it is to where it might wish to be in an environment of constrained or declining resource availability, and therefore is of limited relevance. Despite being a necessary aspect of priority setting, the identification of disinvestment options has attracted limited attention to date. Action with regards to the first form of disinvestments (interventions that are ineffective) is gaining some momentum with, for example, the introduction in 2006 of a new mandate for NICE (UK) to help the NHS identify ineffective interventions (Pearson & Littlejohns, 2007). However, as explained above, identifying and acting on interventions that are ineffective is not priority setting as no choice-making is involved.16 Rather, the second form of disinvestments (potential for more benefits elsewhere) holds more relevance for priority setting processes. Very little work on the second form of disinvestments is being done beyond reinforcing how important these disinvestments are to the management of health care resources (Mitton & Donaldson, 2004a). This research project therefore addresses an  16 No choice is involved unless one considers the choice between stopping doing something that is unproductive and not stopping. Even then, one can argue that no guidance is needed to make that choice (therefore that choice does not involve any form of priority setting activity). 20 area of priority setting processes that has been neglected yet is necessary to the expansion of their role in resource allocation in health care. Study purpose and objectives The current study aims to improve of our understanding of the challenges related to the identification of disinvestment options and on this basis, help make this part of priority setting more pragmatic (i.e. more attuned to the organizational context in which decision-makers operate) so that the priority setting exercises can lead to better results. Underpinning this purpose are three specific objectives: 1) Summarize the evidence on cost containment or reduction strategies in health care and organize it in such a way as to make it user-friendly for decision-makers. 2) Develop an understanding of the barriers to identifying disinvestment options and investigate how this aspect of priority setting processes may conflict with the organizational context. 3) Develop and evaluate potential solution(s) to the issues identified in objective two (i.e. how the identification of disinvestment options conflicts with the organizational context). Overview of dissertation This research project was comprised of three major phases which align with the three objectives stated above. The next three chapters in the thesis describe the goals, methods and results of each phase, as well as specific points of discussion. In the final chapter of the thesis, a summary of the results is offered, as are overall conclusions. A brief overview of these chapters is provided here. 21 Phase 1- Chapter 2 The first phase of the research project was a literature review on the subject of cost containment or cost reduction strategies in health care. The goal for this phase was to assess whether the practical challenges posed by the identification of disinvestment options can be explained by a lack of knowledge on how to accomplish this task. Phase 2- Chapter 3 The second phase of the project involved exploratory research on priority setting in health care with a focus on the activities related to the identification of disinvestment options. Four PBMA case studies were conducted and the successes and challenges regarding the identification of disinvestment options were observed. These observations were analyzed with the goal of producing a plausible explanation for variations in the degree of success, and in the level of challenge faced, in the identification of disinvestment options across the four settings. Phase 3- Chapter 4 In the third phase of this project, the proposed explanation developed in Chapter 3 was used to implement a specific procedural change to PBMA in an unrelated real-life setting. Successes in the identification of disinvestment options were assessed in terms of disinvestment options proposed (and realized), and also in relation to the history of disinvestments in the subject organization. Chapter 5- Conclusion The concluding chapter focuses on contextualizing the key findings from this research project. The relationship of the key findings to existing knowledge is explained, focusing on the specific contribution of these findings. Finally, limitations and next steps are discussed. 22 CHAPTER 2 HOW TO CONTROL THE COSTS OF HEALTH CARE SERVICES - AN INVENTORY OF STRATEGIC OPTIONS Introduction The identification of disinvestment options is a well-known, long-standing and significant challenge to the successful implementation of priority setting processes in health care, and of PBMA in particular (Cohen, 1995; Mitton & Donaldson, 2004b). A logical starting point in the investigation of this challenge is a review of the literature on cost reduction or containment strategies in health care. The purposes of such a review are to dispel the notion that the barriers to the identification of disinvestment options are rooted in a lack of knowledge on how to accomplish this task and to facilitate access to this knowledge for decision-makers. This chapter includes 1) an inventory of the methods or strategies for cost reduction or containment in the health services research literature; and 2) a classification framework to organize this inventory in a manner that is helpful to decision-makers. The availability of a formal classification of cost reduction or containment strategies is seen as an important tool for decision-makers attempting to understand the available strategies, and the differences amongst them, so as to make informed decisions about the most appropriate mix of strategies for a given context. Methods The purpose of this review was not to put together a list of all existing strategies for cost reduction or containment. Rather, the intent was to identify the most common, or ‘popular’, options available. In order to identify such options the Pubmed database was searched on the basis that if a strategy was not the subject of multiple references in Pubmed it would not qualify as a ‘popular’ or ‘common’ strategy. The search terms used were: ‘cost reducing’ OR ‘cost containment’ AND ‘plans’ OR ‘initiatives’ and ‘disinvestment’. All retrieved references were reviewed and those that presented a 23 specific method of cost reduction or containment were retained. The search was limited to papers published since January 1, 1998 (the literature search was completed in May 2008 and papers published until that date were included). The justification for the choice of January 1998 as the cut-off date is a belief that strategies that were implemented more than ten years ago, and not frequently since then, would have very limited relevance for decision-makers because of the significant changes in the economic and technological contexts. In total, 1093 abstracts were retrieved and reviewed. Abstract that contained reports of real life implementation of cost reduction or containment strategies were retained. There were 137 such abstracts, which indicates that most papers on the subject do not report on real life experiences. For each of the retained abstracts, the method of cost reduction or containment was noted. Each relevant reference provided either a new strategy to be added to the list or further information on a previously identified strategy. The frequency of references was used for the purpose of selecting the most common strategies - popularity of publication was used as an indicator of popularity of implementation. Once this initial review was completed and the most popular methods were identified, the “Citation Pearl Growing” search strategy was used to refine the search (Ramer, 2005). Each of the methods identified were entered as a ‘second round’ search term in Pubmed; additional references were reviewed until saturation of information was achieved. Finally, Pubmed was searched for ‘challenges’ in association with each of the methods identified. Once the inventory was established, the strategies were organized by categorizing them in a manner most relevant to healthcare decision-makers in terms of the expected source of resistance to the implementation of cost reduction or containment strategies in a publicly-funded system. 24 Results Overview In total, twelve different strategies17 to achieve cost reductions or containment were identified. These strategies are outlined below and accompanied by a list of references for each (provided within the body of the description of each strategy). More generally, the following characteristics were observed: 1) while each strategy can be applied on its own, many can be used in combination with others; 2) the possible scope of implementation of each strategy is limited, for a given health care organization, by its jurisdictional authority; 3) the primary purpose of each strategy is to reduce costs; they are not designed to improve the quality of health care services delivered, though this can be a by- product. Quality improvements could, however, come from the re-deployment of the freed-up resources; and finally, 4) potential savings do not have to be ‘given up’ by the organization employing the strategy, in order for that strategy in that context to be considered successful. A decision can be made a priori to use the savings, for example, to accommodate a greater number of patients. The strategy can then be seen as successful at containing the cost increase associated with greater volume. The twelve strategies are summarized in Table 3.  17 A strategy in this context is defined as a cohesive set of actions - that is, policies, directives or guidelines - that can be implemented by a decision-maker. 25 TABLE 3   Summary of cost reduction or containment strategies Strategies to reduce or contain health care costs 1. Incentives for providers This approach has historically targeted primarily physicians (Bodrock & Mion, 2008). It can, however, be directed at any individual, group, or class/category of personnel involved in the provision or management of health care services (Hopkins, 1999). The incentives referred to can take one of two forms -- gainsharing (Ketcham & Furukawa, Strategies Summary description Incentives for providers Financial rewards to encourage some changes in behavior from providers Supply management Reducing cost and/or quantity of inputs in delivery of services Compensation system for physicians Use of some non-fee-for-service payment systems for physicians Incentives for patients Providing patients with a financial stake in their treatment decisions Reductions in coverage Reducing range of services provided or restricting eligibility for these services Systematic use of cost-effectiveness analysis All proposed new spending subjected to cost- effectiveness analysis Changing the mix of services currently offered Re-allocation of resources to maximize impact on overall population served Increased emphasis on health promotion and prevention Reducing the impact of specific known health risk factors Investment in information technologies Spending on information technologies to increase productivity Re-structuring of clinical decision- making Establishment of shared decision-making structures to improve clinical decision-making Institutional re-organization focused on service delivery efficiency Re-organization of responsibilities to improve efficiency Strengthening of healthcare workers health and safety programs Introduction of programs aimed at reducing disease and injuries to staff 26 2008), where some proportion of any savings generated is distributed amongst the professionals who are the subject of the incentives, either in the form of direct (monetary) or indirect (additional departmental resources) ‘reward’ (Williams, 2008); or additional pre-determined payments to reward specific behaviors, such as ensuring that treatment follows specific guidelines. This latter form of financial incentives includes pay-for- performance programs (Smith, 2007) and pay-for-quality programs (Young et al., 2007). The rationale for this approach is the presumed behavioral response to the availability of financial rewards.   This strategy can be particularly useful as a means of improving the reach and effectiveness of prevention services, or of reducing the use of costly drugs, devices or treatments. For example, the Lahey Clinic Medical Center, in the United States, reports success in reducing the cost of implants for hip and knee replacement surgeries, without adverse impact on patient outcomes, through a gainsharing program with the surgeons (Healy & Iorio, 2007). Two important challenges in the implementation of this strategy are 1) the management of the relationship between physicians and patients: studies have shown that the introduction of financial incentives for physicians can have a detrimental impact on patients’ trust in their physician (Gallagher, St Peter, Chesney & Lo, 2001; Keating et al., 2007); and 2) ensuring that the financial incentives do not have unintended effects on quality of care by creating conflicting objectives for physicians (Petersen, Woodard, Urech, Daw & Sookanan, 2006). 2. Supply management This strategy focuses on the cost of the inputs into the provision of health care services (Williams, 2007), independent of the utilization decisions made by the providers themselves. For example, this strategy would focus on securing a lab test at the lowest possible cost, independent of whether or not the test itself is necessary or appropriate. Implementation involves making cost minimization part of the mandate for those in charge of ensuring availability of supplies (broadly defined to include, buildings, equipment and labor).  Reducing costs through this strategy can be achieved from one or 27 both of reducing the quantity of inputs required and reducing the acquisition costs (price) of these inputs.  Reductions in the quantity of inputs are essentially improvements in efficiency resulting from the elimination of unnecessary (or unnecessarily complex) processes, or reducing or eliminating raw materials that are not necessary (Gebara, Najarian, Wagner, & Narla, 2004). Such efforts are also known as re-engineering (Schumock, Michaud & Guenette, 1999) or ‘supply-chain’ management (Davis, 2004). Reducing the unit cost of the inputs can be achieved, for example, by having the health care organization, on its own or in concert with others, use its size, and more aggressive price negotiations, to obtain discounts (Frezza, Girnys, Silich, & Coppa, 2000), changing suppliers, or accessing labor contracted at lower prices. Successful implementation of this strategy is challenged by the quality of the purchasing and contracting expertise available within management’s ranks, as these decisions involve more than simply comparing prices, and by the limited range of autonomous providers competing to fulfill health care organizations’ needs (Ham, 2008). The lack of competing providers can be defined broadly to include the challenges around the organization of labor in health care when, for example, most of a certain category of workers can be represented by the same union irrespective of their place of employment . 3. Compensation system for physicians This strategy addresses the core of physician compensation as opposed to incentive programs, such as pay-for-performance, which are typically add-ons targeted to specific objectives (Reschovsky, Hadley & Landon, 2006). The rationale is that a system of compensation that ties remuneration to volume provides, in itself, no incentives for resource conservation. Some form of non-fee-for-service payment system for physicians is a common proposal under this strategy (Goroll, Berenson, Schoenbaum, & Gardner, 2007; Fundamental, 2007; Long, Coughlin & King, 2005). For example, a capitation system, full or partial, offers a built-in incentive to assess the added value of each possible service, which can lead to lower costs per patient. For example, contract capitation has been found to lead to a significant decrease in cataract extraction rates 28 amongst a network of ophthalmologists and optometrists in a pre-post study (Shrank, Ettner, Slavin & Kaplan, 2005). Capitation seems particularly suited for relatively homogeneous patient groups with extensive needs, such as mental health patients (Bloom et al., 2002). An important challenge in implementing this strategy is the determination of equitable capitation rates for different populations (Langton & Crampton, 2008; Ray, Daugherty & Meador, 2003; Rice & Smith, 2000). There is also the issue of participation in the program: if participation is voluntary, or simply not targeted, we can expect that a capitation program would attract mostly those physicians whose practice would benefit financially from capitation payment, which could in fact result in higher overall costs. 4. Incentives for patients The rationale for this strategy is that if patients have no stake in the financial consequences of their choices, financial considerations are likely to be ignored by them in making these choices. Implementation of this strategy revolves often around the increase of co-payments (Babazono et al., 2004; Regopoulos & Trude, 2004; Domaszewicz, 2006). Increased co-payments are seen as incentives for patients to assess from their perspective the value of proposed treatments (especially drugs). There are also, however, many instances of positive incentive programs that involve the provision of specific benefits to patients who exhibit a desired behavior (such as less smoking or weight loss) or to encourage compliance with a specific treatment protocol. For instance, patients who collaborate in the management of their disease can be assured of privileged access to physicians (Reid, 1998), or reduced co-payments can be used to encourage medication adherence (Chernew et al., 2008). A significant challenge in the implementation of this strategy is to address the problem of patients not adhering to required, and potentially life-saving, treatment plans because of financial considerations when increased co- payments are used to reduce costs to the organization (Fendrick & Chernew, 2006). Non- adherence to required treatment would lead to, at best, cost-shifting to another part of the organization or the health care system. 29 5. Reductions in coverage The rationale for this strategy is straightforward: if a health care organization is responsible for the funding of relatively fewer services, its costs will be lower, all else being equal. Coverage can be reduced in two basic ways: fewer types of services can be insured or access to reimbursements can be limited. Reductions in the range of insured services mean that some services will be de-listed for reimbursement18 (Abourjaily, Gouveia, Selker, & Zucker, 2005; Weinstock, 2006). If the service range is preserved, reductions in coverage can take the form of reimbursement caps (Soumerai, 2003) or of other restrictions on use of certain treatments, such as step-therapy (Yokoyama, Yang, Preblick & Frech-Tamas, 2007), which limit volume without changing basic eligibility criteria. Such limitations are often known as “utilization management” (Wickizer & Lessler, 2002). A common focus in this strategy is on services that have been found to provide little or no benefits. For example, the National Institute for Clinical Excellence (NICE) in the UK recently announced a formal policy aimed at eliminating from the NHS “treatments that do not improve health or are poor value for money” (Pearson & Littlejohns, 2007; Elshaug et al., 2008). The ethical issues raised by this strategy pose a significant challenge, however, as those with lesser voice and greater needs, such as geriatric patients (Levy, 2000), may be at risk of having unfair restrictions imposed upon them. For such groups, value of treatments may be underestimated resulting in miscalculation of net benefits and unfair relative spending. 6. Systematic use of cost-effectiveness analysis to screen proposed new spending This strategy rests on the premise that if a health care organization systematically applies value for money ‘screens’ to all proposed changes to purchases and programs, the end result will be lower costs for the same or better services provided. Implementation typically centers on the application of evidence produced by technology, drug, procedure and program cost-effectiveness evaluations. However, unlike the more common threshold  18 No papers were found on the de-listing of procedures in public health insurance programs; the only examples of de-listing in such organizations applied to drugs. 30 approach to funding approvals,19 in this context the cost-effectiveness analysis focus more on the net financial impact, akin to a measure of return on investment (Berg, Thomas, Silverstein, Neel & Mireles, 2004). The expected result of this strategy is the increased use of services that lead to more than offsetting cost reductions in other services. For example, Berg et al (2004) found that for every dollar spent on increasing influenza vaccination rates, $2.21 was later saved in reduced medical services utilization. A variation on this approach is the estimation of the net impact on the cost of treatment per patient per year. Jhaveri, Seal, Pollack & Wertz, (2007), for example, found that a specific treatment for insomnia resulted in a net saving of $1,253 per patient per year. There are two significant challenges in the systematic use of cost-effectiveness analysis. First, is the issue of the substantial resource and expertise requirements to conduct, or even simply contextualize, such cost-effectiveness analysis studies. And second, is the assessment of the validity of predicted benefits, which can be manipulated to produce the desired results (Bottomley & Raymond, 2007). Limited reliable data and outcomes measures are a problem for all strategies and all decisions, but this problem is particularly acute here because the cost-effectiveness measurement is the central focus of decision. As such, this measurement becomes a target around which vested interests can band. 7. Changing the mix of services currently offered This strategy focuses specifically on re-allocating resources through various means including: 1) an analysis of value for money from current practices leading to the determination of treatment protocols that are more efficient20 i.e. where certain high cost/ low benefit practices are avoided (Allen, Hahm, & Polk, 2003); and 2) an increased focus on quality of care leading to changes in practice that would reduce the risk of adverse events (Jackson, Sistrunk & Staman, 2003; Mitton, Dionne, Peacock & Sheps, 2006). The  19 This approach typically focuses the analysis on a cost per unit of benefit, usually a QALY, with approval being recommended for procedures, technologies or drugs costing less than a set threshold per QALY. 20 This analysis for value would likely take the form of a cost-effectiveness analysis but the difference with the strategy number 6 is that the target is current practices. This strategy is aimed directly at current spending as opposed to new spending. 31 rationale for this strategy is that, typically, existing health care practices are not reviewed in terms of their impact for the money spent and the implementation of such a systematic review is likely to highlight opportunities for resource re-allocations that would improve health benefits, allowing more of the demand for services to be met with existing resources. This strategy can be applied at any level of an organization, from an overall application, where some groups of patients may end up as ‘winners’ and some as ‘losers’, to a very localized application, where the choices will be between the different ways to address a given condition for a given group of patients.  An important challenge in this strategy is the identification of the programs or processes to be reviewed since likely not all could be assessed because of the excessive demands this would impose on decision- makers (Elshaug et al., 2008). Currently, as pointed out by Elshaug et al. (2008), most evaluation efforts are aimed at new technologies or programs and even this limited level of review seems to overwhelm decision-makers. 8. Increased emphasis on health promotion and prevention The emphasis in this strategy is on specific, known risk factors for disease such as alcohol, tobacco, illicit drugs, sun exposure or obesity. The two key questions to be addressed are: what risk factor(s) to target (Giesbrecht & Haydon, 2006; Melia, Pendry, Eiser, Harland & Moss, 2000) and what is the best method of action (Speroff, Miles & Matthews, 1998; Olds, Sadler & Kitzman, 2007; Ruser et al., 2005). The rationale is that, overall, the savings from a reduction in downstream cases will more than offset the additional spending to reduce exposure amongst a large population (Tan, 2002). By definition, this strategy can only lead to cost reductions or containment when measured over an appropriate period of time. Therefore, the options under this strategy will be limited by the time horizon where cost reductions are required. Often, over a short term horizon, this strategy will lead to increased spending.  Accordingly, an important 32 challenge is the selection of a time frame appropriate to the context and that is acceptable to decision-makers and payers21 (Moore, 1998). 9. Investment in information technologies This strategy involves focused spending on information technology programs that are expected to lead to more than offsetting reductions in costs.  Although this strategy is all about more efficient use of resources, this is different than supply management where the focus is to reduce prices paid for inputs or reduce unnecessary actions. This strategy involves new spending. As such, this strategy is closer to the systematic use of cost- effectiveness analysis for new spending (strategy number 6).  There are however important differences that lead to this being a distinct strategy. The new spending of strategy number 6 is typically spending aimed at a specific condition, like a drug, a treatment, or a device. Spending on information technologies is not usually targeted that way but is more aimed at systems that are used by many programs. This makes benefits evaluation a very different task. In fact, expected benefits from information technologies spending usually take the form of support service costs reductions. Information technologies are not a health care service in themselves but are about facilitating the delivery of health care services. The expected cost reductions can come in many forms. For example, there can be direct impacts where better information lessens the workload, such as improvements in the interface between different parts of a health care organization. In these cases, information technologies reduce the net cost (usually expressed in time saved) of obtaining necessary information. Massy-Westropp, Giles, Law, Phillips, & Crotty (2005) report on an electronic data linking system between a hospital and a community care organization in Australia. Among many benefits, they found that the system led to a reduction in labor costs as staff in both organizations “indicated they were better  21 The issue of an appropriate time frame ultimately revolves around the estimation of a discount rate. For a decision maker, and some other stakeholders, the discount rate may be very high as the main preoccupation may be on results in the next 3 to 5 years. Society at large may have a much lower discount rate. 33 informed about their client’s medical and disability status” ( Massy-Westropp et al., 2005). There can also be indirect impacts where the information system becomes a tool that permits better management of resources, through improvements in workflow (Antonucci & Bender, 1998) or through improved measurements of outcomes (Bates et al., 1998).  In those cases, information technologies provide new information with a value, in terms of more efficient resource use, greater than its cost. A significant challenge to the successful implementation of this strategy is to ensure that the organizational structure is set up to accommodate the output from the new information technologies (Marchibroda, 2008). For instance, information designed to improve the coordination of care will only be of use if the parties involved have the authority to increase cooperation. It must also be noted that good implementation management seems particularly important with this strategy; there are examples of successful implementation (Massy-Westropp et al., 2005; Bates et al., 1998) but there are many examples of failure, with some arguing that successes are the exception (Sharman, 2007, p.s223). 10. Re-structuring of clinical decision-making This strategy promotes the establishment of shared decision-making structures to improve clinical decision-making. There are two types of such structures: those involving providers and patients and those involving health professionals. Patient- provider structures address mainly specific treatment decisions (O’Connor, Llewellyn-Thomas & Flood, 2004) while health professional structures serve two main purposes: 1) the development of clinical guidelines based on communal expertise (Lewandrowski, 2003; Burns et al., 2003) or 2) improvements in communications, through formal team briefings for example, which lead to the identification and resolution of problems (Lingard et al., 2008).  The overall rationale is that combining the preferences of patients with the knowledge of providers or combining the varied expertise of different health 34 professionals is more likely to lead to improved use of resources. This would translate into savings, limiting the unnecessary use of resources by, for example, avoiding treatments unwanted by patients or preventing duplicate diagnostic services as a result of better information sharing amongst providers. For instance, many randomized trials have shown that involvement of patients in clinical decision-making, through the use of ‘patient decision aids’, “prevents overuse of options that informed patients do not value” (O’Connor et al., 2004). In implementing this strategy, a significant challenge is a common language. For example, research has shown that patients’ understanding of probabilities of events depends on the language used by the physician (France, Keen & Bowyer 2008). 11. Institutional re-organization focused on service delivery efficiency Under this strategy, the organizational structure would be reviewed with a focus on improved efficiency. Possible changes under this strategy could involve a redistribution of responsibilities and/or the creation of new units of service. The rationale is that organizational structures have typically been shaped over time by a series of unrelated events or policies. It is very unlikely that this series of changes has kept operational efficiencies as a main focus. In fact it can easily have had the opposite effect (France, Francis & Lawrence, 2003), which means that there are possible adjustments that would improve operational efficiencies. Implementation of this strategy involves a judicious balance between the devolution of authority and integration of programs. Devolution of authority brings decision-making where the information is, thereby simplifying information exchange and increasing responsiveness (Bellanger & Mosse, 2005). Integration of programs, for example, the integration of some mental health services into primary care (Otis, Macdonald & Dobscha, 2006) and the resulting concentration of authority, facilitates better coordination (Sime, 2005) or can produce economies of scale (Birkmeyer, Skinner, & Wennberg, 2002). Thus, this balance involves assessing, using the example of mental health care and primary care, whether the benefits of having the resource allocation decisions regarding mental health with the health care program 35 outweighs the benefits of better coordination of services when authority over such decisions is within the broader primary care program. The establishment of endoscopy ambulatory surgery centers is an example of units created through institutional re-organization that has been brought about, at least in part, to specifically address cost containment. These units, as opposed to the hospital or office settings, have been found to provide the best value for patients, physicians and payers, in part because of their adaptability (Frakes, 2002), a feature associated with self-contained units. A main challenge for this strategy is the acknowledged difficulties that health care organizations have with managing institutional change (Mercer, 2008). Thus, in re- organizing, three battles are faced: 1) selling the idea that change is necessary and beneficial and 2) identifying where integration is preferable and where devolution may be better and 3) selling the specific proposed change. 12. Strengthening of health care workers’ health and safety programs This strategy explicitly recognizes that compensation of health care workers is a major part of total costs.22 It also recognizes that diseases and injuries to health care workers represent a substantial ‘overhead’ cost for all health care organizations (Hudson, 2005). Therefore, this strategy suggests that changes in policies and practice, targeting specific work-related risks, can lead to significant cost savings. Hand-washing practices, for example, have been identified as a “crucial component of risk management for both health care workers and their patients” (Larmer, Tillson, Scown, Grant, & Exton, 2008). Accordingly, implementation of this strategy would involve ensuring that policies around hand-washing reflect current best-evidence. The savings come from two main sources: less absenteeism and reduced risk of infections for patients. Targeted risks are either risks of disease transmission (Mirtskhulava et al., 2008) or risks of injury in the performance of duties (Morgan & Chow, 2007). A key challenge in the implementation of this strategy is the recognition that staff safety is a holistic issue (Possamai, 2007). Therefore,  22 Labor costs as a proportion of total health care costs are estimated at 60 to 80% in Canada in 2007 (Canadian Institute for Health Information, 2007, p9). 36 attention must be paid to the system response to changes made in that some of these responses can nullify the intended impact (for example through the development of ‘shortcuts’ that circumvent new safety regulations but address time pressures). Classification framework Classification frameworks for these strategies were also noted. ‘Classification frameworks’ are defined as any methodology that results in the grouping of homogeneous strategies within categories. Very few such frameworks were found and in all cases, the categories seem to be defined for reasons of convenience (related to the analysis being undertaken) as opposed to being the result of a normative approach. For example, Ginsburg (2004) talks of four broad options (or categories of strategies) to contain costs: “one can increase the efficiency of health care delivery; increase the financial incentives for patients to limit their use of medical services; increase the administrative controls on the use of these services; or limit the resources available to the health care system” (p.1592). Cost containment strategies are categorized this way by Ginsburg because his analysis focuses on the use of incentives aimed at different agents in the system: the administrators, the patients, the physicians or the funders. In another example, Marmor, Oberlander and White (2009) break down the cost containment strategies into two categories: 1) Improving medical practices and health outcomes and 2) Restructuring the health insurance marketplace. This classification reflects a distinction between measures focusing on medical issues (e.g. prevention, chronic disease management) and those focusing on organizational factors because the authors believe that, for the U.S., addressing organizational factors is the only real solution to cost containment.  From the perspective of this thesis project, the most convenient classification criterion for healthcare decision-makers in Canada is the impact on services because this is a significant source of resistance to the implementation of a cost reduction or containment strategy in a publicly-funded system (see for example, Hansen 2009). Accordingly, it is proposed to organize the strategies in terms of their expected impact on the make up of 37 the healthcare services provided by an organization, from the perspective of the population served. Impact of strategies on services The first step in organizing the strategies according to their expected impact on services is to recognize that the implementation of cost reduction or containment strategies can have one of three consequences on the make-up of the services provided by an organization: 1) there could be no changes to the services; 2) there could be changes within programs (where programs are defined as a set of activities designed to address a specific health issue) which are changes in the way the programs are delivered, i.e. changes in the way the population, or patients, interact with the programs, for example, who they see, where they have to go, etc; or 3) there could be changes between programs i.e. decisions that affect the relative availability of the programs, for example, where one or some programs are discontinued or their accessibility is reduced. It must be noted that the programs affected in such a way (except for the cases of discontinuation) may, in turn, attempt to alleviate the impact on their patients or population by introducing changes in the way they are delivered (changes within programs). These three possible types of impact on services are suggested as our classification categories. Strategies classified by impact on services Category 1: No changes between programs or within programs This category includes cost reduction or containment strategies where the results can be achieved without any changes in the relative scope of programs or in the way patients experience, or interact with, the programs (i.e. with no changes, as we defined them, within programs). For example, if we use the case of a hip and knee replacement program, there could be cost reductions in this program without any within or between programs changes by increasing the efficiencies in some of the program processes. Such improvements could address scheduling practices for the operating rooms or for the use 38 of diagnostic equipment (assuming the appropriate share of overall diagnostic equipment costs is charged to the program), or streamlining of supply purchases. These changes are not changes in the relative scope of the program because they do not change the overall budget for the program. They also are not changes within the program because the patients still see the same providers, at the same locations and receive the same services from each as before. The result of such changes would be to perform more interventions within a fixed budget. We have an illustration of such cost reductions at a hospital in Richmond, British Columbia, Canada, where a high volume hip and knee replacement surgery center was established. Key changes are described as follows: “staggering operation start times and schedules between two rooms, so surgeons could “swing” between rooms as their patients were ready. This, in addition to coordination with surgical units post-op, allowed operating teams to complete eight joint replacements or reconstructions per day instead of three. Surgical procedures and clinical practices were standardized, using one single type of prosthetic device, which made work smoother for nurses and allowed the hospital to negotiate better deals on bulk purchases. Together, these measures were able to increase operating room efficiency by 25%” (Health Innovation Forum)23. Three strategies fit in this first category: 1-Supply management, 2- Investment in information technologies, and 3- Strengthening of health care workers health and safety programs. The changes to the hip and knee replacement program at Richmond Hospital that were described above are an example of supply management- these changes improved the productivity of some inputs and reduced the cost of other inputs. These three strategies are the only strategies that can lead to cost reductions without changes between programs or within programs. This does not mean that implementation of these strategies will never lead to such changes. This, in turn, means that the decision-maker will have to take care that the context (here it would be that no changes between or within programs are allowed) is reflected in the implementation plan for the chosen strategy. For example, investing in information technologies could lead to changes within programs or  23 We assume here that the prosthetic device selected was one of many devices of similar quality as there is no indication that this is not the case. If a device of lesser quality was selected amongst a wide range of possibilities, then that would make this a change in process. 39 even between programs, but the impact can be restricted to cost reductions without such changes. Category 2: Changes within programs In this category, we will have all strategies that can be expected to produce cost reductions through changes within programs, or changes in the processes that make up the programs. We have an example of this in Edmonton, Alberta, where two processes in the hip and knee replacement program were changed in what became a new model of care: the role of primary care physicians was increased and central assessments clinics were established (Alberta Bone & Joint Health Institute, 2008). These are “changes within programs” because they changed the way the patients experience the program; for example, the patients are now referred to newly created assessment clinics by their GPs, which means that they go and see somebody previous patients did not see. These changes did not, however, change in any way the relative scope of the program- those who were eligible before remained eligible, the intervention was still fully covered by Medicare and the total program budget was unaffected; these changes in themselves neither increased nor reduced the relative scope of the hip and knee replacement program. There are six strategies that can get to cost reductions through changes within programs without imposing changes between programs: 1-Incentives for providers, 2- Incentives for patients, 3- Compensation system for physicians, 4- Systematic use of cost- effectiveness analysis, 5- Re-structuring of clinical decision-making, and 6- Institutional re-organization focused on service delivery efficiency. The case of the Alberta Bone and Joint Institute represents an example of institutional re-organization where some roles were redefined and new service centers were created. There is also one strategy from the first category that can be used to reduce costs through changes within programs: Investment in information technologies. And, the same caveat must be added in this category of strategies: some of these strategies can lead to changes between programs unless they are specifically limited to changes within programs in their implementation. 40 Category 3: Changes between programs This final category includes all strategies that address costs through changes in the relative scope of programs- either increasing (including introducing new programs) or decreasing (including eliminating) the scope of some programs. If we stay with the example of the knee and hip replacement program, examples of between program changes would be changes to the eligibility criteria (related, for example, to age, repeat surgery, monitored waiting) or simply a relative change to the total annual funding available for such interventions24, which would result in changes to wait times. These examples would lead to some individuals not having the same access to the program anymore, reflecting a change in the relative scope of the program. Three strategies fit this last category: 1-Reductions in coverage, 2-Changing the mix of services currently offered, and 3-Increased emphasis on health promotion and prevention. The examples presented above for a knee and hip replacement program could be illustration of a strategy of reductions in coverage. There are also three strategies from category 1 and 2 that can be used to reduce costs through changes between programs: Incentives for providers, Investment in information technologies, and Re-structuring of clinical decision-making. The potential impact on health care services of each strategy, by category, is summarized in Table 4.  24 This means a change in the annual budget of the program that is different than the organization’s overall budget change. The “organization” in this context may be a provincial Ministry of Health or a regional health authority. 41 TABLE 4.  Potential impact on the make up of services of each strategy Strategic options for decision-makers From the perspective of the decision-maker, the categories of impact are cumulative. If a decision-maker is willing to accept changes within programs (category 2), he or she will gladly accept the same cost reduction in a form that is not impacting patients (category 1). Similarly, if changes between programs (category 3) are acceptable, similar cost reductions from within program changes will be gladly accepted because these will not involve reductions in the relative scope of any programs and a change in process is easier to defend than a restriction in eligibility. The trade-off for the decision- maker is clear: to access progressively more possible strategies, he or she must be ready to expose the population and patients to more significant changes. On one hand, more choices are preferable because each strategy attacks the problem from a different angle Cost reduction strategies Potential Impact on Services Category 1 Category 2 Category 3 Incentives for providers X X Supply management X Compensation system for physicians X Incentives for patients X Reductions in coverage X Systematic use of cost-effectiveness analysis X Changing the mix of services currently offered X Increased emphasis on health promotion and prevention X Investment in information technologies X X X Re-structuring of clinical decision- making X X Institutional re-organization focused on service delivery efficiency X Strengthening of health care workers health and safety programs X 42 and being restricted to fewer ‘angles’ increases the likelihood that a specific ‘well’ may have dried up. For example, there are only so many price concessions that can be obtained from suppliers through more aggressive negotiations in the context of a strategy of ‘supply management’. And program managers complain frequently of being down to the bone in terms of cutting out the fat. On the other hand, we know that change is seen, a priori, as a negative in health care management, so there are pressures to limit the impact on services to a minimum25. The framework proposed to organize cost reduction or containment strategies, as shown in Figure 2, can be used to inform the decision-maker as he/she wrestles with these opposing forces.  25 In the words of Michael Decter, in the chair’s message introducing the Health Council of Canada’s 2005 Report, “Historically, we have learned two hard lessons in health care. The first lesson is that change is difficult, even when all the evidence points to its necessity. Health care is a large and complex system. It is easier not to change than to change”.  (Health Council) 43 FIGURE 2  Strategic options available to decision-makers                                           Acceptable impact                Strategic options available                                                on services What kind of impact on the services provided would be acceptable at this time? Category 1 No changes within or between programs Category 2 Changes within programs but not between programs Category 3 Changes between programs -Supply management -Investment in information technologies - Strengthening of healthcare workers health and safety programs -Incentives for providers -Incentives for patients -Compensation system for physicians - Systematic use of cost- effectiveness analysis - Re-structuring of clinical decision-making - Institutional re- organization focused on service delivery efficiency - Reductions in coverage - Changing the mix of services currently offered - Increased emphasis on health promotion and prevention 44 Discussion Key messages This chapter provides an inventory of cost reductions and containment strategies as well as a framework where these strategies are organized in terms of their potential impact on services from a patient perspective. This inventory shows the depth of the existing knowledge on cost reduction or containment strategies in the health service research literature. It also shows the importance of choosing the right strategy (or strategies) for a specific context. Each strategy has a different rationale, different impact and involves different potential implementation challenges. Organizing the identified strategies as is proposed offers two important benefits: 1) it helps decision-makers looking to reduce or contain costs to focus on those strategies that suit their context in one very important respect, the acceptable impact on services; and 2) it informs the actual implementation of these strategies. For example, if a decision-maker determines that no changes between or within programs can be made in the current search for cost reductions, then supply management is one of only three options to consider. Alternatively, it would be a mistake for a decision-maker to turn to expanded use of cost- effectiveness analysis when changes between or within programs are not realistically possible. The context would interfere with the proper application of this strategy leading to disappointments and, often, inappropriate criticism as expectations would not be met. Potential limitations There are four potential limitations to this research phase. First, the search was not a systematic review and some specific strategy that may have been used successfully in a given context may have been omitted from the inventory. Second, each of the twelve strategies for cost reduction or containment represents an entire, evolving field of research on its own. In summarizing each in a few sentences, some details that could be important to decision-makers in their operational context were left out. Third, the 45 segregation of the strategies across the classification categories was only based on their potential impact on the make up of services. Doing so ignores other factors that may inform the selection of an appropriate strategy within a given category. Finally, possible implementation methodologies were not addressed, other than through examples. The purpose here was in outlining a framework for strategies related to cost reduction and containment. Future work could naturally address issues around implementation. Conclusion This chapter served two distinct objectives. First, it established that there is a strong knowledge base in the health service literature on the topic of cost reduction or containment strategies. The review presented above found twelve distinct strategies that have been proven effective in some real-life settings. Of those twelve, ten could be used to generate disinvestments, as we defined them in the introduction (i.e. only two strategies can only be used to generate efficiency gains). The implication of this finding is that the challenges around the identification of disinvestment options can not be explained by a lack of existing knowledge on how this task can be accomplished. The second objective was to compile the existing knowledge and organize it in a way that is convenient for decision-makers. This objective addressed the possible problem that decision-makers do not know of some, or most, of the existing knowledge on cost reduction or containment strategies. In this sense, the inventory of cost reduction strategies and the framework organizing them into categories can be useful tools for the decision-maker searching for disinvestment options while maintaining access to programs and services in which the public, patients, and/or staff may be heavily ‘invested’. 46 CHAPTER 3 EXPLORATORY RESEARCH ON THE TOPIC OF DISINVESTMENTS IN THE IMPLEMENTATION OF THE PROGRAM BUDGETING AND MARGINAL ANALYSIS PRIORITY SETTING PROCESS Introduction Despite more than thirty years of research on priority setting in health care and resulting processes that provide “a theoretically valid, ‘rational’, and systematic set of principles for conceptualising priority setting” (Bate et al., 2007, p.259), recent research in the NHS found that resource allocation decisions are still largely driven by political, historical and clinical issues (p.258). This finding has been confirmed by Peacock et al. (2009): “In recent literature, economic approaches to priority setting have been criticised on the grounds that they have only had limited success in practice”. Both groups of authors explain the lack of use and impact of formal, structured priority setting processes as the result of practical challenges to their implementation. Issues of challenges to implementation are not new to priority setting research and part of the work on this has focused on the identification of ‘success factors’ for formal priority setting in health care (e.g. Mitton & Donaldson, 2003; Mitton & Donaldson, 2004; Gibson et al., 2006; Peacock et al., 2007). As mentioned in the Introduction, this thesis project builds on the work on ‘success factors’ while focusing on one specific challenge to priority setting implementation: the identification of disinvestment options. In the previous chapter, we determined that the barriers around the identification of disinvestment options were not related to an absence of knowledge on how to do this. A review of the literature on cost containment or reduction strategies (strategies to generate 47 disinvestments) found ten strategies, besides different forms of efficiency gains, which have been used across different settings (see pages 28 of chapter 2). This suggests that the challenges around the identification of disinvestment options are related either to access to this knowledge or to the decision to use some of this knowledge. This chapter will report on four case studies of exploratory research. The objective of this research, building on the work of Bate et al (2007) was to generate a possible explanation for the level of challenges faced in identifying disinvestment options that is based on organizational behavior (where organizational behavior is defined as that aspect of individual responses to the demands of the PBMA process that is shaped by the organizational context). Thus the focus was not on adding to the list of success factors found in the literature but rather on better understanding how organizational behavior affects the extent to which these success factors come to bear on the results of PBMA. The PBMA process was implemented in four different settings with the exercises following standard procedures found in the literature (described in Appendix A, page 115). The analysis focused on the parts of the process related to the identification of disinvestment options, the decisions to act on those options, and specifically on the organizational behaviors that seem to facilitate or impede this task. Research questions The exploratory research questions for this phase are as follows: 1. For each of four PBMA case studies, what were the observed results in terms of changes to the resource allocation pattern and challenges to the identification of disinvestment options? 2. For each of four PBMA case studies, amongst known success factors, what were the key barriers or facilitators in identifying disinvestment options? 3. What organizational behavioral characteristics emerged as potential determinants of the challenges to the identification of disinvestment options? 48 Methods Context and settings PBMA was implemented in four different settings: Vancouver Island Health Authority (VIHA) across the entire organization; 7North Geriatric Assessment and Rehabilitation Unit in Victoria (part of VIHA); Mental Health and Addiction Services- South Vancouver Island (part of VIHA); and Menno Place. The settings involved were not targeted and the sample size was not pre-determined. This was a sample of convenience. Once the decision was made that the next research step would involve the implementation of the standard PBMA process, some implementation settings were required. To be eligible, a possible setting had to have some priority setting issues at the time of the research project and an interest in using a structured priority setting to address these issues. The four cases selected were the first four cases identified that met these eligibility criteria. The exploratory research was stopped at four cases because no other cases were identified at that time and the four cases were considered sufficiently different to provide a broad exploration of the issue of disinvestments. In each of the four case studies, my role was that of educator, advisor and, to some degree, participant. Over time, in each case, this role lead to becoming a trusted member of the team and, as such, being regularly consulted both formally (as part of internal team meetings) and informally (through phone calls, emails and ‘hallway chats’). In no cases did this role involve having a vote at any decision-making meetings. The role was very much that of a neutral advisor whose opinion on the method and nuances of process implementation was sought. VIHA is one of five regional health authorities in British Columbia. VIHA serves a population of approximately 750,000 and has an annual budget of about $1.6 billion, which does not include the bulk of physician compensation. PBMA was introduced at VIHA in the fall of 2005 and was part of the budget-making process for two consecutive years (budget years 2006/07 and 2007/08). VIHA is comprised of two organizational structures: Integrated Health Services, and Corporate and Strategic Services, both under 49 the authority of a single Executive Team. Integrated Health Services includes eight portfolios while Corporate and Strategic Services has seven portfolios (a portfolio typically includes a number of related departments). VIHA has a dual leadership structure with each portfolio managed by an executive director and an executive medical director and each department having a director and a medical director. The composition of each structure is presented in Appendix B. PBMA was implemented at VIHA to provide guidance on overall resource allocation decisions as part of ongoing efforts to ensure the highest quality management of the scarce resources available. Following standard PBMA procedures (described in Appendix A on page 115), each portfolio across the organization was required to submit both disinvestment and investment proposals (i.e. identify disinvestment and investment options) to achieve the aim of re-allocating resources to better meet pre-stated organizational objectives. 7North is a 36-bed in-patient geriatric assessment and rehabilitation unit within VIHA, located in Victoria. This unit falls under Integrated Health Services and resides in one of its eight portfolios (Continuing Health Services). Within Continuing Health Services, the unit is the responsibility of the Seniors’ Health Department which is managed jointly by director and a clinical director. PBMA was implemented at 7North starting in February 2008 with the specific aim of assisting in the possible redeployment of geriatrician services within the Victoria region. The basic problem was a large variation in access to geriatricians in the Victoria area. Specifically, seniors treated at the 7North unit had much better access to geriatrician services than seniors accessing these services elsewhere. Increasing the number of geriatricians FTEs in the region was not an option, leaving the redeployment of existing capacity as the only avenue for change. Despite the recognition of the access problem, the management of Seniors’ Health had been unable to reach an agreement with geriatricians on possible solutions, which led to the decision to implement PBMA. The PBMA implementation at 7North followed standard procedures with one context-specific limitation: the leadership of Senior’s Health had stated that the only possible disinvestments were geriatrician FTEs at 7North. No other reductions in services were to be included in the process. 50 Mental Health and Addiction Services (MHAS) is a department within Continuing Health Services, which is one of the eight portfolios of Integrated Health Services at VIHA26 (see Appendix B). MHAS - South Vancouver Island is responsible for all programs in the southern part of Vancouver Island that are related to mental health, whether the conditions are caused by accidents, diseases or addictions. The programs range from prevention to treatment and cover a wide range of interventions from sobering and assessment centers to dementia consultation to emergency psychiatric consultations. The program is led by a director and a medical director. The PBMA implementation at MHAS also followed standard procedures although it too had one key context-specific difference: in this case no disinvestments of any sort (i.e. no service reductions) would be considered. As such, the goal of the implementation was to produce a ranked list of spending initiatives. Leadership within MHAS felt that new uncommitted funds might be made available thus they wanted to have ready a ranked list of spending proposals. Despite this, this case is included in this chapter because the decision to forgo, at the outset, consideration of any disinvestments can provide some insights into the challenges of identifying disinvestment options. In this case the challenges are not around the implementation of PBMA per se but rather relate to the broader context of decisions leading up to a PBMA exercise. Menno Place is the combination of Menno Housing, Menno Home, and Menno Hospital. Menno Place is located in Abbotsford, British Columbia, and is under the jurisdiction of the Fraser Health Authority. PBMA was introduced at Menno Place in August 2007 in order to assist in the development of a deficit reduction plan for Menno Home and Menno Hospital. These two components of the organization had been running on and off in a deficit situation for the last three years and the accumulated deficit was reaching a size that was becoming a concern to the Mennonite Benevolent Society which operates Menno Place. The total budget for Menno Hospital for 2007/08 was $12.49 million with a projected deficit of $191,000 which would lead to an accumulated deficit at the end of the period of $545,000. For Menno Home, the budget was $11.7 million, the projected deficit was $355,000 and the accumulated deficit was $387,000. So, combining  26 The department has since been moved to a portfolio named Population and Community Health. 51 the home and the hospital, the forecasted deficit was $546,000 on a total budget of $24.2 million, which would lead to an accumulated deficit of $932,000. Again, a standard approach to PBMA implementation was taken. Data collection and analysis Available data varied by case and can be broken down into two types: process data including documentation and observation, and decision maker interviews. The first type of data was attained from documents that accrued during the process and observations of the process. Throughout each of the four process implementations, all documentation and materials related to the disinvestment procedures were collected. This included documents describing disinvestment procedures, disinvestment proposals or options and memos summarizing the results of proposal ratings and publicizing decisions. Documents were used to determine what happened in each case in terms of disinvestment options identified and acted upon- the documents provided factual records of the results of the implementation in terms of disinvestments. Attending PBMA meetings either by phone or in person provided opportunities for observation in each of the case studies. Observations provided original data as well as opportunities to check the findings from the document review. For analysis, information from relevant documents and observations was organized on the basis of known success factors for PBMA implementation (specifically: culture to learn, strong leadership support, organizational stability, high level champion for the process, having resource allocation decisions that must be made, and strong relationship between physicians and managers) using the framework presented in Figure 1. The factors that emerged in this analysis as most relevant for each case are reported in the results. Data that did not fit those success factors were classified under ‘other organizational context’. The second source of data was formal evaluations conducted at VIHA (with respect specifically to the VIHA implementation across the entire organization) and at Menno Place. The evaluations were based on interviews conducted once process implementation was completed. There were two evaluations at VIHA (one after each of 52 the two years of implementation). In the first evaluation, 22 decision-makers were interviewed. For the second year, there were also 22 respondents, which included seventeen from the first year. The evaluation at Menno Place involved nine decision- makers. In both organizations, participants were purposively selected to provide a broad range of health care management experience (i.e. both clinical and managerial leadership), varying levels of authority within the organization and differences in the degree of involvement in the process. Evaluation interviews were semi-structured, lasted about 45 minutes each and were audiotaped. The interview contents were coded using a thematic analysis approach (Braun &Clarke, 2008). The primary researcher performed the coding. Summary findings from the thematic analysis were reviewed by some interviewees, a form of feedback and validation often referred to as member checking (Schwartz-Shea, 2006, p.104). All research work at VIHA and Menno Place was approved by the UBC Behavioural Ethics Board. Results The results are presented in the order of the research questions, by case study. All data available (i.e. process data and interview answers) were combined in addressing each question and the results summarize the overall findings for each case. Observed results in terms of changes to the resource allocation pattern and challenges in the identification of disinvestment options 1) Vancouver Island Health Authority (across the entire organization) PBMA was implemented at VIHA across the entire organization for fiscal years 2006/07 and 2007/08. In the first year, 54 disinvestment proposals were submitted with a total value of $15 million (all figures are rounded) and in the second year there were 13 proposals worth $10 million. The target for disinvestment proposals in the first year was 53 2% of the previous year’s budget or about double what was submitted. The amount of resources actually re-allocated was $14 million in the first year and $4.5 million in the second year. The extent of the impact of PBMA can be described this way: over the two years of implementation, about one half of one percent of the spending at VIHA was allocated differently because of the PBMA process. There were some significant challenges in the identification of disinvestment options at VIHA. “It’s an impossible task, putting up some possible resource releases for us right now. It’s an impossible task because there are wait times for all areas of our portfolio”. (From the leader of a medicine department) The results in terms of the nature of the proposed options were disappointing to some: “We did very poorly, I think, as an organization in identifying disinvestments, it was much, much of the old stuff we’d heard for years of decreasing overtime…” “I think the other dimension to this is the disinvestment process and that’s where people were to bring forward potential savings in their portfolios. In my opinion it was very much the same sorts of things that people offered in previous years in terms of budget savings such as reductions in overtime or efficiency in this area or in that area…” The challenges were such that in the second year the formal requirements for disinvestment proposals were removed: “There is no blanket requirement or target for programs to submit resource release or revenue generation opportunities this year.” (From instruction memo for year two of PBMA, December 1, 2006) 54 2) 7North Geriatric Assessment Unit At 7North, the impact of the PBMA process on resource allocation was very simple: no changes were made to the way the unit operates and therefore there was no impact on the resource allocation pattern. The PBMA process implementation resulted in one specific recommendation that involved a reduction of the role of geriatricians at 7North and a redeployment of their services to other settings27. This recommendation was ultimately rejected by the Executive Team.  In terms of disinvestment options, since the process at 7North was limited to the assessment of reallocation options for geriatricians and the form of the possible reallocations being considered had been discussed before, the development of the disinvestment options was a straightforward task. 3) Mental Health and Addictions Services (MHAS)- South Vancouver Island Given that the PBMA process, as implemented at MHAS, focused on investments and did not consider disinvestments, the impact of PBMA on resource allocation patterns was settled at the outset: there would be no immediate impact because no new money was available and no re-allocations could take place. Should new money be available, there would be an impact as the ranked list of investment initiatives would then be funded according to the level of funds available. The identification of disinvestment options posed no direct challenges because disinvestments were excluded from the PBMA process. However, the full assessment of the challenges to the identification of disinvestment options at MHAS must address the reason why disinvestments were excluded from the PBMA implementation. Disinvestments were excluded because leadership of MHAS made it clear that they believed absolutely no reductions to services would be accepted by the Executive Team. This belief can be traced directly to an earlier refusal by the Executive Team to endorse a  27 The recommendation was that geriatricians act only as consultants with the patients of 7North while the Most Responsible Physicians would be a family physician acting as a clinical associate, (likely two family physicians would be necessary) paid on a service contract. 55 proposed re-allocation of funding across geographical regions for some of MHAS’s services. The proposal was a managerial initiative of MHAS designed to improve the overall regional distribution of its services. The plan involved reductions to some services in South Island (despite the acknowledgement that theses services were insufficiently funded to deal with the workload) and increases in the volume of some services in Central and North Island where the funding deficiencies were considered worse. The plan had been developed after extensive consultation with stakeholders and involved difficult choices. 4) Menno Place Ninety-three disinvestment proposals were submitted at Menno Place with a total value of approximately $888,000. All the disinvestment proposals were ultimately approved. Of this amount, approximately $480,000 was re-allocated to new investments and about $408,000 went to deficit reduction (leaving a forecasted deficit of about $138,000 for fiscal year 2008/09 as opposed to the original estimate of $546,000). We can assess the magnitude of the process’ impact on the resource allocation pattern in two ways. First, the value of the implemented disinvestments was about 3.5% of total spending. This means that about 3.5% of the money that would have been spent had services stayed the same was not going to be spent the same way this coming year. In fact, close to half of this money was not going to be spent at all. And second, approved disinvestment proposals covered a wide range of programs and were not focused on support programs28 or some other specific area of services (i.e. changes were widespread throughout the organization). Examples of approved disinvestments are provided in Appendix C. Approved disinvestments covered such disparate areas as nursing hours, nursing staff composition, recreation programs, food services and equipment management.  28 Disinvestments in support programs are often efficiency gains because they typically do not involve a change in clinical services (or at least the intent is to not affect clinical services). If most disinvestments are in support services, the PBMA process simply becomes a cost cutting process where the objective is ‘doing more with less’. 56 The disinvestment proposal process was open to all staff at Menno Place and there was widespread participation in the process. The process met the targets for disinvestment option values without any stated, or observed, significant challenge. In summary, the impact on resource allocation was very limited at VIHA for the organization-wide case study, relative to the size of the overall budget, null at 7North, not applicable at MHAS and significant at Menno Place both in relative terms compared to the budget and in terms of the range of changes implemented. There were significant challenges to the identification of disinvestment options at VIHA, no apparent challenges at 7North, no search for disinvestment options at MHAS and no significant challenges at Menno Place. So, while the same process was implemented in each of the four settings, and with the help of the same advisor in all cases, the observed results varied considerably between cases, both in terms of impact on the resource allocation pattern and the degree of challenge faced in the identification of disinvestment options. Key barriers or facilitators to the identification of disinvestment options Organizational integration has been a broad area of focus within research on priority setting in health care. One specific aspect of this area has been the identification of success factors for priority setting, or those organizational features that seem to be present when priority setting is successful (i.e. when disinvestment options are identified and acted upon). Many such success factors have been documented (for example: culture to learn, strong leadership support, organizational stability, high level champion for the process, having resource allocation decisions that must be made, and strong relationship between physicians and managers) and this section will identify those that have emerged as key facilitators (i.e. had a strong presence) or key barriers (i.e. were largely absent) for each of the four cases. 57 1) Vancouver Island Health Authority- across all programs The main facilitators at VIHA were the strong support of the process by the Chief Operating Officer (COO) and organizational stability. The COO was a vocal advocate of the process: “The thing is I’m not sure what else is out there that’s even close to being better (than PBMA)…any process has limitations…but the fundamentals (of PBMA) are incredibly sound.” The COO was observed repeatedly at PBMA meetings explaining how managers have to put on their ‘corporate hats’ when dealing with PBMA and not concentrate narrowly on concerns specific to their departments. This is particularly relevant to the identification of disinvestment options because these options are proposals that are expected to have a relatively limited impact on corporate objectives, not on the specific objectives of a given department. Therefore a ‘corporate hat’ changes the search and may well end up making it easier (at a minimum, it increases the number of possibilities). In terms of encouragement and specific instructions, strong support from the COO emerged as an important facilitator for the identification of disinvestment options for the PBMA process across the entire VIHA organization. The other key facilitator identified in the analysis was corporate stability. The search for disinvestment options typically reflects a novel approach to resource management and it has been found that organizational stability (especially amongst senior staff) facilitates its integration into the organizational management culture (Donaldson, 2001). During the two years of PBMA implementation at VIHA the Executive Team stayed the same save one change (the Chief Medical Officer). The main barrier to the identification of disinvestment options seems to have been the lack of a formal requirement to change the resource allocation pattern.  PBMA was brought in at VIHA as part of ongoing efforts to improve the management of the scarce 58 resources available. However, there was no obligation to either justify the current service mix or change it, which means that there was no outside force that addressed the natural resistance to change. And since disinvestment options are proposals for change, they are directly affected by this lack of pressure to act. This ended up being a significant barrier because it effectively removed any significant changes to services from the possible targets for disinvestment options, leaving managers to search for disinvestments that have little impact on services (which is a difficult search). 2)  7North Geriatric Assessment Unit In the case of 7North, there initially appeared to have been no challenges to the identification of disinvestment options because the scope of the process was about redeploying geriatricians away from the unit and no other service reductions at 7North were to be considered. The disinvestment options involved different levels of reduction to the geriatrician services offered at 7North and flowed directly from the context in which PBMA was being applied. However, as was mentioned above, no changes were made to the operations of 7North as a result of the PBMA process. In the end, the recommendation from the process was rejected because the Executive Team was not willing to get into a fight with the geriatricians. Since all the disinvestment options involved reductions to geriatrician services at 7North, none of them would have been acceptable to the Executive Team because geriatricians were opposed to any reductions in their services at this unit. Therefore, while on the surface there appeared to have been no challenges to the identification of disinvestment options at 7North, it was found after completion of the process that there seemed to have been barriers to the identification of ‘acceptable’ disinvestment options (acceptable to the Executive Team). And the main barrier seems to have been the poor relationship between physicians and managers. In the context of this relationship, the PBMA process became a testing mechanism for the validity of each side’s claim (geriatricians claimed that no changes should be made and management had the opposite view). Therefore the poor relationship between physicians and managers caused the process to focus on the disagreement between the two groups 59 and to isolate itself from the broader organizational perspective which ultimately affected negatively the identification of disinvestment options 3) Mental Health and Addiction Services At MHAS no disinvestment options were included in the process. There was no attempt to identify any possibility of saving resources. Interestingly, this was the scope of the process despite the observation that most success factors for PBMA were present at MHAS. There was the strong support of VIHA’s COO for PBMA, organizational stability, a culture to learn and an openness to change in the department, a strong relationship between physicians and management and a strong desire to change the resource allocation pattern. So the decision to exclude disinvestments from the process could not be explained by the known success factors and none of them (i.e. their absence) can be considered a significant barrier in this case. 4) Menno Place The main facilitators in the search for disinvestment options were the strong leadership support, a high level champion, and the desire to fix the financial problems of the organization. PBMA was introduced at Menno Place by the new COO. This was one of his first major initiatives. Because he had just been selected for the position by the Chief Executive Officer (CEO) and the Board, he enjoyed at that time their full support. Therefore PBMA was openly supported by the Board, the CEO and the COO. At the initial PBMA training session, the Chairman of the Board and the CEO said a few words which publicized this endorsement. The COO quickly became a champion for the process and vocally defended it whenever necessary. At one point, the director of Support Services announced that he refused to participate in the search for disinvestment options, preferring to manage his own budget independently as he had done for years- as opposed to the organization-wide approach of PBMA. This director threatened to resign if forced to comply with the 60 disinvestment requirements and the COO decided to accept his resignation. This decision sent a strong signal through the entire managerial ranks. A third facilitator was the impetus to address the deficit situation of the organization. There was no formal requirement for Menno Place to eliminate its deficit or reduce it by a set amount. However, the deficit situation was well-publicized and there was general acceptance in the organization that it had to be addressed which means that there was an openness to changes to the resource allocation pattern. The main barrier at Menno Place was the organizational instability.  The arrival of the new COO heralded a reorganization of the leadership of the organization. The Support Services director was the first to leave. He was followed by the director of Resident Services (by far the largest department in the organization) and then a new position of Chief Financial Officer was created. Therefore there was significant turnover of those people directly in charge of the resource allocation decisions. In summary, amongst known success factors for PBMA, those that emerged as key facilitators or barriers to the identification of disinvestment options provide valuable insight in explaining the degree of challenge experienced in each case in the accomplishment of this task. However, the differences in observed results are not mirrored by differences in success factors. At both VIHA and Menno Place there was strong leadership support for the process. VIHA had organizational stability which Menno Place did not have. On the other hand there was no formal requirement to change the resource allocation pattern at VIHA while there was a definite impetus to move in that direction at Menno Place. So while the analysis of known success factors helps understand how the search for disinvestment options unfolded in each case, it is not sufficient to explain the differences between the cases at VIHA (where significant challenges were found in each case) and the case of Menno Place (where this task did not face significant challenges). 61 Organizational context factors that emerged as potential determinants of the challenges to the identification of disinvestment options Themes from the data analysis that were not directly related to the ‘known success factors’ for PBMA were classified under ‘other organizational context’. The dominant themes in this category are presented by case study. These represent the most significant ways in which the organizational context affected the PBMA implementation other than directly through the known success factors. 1) Vancouver Island Health Authority (across the entire organization) The organizational context theme that emerged most strongly at VIHA was the apparent disagreement about the nature of possible disinvestment options between clinical leaders and managers (as mentioned above, all portfolios and departments at VIHA have dual leadership with a director and a clinical director). For example, one physician leader described the search for disinvestment options this way: “There is a sense that we are trying to get blood from a stone, it’s very challenging to do that (identify resource release options) without big service cuts.” Implied in this comment was the belief that big service cuts were not something that could be done, which meant that this physician was saying that this was a very frustrating task. In fact, this respondent displayed frustration in talking about the search for disinvestment options. This statement can be contrasted with this description, by a director, of how disinvestment options were identified. “We did have a portfolio discussion about the disinvestments to make sure we got to the right number, which was the 2% cut across the board …where our 2% was going to come because it wasn’t coming equally from everybody and decided 62 what we felt was safe to put forward and what was not safe. We decided what we would put forward and what was foolish to put forward.” This director speaks of ‘safe’ or ‘foolish’ proposals, which implies some kind of reference against which a possible option is assessed. This statement also suggests that it was possible to identify sufficient disinvestment options that were ‘safe’ in terms of this reference. This example was repeatedly observed. Physicians thinking that disinvestments could not involve service reductions because most or all services were operating at full capacity and beyond and managers that wanted to challenge service volumes. Given the dual leadership structure, this disagreement on what can be a disinvestment option likely affected the level of challenge to the identification of such options. 2) 7North Geriatric Assessment Unit At 7North, in trying to go beyond the lack of collaboration between geriatricians and managers in the development of disinvestment options, the organizational context factor that emerged as a possible root cause of the challenges in the identification of ‘acceptable’ disinvestment options was the lack of communication between the Executive Team and managers of 7North regarding the type of service reductions that could be considered in priority setting activities (i.e. that could ultimately be endorsed by the Executive Team). In the course of the process, there was no challenge to the nature of the disinvestment options that emerged. However, as it turned out, the position of 7North management on this issue was not aligned with that of the Executive Team. When asked how this came about, the leadership of Seniors’ Health explained that they suspected the Executive Team would not want to challenge the geriatricians. However, they added that they were hoping that if the recommendation to make changes to geriatrician services came as a result of a PBMA process implementation, the assumed reluctance of the Executive Team to support such a change would abate. Therefore, in the end, it seems that the leadership of Seniors’ Health misinterpreted, or misjudged, the Executive Team’s steadfastness in avoiding conflicts with geriatricians. 63 3) Mental Health and Addiction Services The situation at MHAS with regards to organizational context factors affecting the identification of disinvestment options is very similar to that at 7North. On the surface, the issue of the challenges to the identification of disinvestment options seemed irrelevant to the PBMA process at MHAS. On further investigation, it was discovered that the leadership of MHAS embraced the concept of re-allocation of resources and the implied requirement for disinvestments. However, the leadership reluctantly concluded that no reductions of any sort to mental health or addiction services were acceptable to the Executive Team because of the political sensitivity around such changes. In follow-up questions on this issue, VIHA’s COO stated that there was no overarching ban on reductions in mental health and addiction services but that the specific reductions involved in the plan previously proposed by MHAS were not acceptable to the Executive Team. This explanation by the COO points to a similar barrier at MHAS as was observed at 7North: a lack of communication between the leadership of MHAS and the Executive Team on what service reductions would be acceptable to the Executive Team. This lack of communication resulted in the abandonment of the search for disinvestment options at MHAS. 4) Menno Place At Menno Place, the search for disinvestment options can be considered successful as targets were met and there were no concerns expressed in the evaluation interviews with the quality of the options or with the difficulties in accomplishing this task. The main organizational context factor that emerged as a facilitator was the high level of buy-in or commitment to the process through-out the organization. The evaluation found unanimous support for the process amongst respondents- which included the ex-head of Support Services (who, in the end, endorsed the process despite admitting that he could not work within it). An important contributor to the commitment level was the transparency of the process through, for example, the inclusion on the 64 PBMA Advisory Panel of the union and association representatives. Importantly, with regards to the identification of disinvestment options, the transparency of the process reflected an openness to the changes that could result from the process. For example, there was no possibility in the process, as it was applied at Menno Place, for senior management to reject a disinvestment option because it was ‘out of scope’ or involved a change that it was not willing to accept and defend. Because proposal submission was opened to all staff, a formal feedback process was instituted. Proposals were returned to staff within a week with one of three responses. Proposals were either considered eligible for business form development, returned for additional work (clarification needed) or rejected because they were not feasible (and why they were not feasible was explained). Feasibility was defined only in terms of obligations of the organizations, for example, labor laws and contracts, regulations or guidelines from Fraser Health Authority. In summary, the other organizational context factors that emerged as affecting the level of challenges in the identification of disinvestment options in the four cases seem to revolve around the definition of the type of disinvestment options that are acceptable to senior management and/or the communication of this information. In the three cases at VIHA, there appeared to have been either misconceptions or disagreements (between managers and senior management or amongst managers) as to what is acceptable. At Menno Place, the search for disinvestment options seems to have been perceived as a wide open search by staff and managers and such a search seems to have been encouraged and supported by senior management in that the process, as was applied there, did not make allowances for rejecting proposals because they were not what senior management was prepared to implement. Discussion The PBMA implementations in the three settings at VIHA and at Menno Place led to very different results in terms of the impact on resource allocation patterns. In the two years of implementation across all programs at VIHA, the sum of all resource re- allocations from the process amounted to about one half of one percent of the total 65 budget. At 7North and MHAS there was no impact on resource allocation. On the other hand, the process recommendations at Menno Place resulted in changes to 3.5% of the expenditures in the first year of implementation. There were also significant differences in the challenges faced in the identification of disinvestment options, with these challenges being more prominent in the VIHA cases. This is not surprising as we expect that greater challenges in the identification of disinvestment options would result in less impact on resource allocation for the process. The known PBMA success factors provide useful information in explaining the degree of challenge in the identification of disinvestment options and the process impacts observed in each case. However, they do not provide a complete explanation. Or, put another way, they are not sufficient to provide a plan to facilitate the identification of disinvestment options in different settings- there is no common thread across all cases. To further inform the search for determinants of the degree of challenges in the identification of disinvestment options, the analysis turned to ‘other organizational context factors’ that shaped the behavior of decision-makers with respect to PBMA implementation. In each of the cases, the ‘other organizational context factor’ that emerged as the main determinant of the level of challenges to the identification of disinvestment options seemed to be related to the degree of communication between senior management and managers on what can be a disinvestment option. Communication was poor in the VIHA cases and seemed to be strong at Menno Place. It is not possible to determine on the basis of the information available whether the lack of clarity at VIHA on what can be a disinvestment option was due to a lack of message (i.e. whether senior management knew from the outset what could be acceptable or was planning to make decisions as proposals were submitted) or whether this was simply a case of poor communication. The objective of the exploratory research was to generate some hypotheses relating organizational factors to the challenges in the identification of disinvestment options. However, because issues around the communication of acceptable disinvestment options emerged as the dominant theme in the analysis of other organizational factors and because evaluating more than one hypothesis at a time in a real life setting could 66 confound the results, a decision was made to focus on one hypothesis at this time. In this context, the results presented clearly point to a possible connection between clarity regarding what can be a disinvestment option and the degree of challenges that can be expected in the search for disinvestment options. This possible connection can be expressed as the following hypothesis: The determination and communication of acceptable service reductions, at the outset of process implementation, facilitates the identification of disinvestment options and increases the likelihood that any of these options will be acted upon. The main limitation to this phase of the research project is the fact that it is based on only four PBMA implementations and that, because of their specific context, these four implementations were different in nature. One implementation dealt with an organization with a $1.6 billion budget. One dealt with only one specific type of disinvestment. One did not include any disinvestments. And finally one was done in an organization with a $25 million budget and some significant financial challenges. In all likelihood, there are other possible explanations for the observed differences in the impact of the process and the challenges to the identification of disinvestment options. For example, one might suggest that the financial difficulties at Menno Place were a key factor in the observed successes. However, it must be noted that there was no formal requirement at Menno Place to reduce or eliminate the deficit. There was a desire to do so, just as there was a desire by the leadership of 7North to improve the allocation of geriatrician services. Interestingly, building on the success of PBMA at Menno Place, the Board of Directors developed and implemented a debt elimination plan starting with the next budget year. This plan included formal requirements to produce a specific annual surplus for the next four years, to be applied to the debt until it is eliminated. On the logic that financial obligations are a key facilitator for the identification of disinvestment options, one would expect at least as good successes in year two of PBMA at Menno Place given that the financial obligations were even stronger (PBMA has been implemented at Menno Place each year since the original case study was conducted). Results proved quite different. The year two process fell apart because of a lack of 67 disinvestment options. The process had to be rescued at the last minute. An analysis of what happened in year two at Menno Place found that the establishment of the debt elimination plan negatively affected other success factors (notably strong leadership support and buy-in) which on balance led to much more significant challenges to the identification of disinvestment options to the point where the targets were not close to being met29. It must also be added that while there likely are other possible explanations, the overall objective of the research project is to provide new information that will facilitate the use of priority setting processes in health care, therefore the emphasis is on explanations that involve organizational factors that can be acted upon, unlike the financial situation of an organization which is what it is found to be (one could hardly suggest worsening the financial situation to facilitate the identification of the disinvestment options needed to address the now poor financial situation). While the variability in the nature of the settings and the scope of implementation across the cases is a limitation to this research, it is also a justification for the use of case study research in this thesis project as it illustrates how context dependant PBMA implementation is in real life. This intrinsic variation across PBMA implementations, in itself, favors a case study research design over an experimental research design. In addition to the limited number of case studies and the variability in the context of the implementations, there are other potential limitations to this exploratory research. For example, VIHA and Menno Place are two organizations of vastly different size. It could be that the problems with communication around disinvestments at VIHA can be explained in large part by the challenges around the communication of new ideas in such a large organization. Further, the analysis was done after the data collection so it was not possible to investigate specifically the causes of the problems in communication as this issue only became clear after data analysis. However, despite these limitations, the  29 Another recurring theme in the investigation of the problems with year two at Menno Place was that knowing in advance, in the form of a formal target, how much money had to go to the debt acted as a disincentive to identify disinvestment options because it was pre-determined that a large proportion of the disinvestments would be lost to operations (i.e. would go to the debt instead of re-allocations however productive the investment options were). 68 hypothesis generated offers a plausible explanation for the differences between the cases in implementation results and in the challenges faced. Moreover, the suggested linkage between any form of clarity or certainty around what is acceptable as a disinvestment option and the degree of challenge in the identification of disinvestment options has a theoretical underpinning based on an application of Herbert Simon’s bounded rationality theory (Simon, 1990). As this theory suggests, in the absence of instructions regarding what disinvestments would ultimately be acceptable, participants in the PBMA process are expected to make their own rational determination, as they do for all factors where they have incomplete information. Therefore, we have what we can call the ‘real’ range of acceptable disinvestment options, or service reductions (what the leadership of the organization is really prepared to accept and publicly defend) and the ‘interpreted’ range of such reductions (what managers rationally feel the leadership is willing to accept and defend based on the information available and other considerations, as necessary). This has two effects. First, in their search for disinvestment options, managers will only look within the ‘interpreted’ range of acceptable service reductions. This range tends to be very limited and only include services that are ‘under-used’ or changes that do not affect services directly, like reductions in overtime, as it was in the case of VIHA across all programs (or even worse in the case of MHAS). A limited range makes the search often frustrating as managers are searching for something that barely exists (i.e. once it is ‘rationally’ established that one is to search for under-used services, how many of those are there?). Second, only the proposed changes that fall at the intersection of the two ranges will ever be implemented and the more participants are left to develop their own interpretation of what is acceptable, the greater the likelihood that this intersection is small (the PBMA recommendation at 7North fell outside of this intersection, for example)30.  30 Of course, if the ‘real’ range of acceptable disinvestments is very small, then the intersection with the ‘interpreted’ range will also be very small even if there was a perfect match. However, one would wonder why PBMA would be implemented in such a situation given that this is a process focused on re-allocations (Mitton & Donaldson, 2004b, p.69). 69 The next step in this research is to prospectively evaluate the hypothesis generated in this phase of the project. The strategy selected to do this was to incorporate the hypothesis in an application of the PBMA process and analyze its impact on the process. The simplest way to incorporate the hypothesis in the process is to translate it into a specific action that is part of the procedures. As PBMA stands now, there is no step that calls for a determination of acceptable service reductions at the outset of the process in any of the seven stages (presented in Appendix A). The closest that the process comes to addressing this issue is in stage one: “Determine the aim and scope of the priority setting exercise—Will the analysis examine changes in services within a given program or between programs?” One would expect that the decision as to the aim and scope of the PBMA exercise would be based at least in part on the acceptability of changes to programs however the description of this stage clearly puts the focus on re-allocations across programs and the acceptability of that kind of re-allocations. The literature review on PBMA did not provide any instances (case studies or theoretical papers) where this stage of the process has been interpreted to include even a general assessment of acceptable service reductions. The hypothesis can be incorporated in the PBMA process by amending the first stage so as to include a determination and communication of acceptable service reductions as part of this stage. The revised stage 1 would then be expected to lead to some clarity throughout the organization, at the outset of the process, regarding where to look for disinvestment options (i.e. what can be a disinvestment option). And this clarity would be expected to facilitate the identification of disinvestment options and lead to more of these options being acted upon by leveraging existing success factors. This means that this clarity would be expected to lead to PBMA having a greater impact on resource allocation for any given level of success factors observed. Conclusion The strategy for this research project is to build on the existing work on success factors for PBMA and to do so by focusing on the challenges around the identification of disinvestment options. The exploratory research presented in this chapter has produced a 70 novel hypothesis relating organizational factors to the challenges in the identification of disinvestment options.  This phase has therefore moved the research project ahead while confirming an important assumption that underpins the strategy: in the case where the identification of disinvestment options was the most successful, the observed impact of the process on resource allocation was the most significant. The next phase of the project is to evaluate the new hypothesis in a real life setting. 71 CHAPTER 4 IMPLEMENTATION OF A MODIFIED VERSION OF PROGRAM BUDGETING AND MARGINAL ANALYSIS AT ST. JOSEPH’S HEALTH CENTRE, TORONTO Introduction Practical challenges to the implementation of priority setting processes in health care have been the subject of much research in the form of case studies and literature reviews (e.g., Mitton & Donaldson, 2003; Mitton & Donaldson, 2004; Gibson et al., 2006; Peacock et al., 2007). This work has advanced our understanding through the identification of ‘success factors’ related to implementation of formal approaches and uptake of resulting recommendations for resource allocation. Nonetheless, practical challenges remain and continue to limit the use of formal priority setting approaches in health care organizations in most countries. It was in this context that exploratory research in the Vancouver Island Health Authority (one of five regional health authorities in British Columbia) and at Menno Place (a seniors’ care facility in Abbotsford, British Columbia) was conducted. This research involved the implementation of the Program Budgeting and Marginal Analysis (PBMA) priority setting process (the process is described in Appendix A) and focused on a well-known, practical challenge to implementing PBMA: the identification and realization of disinvestment options (Cohen, 1994; Mitton & Donaldson, 2003; Mitton & Donaldson, 2004b). Observations and interviews were used to generate data that allowed for an in-depth analysis of each site as well as comparisons across the sites. 72 Importantly, the observed differences in the results of the PBMA process across the case study sites were not fully explained by variations across sites in key success factors reported in the literature, such as leadership and organizational buy-in. However, one prominent organizational feature that was present in the more successful case but absent in the less successful ones was the determination and communication, at the outset of the process, of the areas of the organization where service reductions were acceptable. This observation pointed to a possible connection between clarity regarding areas of acceptable service reductions, the ability to identify meaningful disinvestment options (or develop disinvestment proposals), and the extent to which the priority setting process affects actual resource allocation decisions. Building on this exploratory research, the work described in this chapter sought to explore the potential connection identified in the exploratory research phase of the thesis project. Specifically, this chapter investigated whether determination and communication of areas of acceptable service reductions, included as a formal step in the PBMA process, can facilitate the identification of disinvestment options. Such an approach has not been previously reported in the priority setting literature. The research described in this chapter involved implementation of the PBMA process with the determination, at the outset, of areas of acceptable service reductions. The scope of acceptable reductions was then communicated widely across the organization. The process took place at St. Joseph’s Health Centre in Toronto in 2009. St. Joseph’s was selected because the researcher became aware of the interest of senior leadership in PBMA at the time when a site for a real life implementation of the process was needed. As such, St. Joseph’s met the eligibility requirements of having priority setting issues and an interest in using PBMA to address these issues. 73 Research questions The research questions are as follows: 1. As part of a PBMA implementation at St. Joseph’s Health Centre that included the requirement for an explicit determination and communication of acceptable service reductions at the outset of the process, were proposals for disinvestment identified and acted upon? 2. In the context of this PBMA exercise, how challenging did decision-makers at St. Joseph’s Health Centre find the task of identifying disinvestment options? Methods Context and setting SJHC is a 371-bed community teaching hospital in Toronto. In 2007/08, the hospital had over 21,000 admissions and almost 220,000 ambulatory care visits. The annual operating budget for 2007/08 was ~$230 million. Senior Management at SJHC is comprised of the President/CEO and two Executive Vice-Presidents while the Senior Executive Team is comprised of Senior Management and six other members. The hospital is overseen by a 19 member Board of Directors with ultimate budget and planning authority within the context of the Service Accountability Agreement between the hospital and the Toronto Central Local Health Integration Network (LHIN). The executive sponsor of the PBMA process at St. Joseph’s was one of the two Executive Vice-Presidents. The annual budgeting process at St. Joseph’s includes the development of an Operating Plan where decisions are made regarding departmental budget adjustments and service changes. The Operations Committee is responsible for the development of the 74 annual Operating Plan and is comprised of senior managerial and clinical representation from across all hospital departments. The Operations Planning Committee is a permanent sub-committee of the Operations Committee and is charged with the detailed work necessary for the development of the Operating Plan. The 2010/11 Operating Plan process was split into three streams: 1) Structural budget adjustments dealing with departmental imbalances between service volumes and funding allocations (i.e. structural deficits); 2) Potential efficiency gains identified through benchmarking with other similar-sized hospitals in Ontario; and 3) PBMA implementation to determine how resources might be reallocated to improve overall benefit at a time when new funding is very limited. The addition of the PBMA process to the Operating Plan was a strategic decision by Senior Management designed to fill a perceived gap in the tools available to the Operations Committee in meeting future budget challenges. Accordingly, the main goal in the first year of implementation was to establish the process within the organization’s budget making structure. Two steps were taken to facilitate this introduction. First, PBMA, for this first year, was isolated from the immediate financial pressures faced by the organization. Specifically, PBMA was not asked to contribute to the plan to address the projected 2010/11 deficit (i.e., no disinvestments were required to go to deficit elimination and all were to be available for re-allocation). And second, the implementation had no pre-set objective in terms of value of re-allocations; the goal was to see some service changes take place as a result of the process. PBMA implementation at St. Joseph’s PBMA implementation at St. Joseph’s followed standard procedures described in the literature (Peacock et al., 2006; Mitton & Donaldson, 2004b), except that an addition to Step one was made. This addition involved the determination and communication, 75 before the search for disinvestment options began, of the areas within the organization where service reductions would be acceptable to Senior Management31 (Table 5). TABLE 5   Steps of PBMA process at St. Joseph’s  31 Service reductions might not be acceptable in some areas of the organization for a number of reasons. For example, some services may be mandated, or subject to funding decisions outside of the control of the organization, some services may be too politically sensitive, some may be subject to collaboration with another organization which would make unilateral action illegal. The communication around acceptable service reductions should include a reason why reductions in some services are NOT acceptable. Steps Activities PREPARATION PHASE 1 Determine the aim and scope of the priority setting exercise Will the analysis examine changes in services within a given programme or between programmes? ADDITION TO STEP 1: Determination, and communication of the areas within the organization where service reductions are acceptable. 2 Compile a programme budget The resources and costs of programmes combined with activity information. 3 Form a marginal analysis advisory panel The panel should include key stakeholders (managers, clinicians, consumers, etc.) in the priority setting process. IMPLEMENTATION PHASE 4 Determine locally relevant decision making criteria The advisory panel determines local priorities (maximising benefits, improving access and equity, reducing waiting times, etc.) with reference to national, regional, and local objectives. 5 Identify where services could grow and where resources could be released through improved efficiency or scaling back or stopping some services The panel uses the programme budget along with information on decision making objectives, evidence on benefits from service, changes in local health care needs, and policy guidance to highlight options for investment and disinvestment. 6 Evaluate investments and disinvestments Evaluate the costs and benefits for each option and make recommendations for change. 7 Validate results and reallocate resources Re-examine and validate evidence and judgments used in the process and reallocate resources according to cost-benefit ratios and other decision making criteria. 76 To simplify the integration of PBMA in the organization, the process was organized to fit within existing structures. Accordingly, the Operations Committee served as the ‘Marginal Analysis Advisory Panel’ (step 3 in Table 5). Further, the Operations Planning Committee was renamed the ‘Resource Allocation Committee’ for the purposes of the PBMA exercise.  The Resource Allocation Committee was charged with carrying out tasks as directed by the Marginal Analysis Advisory Panel with the Advisory Panel retaining final ‘approval’ and responsibility for making recommendations to the Senior Executive Team (see Table 6). TABLE 6  Organizational structure for PBMA at St. Joseph’s WHO DUTIES Resource Allocation Committee (existing Operations Planning Committee plus some additional representation) Actual implementation of PBMA, e.g., establishing criteria, deciding on process guidelines, rating of proposals PBMA Advisory Panel (existing Operations Committee which has managerial and clinical representation from across the hospital) Oversight of PBMA implementation, approval of the process guidelines, criteria and ratings, and making resource allocation recommendations to the Senior Executive Team Senior Executive Team (nine members including CEO and two Executive Vice- Presidents) Review of the process recommendations and decisions on these recommendations Senior Management (CEO and two Executive Vice-Presidents) Final funding decisions (subject to Board approval) The requirement for making a determination, at the outset of the process, of the areas within the organization where service reductions would be acceptable was explicitly introduced at St. Joseph’s at the initial PBMA planning session with Senior Management (the purpose of this planning session was to set the broad parameters of PBMA implementation at St. Joseph’s). Senior Management subsequently decided that all services would be eligible for consideration as possible disinvestment options (i.e. managers can propose reductions in any of the services under their authority and Senior Management is willing to implement and defend service reductions, including possible termination of some services, in any of the service areas). This decision was clearly 77 communicated to managers at the Management Forum where the process was launched and also at both Medical Advisory Committee and Inter-professional Advisory Committee meetings. Case study research design The case study research design was selected for this project because the focus was as much on the understanding of the way the change to PBMA procedures affected the process as it was on the actual disinvestments results obtained at St. Joseph’s. The two research questions reflect the dual focus of this research. Question one addresses the ‘what was the result’ (in terms of disinvestment options identified and acted upon) of a PBMA process that included a formal determination and communication of acceptable service reductions at the outset. Question two deals with the ‘how and why’ of this result in relation to the change in procedures. The research methodology best suited to addressing the how and why questions is case study research, or specifically the explanatory case study research methodology: “case studies are the preferred strategy when ‘how’ and ‘why’ questions are being posed…and when the focus is on a contemporary phenomenon within some real-life context” (Yin, 2008; p.1). Data collection and analysis There were two sources of data in this study. The first data source was documents that accrued during the process, and observation of (parts of) that process. Throughout process implementation, all documentation and materials directly related to the disinvestment procedures were collected. These were the documents that were produced as a result of the process implementation; there was no search for documents as part of this research, only the collection of documents produced in the course of the process implementation. . The documents collected included those describing each disinvestment proposal as well as the briefing notes from the Resource Allocation Committee that accompanied the proposal ratings submitted to the Marginal Analysis Advisory Panel. Documents were used to determine factually the results of the PBMA process at St. 78 Joseph’s in terms of disinvestment options identified, their ratings, the process recommendations and the decisions regarding disinvestments. Opportunities for observation were provided by attending most of the PBMA meetings either by phone or in person. Observations were used to check the findings from the document review. The second source of data was a set of individual interviews. Once the resource allocation decisions related to the PBMA recommendations were made, semi-structured interviews were conducted with 14 decision makers who were involved in some manner with the PBMA process. The interview schedule was developed on the basis of past experiences with evaluation of PBMA implementations, adjusted for the specific interest in disinvestments (and the related challenges). The interview schedule was not pre-tested as it drew from relevant experience (see Appendix D for interview schedule.) Participants were purposively selected32 to provide a broad range of health care management experience (e.g. both clinical and managerial leadership), as well as of levels of authority within the organization and of degree of involvement in the process, from direct participation to more peripheral involvement. Evaluation interviews lasted about 30 minutes each and were audiotaped. The interview contents were coded using a thematic analysis approach (Braun &Clarke, 2008). The primary researcher performed the coding. Summary findings from the thematic analysis were reviewed by some interviewees, a form of feedback and validation often referred to as member checking (Schwartz-Shea, 2006, p.104). Findings were also verified through triangulation with the results from the document review. Ethics approval for this evaluation was obtained from the UBC Behavioural Research Ethics Board. Results Results are organized according to the two research questions. First, the outcomes in terms of disinvestments are described. Then the issue of the challenges faced by the  32 The selection was done in collaboration with Senior Management. 79 decision-makers tasked with the identification of disinvestment options is addressed, in the context of the impact of the procedural change that was introduced. Disinvestment options and resource allocation decisions In total, fifteen disinvestment proposals were submitted to the Resource Allocation Committee for rating. Out of the fifteen proposals submitted, eleven were service reductions, three were efficiency gains presented as service reductions and one was a reduction in a planned service increase (i.e. it did not involve any reduction to existing services). So, eleven genuine disinvestment options (i.e. proposals involving reductions in the quality and/or volume of services) were identified. Of these, eight were rated by the Resource Allocation Committee, which noted that three proposals were actually variations of the same proposal (only one of the three went forward for rating), and one proposal involved a change to Mental Health and Addictions services which was dropped from the process due to anticipated direct funding from another Government source. The ratings and ranking of the disinvestment options is presented in Table 7. Of the eight proposals that were ranked, the Resource Allocation Committee recommended that the top six be implemented. That is, these six proposals were identified as the disinvestment options where the estimated loss of benefit resulting from the service reduction was deemed to be less than the estimated gain in benefit associated with the service expansions made possible by the re-allocation of resources. As the focus here is on disinvestments, the proposed investment options and details of the re-allocation exercise are not provided. Senior Management accepted the recommendations as presented. As such, six services that were currently used by patients, and producing benefits, would be either 80 eliminated or have their volume and/or quality reduced once the decisions are fully implemented33. TABLE 7  Ratings for the disinvestment proposals at St. Joseph’s* Proposal number Proposal name Score 1 Eliminate a clinic 37.4** 2 Eliminate b clinic -4.9 3 Eliminate c Rehabilitation program -6.7 4 Eliminate d Health Centre -13.3 5 Reduce e Supplies to units -14.4 6 Eliminate f clinic -16.0 7 Eliminate g services -87.0 8 Disband h team -91.3 *This table was extracted from the Briefing Note sent by the Resource Allocation Committee to the Operations Committee after the proposals had been rated and ranked. Names of the services have been removed as these have not yet been publicly released. **A positive score for a resource release proposal indicates that cutting part, or all, of the program would actually increase benefits produced by the organization (i.e. improve service). In summary, the PBMA process as implemented at St. Joseph’s resulted in the identification of eleven disinvestment options involving service reductions. Of these, eight were formally rated and six were acted upon, in accordance with the process recommendations, by the organization’s leadership.  33 The total value of the resources made available for re-allocation to service expansions from these six disinvestments is approximately $1.05 million. 81 Challenges in the identification of disinvestment options Ten of the fourteen evaluation interview respondents were involved in the development of disinvestment proposals. These respondents were asked whether or not they found the identification of disinvestment proposals to be challenging. Six of them answered in the affirmative, and were asked to explain why. In four cases, the task was described as challenging because the respondents were in charge of support service programs and, as such, felt they did not have much, if any, control over service quality or volume. In their view, their role was to respond to the demands of those that control the clinical activities (#s in brackets after the quotations are the identifier for the respondent). “Q: Was [finding disinvestment options] something that was a challenge to do? A: It certainly was because, it’s because of the department I’m in. We do what other people ask us to do. We can’t just drop a service. If we do, we have to find someone else to do it for us. So, as long as the hospital is doing certain clinical activities, we have to support them” (#2) “Q: When you were looking for options for resource releases, did you find it challenging? A: Yes, and I think it was because I don’t have a clinical service, it’s more like a support service.” (#4) “Q: …did you find it to be a challenge? A: Yes, very much so because again it’s all about volume reduction and I don’t produce a unit of volume.” (#6) “We are a support service, so as a support service, we can only change depending on what is going to happen in the rest of the hospital.” (#7) 82 The fifth respondent found it difficult to identify service reductions with a value above the minimum required for disinvestment proposals34: “Q:…did you find that challenging? A: Very, we have a fairly small budget in x, just over $3 million I think…the challenge was that each release had to be $50,000, and that’s a rather large release looking at a budget of about $3 million.” (#8) 35 In the case of the sixth respondent, the explanation given was of the type expected when there is no clarity around acceptable service reductions: “I think the goal was to say what is it within the services that encompass the program that we feel are not core for a community hospital, and within the medicine portion of my portfolio there was not an easy answer at all. Looking for stuff that we could say ‘we no longer really need this within medicine’ in the x care program was difficult.” (#14). Overall, only one of ten respondents involved in the identification of disinvestment options said he or she faced challenges that are related to uncertainty around what can be a disinvestment option (the stated issue with support service managers was one of control, not one of whether or not any given clinical service could be reduced). This finding suggests that amongst those who were interviewed there was clarity around acceptable service reductions. Some clarity (or instructions) around acceptable disinvestment options was the main purpose of the key modification to PBMA introduced at St. Joseph’s for the purpose of this project. The expected impact of this procedural change on the identification of disinvestment options, and ultimately on the actual disinvestments being implemented, depends on its success in establishing this clarity.  34 In order to limit the time demands within the PBMA process it is common to establish minimum values for disinvestment proposals as part of the process guidelines. At St. Joseph’s the minimum value was $50,000. 35 The department in question covers a large number of small programs therefore a disinvestment proposal of a minimum of $50,000 represents a significant cut to any of those programs. 83 In order to confirm that there was clarity around acceptable disinvestment options, the respondents were also asked specifically about their understanding on the issue of the type of service change that could be proposed for disinvestment. Most respondents seemed to have understood clearly that the PBMA process was about service changes (as opposed to efficiency gains) and that all services were eligible in the search for disinvestment options. Examples of responses/comments include: “Q: Was it clear to you that there were no sacred cows? A: Yes it was very clear.”(#1) “Q: So, in your specific case, when you were putting together your proposals (for disinvestments) your understanding was that this is an opportunity to look at making some changes. A: Yes, yes, in a thoughtful way.” (#4) “We can’t be everything to everybody anymore, I believe that.” (#7) “Q: Was your understanding that what you were looking for at that time was changes in services as opposed to trying to get more efficient? A: Yes, that was made clear, I think, throughout the process.” (#8) “I looked at my services and I thought about which one can I afford to lose without jeopardizing the delivery of services that we provide, it would be at a different level, but without jeopardizing the crux of why we are here.” (#10) Only one of the respondents still seemed to consider some areas of service out of bounds in the search for disinvestment proposals: “In my mind, that is all I can come up with because of my areas. I have x department and y department. I can’t do anything with y department, I can’t close 84 stretchers. I can’t do resource releases in y department, it doesn’t make any sense. So I had to do it in x department.” (#3) There were no instructions in any of the information related to the PBMA implementation that stated that some specific service reductions were not ‘possible’. In fact, the messaging was consistent that all services were eligible for reductions. It seems therefore that this respondent interpreted the communication regarding acceptable service reductions through a personal ‘filter’ of overarching views on what services can be reduced. The evaluation also asked respondents to compare PBMA to the way decisions about disinvestments were made in previous years. The previous processes around disinvestment decisions were described by respondents as limited in scope (only focused on specific parts of the organization), informal, controlled by Senior Management, and not inclusive (i.e. only those directly affected by the initiatives introduced by Senior Management were involved). “The other service reductions that happened before…it was not as formalized [as the PBMA process], again it came more from a senior level down. It did not fit with the strategic direction of the organization…. It’s not that it was not transparent…people were aware what was going on with those, to the extent that they needed to be aware.” (#6) Q: Looking for disinvestment options, was it done before? “ A: If it was done, it was done informally I think, in the programs themselves, it was not done collectively, as a group” (#10) “The discussion (around disinvestments) would inevitably go to ‘well senior management needs to make a decision about what is strategic and not strategic’, right. Now we acquired a tool to help us understand the relative value of different things.” (#12) 85 The PBMA process on the other hand was described as a collective exercise (with wide representation) and certainly a different, and preferable, way of approaching disinvestments. “We have the potential to move resources. I don’t think that was thought of before that way.” (#10) “Did [the PBMA process] make people look at [service changes] in a different way? Absolutely. I don’t know if people really thought about the fact that we could not be in the business of doing program x anymore before, necessarily, or if they did, they did not know how to present it in a way that made sense, so that people could buy in.” (#11) The assertion that the previous processes around disinvestments were informal and not inclusive seems to be corroborated by the fact that three respondents said they were not aware of a history of service reductions at St. Joseph’s. “Q: That is something (looking for changes in services) that you were not doing as part of the budgetary process in previous years? A: That’s correct because generally we have only looked at efficiencies.” (#8) The difference between PBMA and previous processes was particularly significant to one respondent who described it as: “a culture shift in terms of the way we use resources and I think what this does is it helps people understand the opportunity cost better, which is what this is always all about, right.” (#10) The proposed procedural change seems to have provided clarity around the issue of acceptable disinvestment options. In the absence of instructions as to what is 86 acceptable in terms of service reductions in the organization, managers are left to their own interpretation of what service reductions leadership is ready to implement and defend. In this sense, some clarity would be expected to facilitate the search for disinvestment options simply by eliminating the need to guess what can be a disinvestment option. At St. Joseph’s, only one respondent found challenges in the search for disinvestment options that were more than technical issues that can be addressed with simple adjustments to the process. Moreover, the PBMA process as it pertains to disinvestments was generally viewed positively and was considered, by most respondents, to be an improvement over previous processes as opposed to being described as a significant practical challenge or an ‘impossible’ task. Discussion Factors such as strong leadership, organizational stability (low staff turnover, especially in strategic planning divisions), a high level champion for the process, an organizational culture to learn, having resource allocations decisions that must be made (i.e. where the status quo is not an option) and a strong relationship between physicians and managers have all been identified as key success factors for priority setting implementation in health care (Donaldson, 2001; Mitton & Donaldson, 2003; Peacock et al., 2009). Yet practical challenges to such implementation remain. It is in this context that the current research addresses a specific practical challenge to the implementation of PBMA, namely identifying disinvestment options. Each of the success factors identified in previous research affects the identification of disinvestment options in some way. The proposed procedural change to PBMA can be seen as a medium through which these success factors can exert a more direct influence on this particularly challenging part of the process. In this sense, the proposed change builds on and complements existing knowledge and at the same time adds an explicit step that has not been considered previously. Evaluating the procedural change at St. Joseph’s produced results in terms of identification of disinvestment options, implementation of recommendations, and rigor 87 and transparency in decision-making that were as anticipated. These results are consistent with the hypothesis generated in chapter 3 that clarity around what areas of services are acceptable for reductions can facilitate the identification of disinvestment options. Clarity can have this effect because it seems to attenuate the conflict that the request for disinvestment options, as part of the PBMA process, creates for decision-makers. This request for disinvestment options goes against some common organizational incentives, such as the positive implications of an ever-increasing departmental budget or the expectation to be an advocate for one’s population. These organizational incentives and the requirements for disinvestment options are, in this sense, two opposing forces, which means that participation in the PBMA process tends to put managers in an uncomfortable position. Some direction as to what areas of services are acceptable for retrenchment can reduce the pressures of these opposing forces by making service reductions a realistic possibility, at least in some circumstances. The results at St. Joseph’s also provide some lessons that can be applied to future work on improving the effectiveness of the PBMA process. The first lesson relates to support service departments. It became clear in the evaluation phase of the project that simply identifying the areas where service reductions are acceptable is not sufficient to permit full participation in the development of disinvestment proposals by everyone involved in the provision of these ‘eligible’ services. Having broader participation in the search for disinvestment options can be desirable because more perspectives bring more information and expertise to the task and can therefore lead to better informed proposals. In the case of St. Joseph’s, most of the support service managers limited their search to those very few support services not directly linked to clinical services. This behavior was not explicitly dictated by any part of the process, but rather seemed to be the result of a feeling amongst support service managers that they did not have the jurisdictional authority to offer up those of their services that would impact clinical services’ capacity. It would be advisable in future case studies to raise this issue at the outset with the leadership of the organization. This would lead to an explicit choice as to the role of support service managers in the identification of disinvestment proposals. 88 A second key lesson from the St. Joseph’s case study is the need to be explicit in the communication of acceptable service reductions, especially when these include areas of the health care system where services are not commonly reduced (i.e. the degree of clarity seems to matter, at least in some cases). One of the evaluation interview respondents seems to have understood the message regarding acceptable service reductions in the context of common practice around such reductions, which typically excludes certain areas of service. If some, or all, of these ‘typically-excluded’ areas (e.g. emergency services) are to be included in acceptable service reductions, they should be specifically identified as such. The findings of this study are limited to the site at which the data were generated and no causal inferences can be asserted on the basis of this particular case (i.e. the findings from this research cannot be generalized). However, the study did yield in-depth descriptions about the particular experiences at the site, which provided insights into the use of PMBA in a real life setting in ways that conventional experimental research, with its requirement for controlled conditions, may not be able to provide (Pine, 2009). While the results at St. Joseph’s support the hypothesis developed in chapter 3, it must be acknowledged that factors specific to the implementation at St. Joseph’s could have led to these results independently of the procedural change that was introduced. For example, the fact that PBMA had a very limited role in terms of resource allocation decisions may have facilitated in itself the identification of disinvestment options because decision- makers knew the ‘risks’ were very limited (i.e. no large changes to services were anticipated to result from this PBMA implementation). It might have also been that in the first year of implementation, the process provided a forum for some service reductions that had been discussed informally for some time to be aired publicly (thereby providing ready-made disinvestment options). As plausible as these other explanations are, it seems however, that the procedural change still, at a minimum, set the stage for them and therefore played a necessary role. The overall thesis project follows the five phases of action research as described by Susman (1983): diagnosis, action planning, taking action, evaluating and specifying 89 learning. The diagnosis followed from the observations and findings in the exploratory case studies at the Vancouver Island Health Authority and at Menno Place which built on a broad review of the literature on priority setting. On the basis of this diagnosis, specific action was planned, this action was implemented at St. Joseph’s and, finally, the results were evaluated. In addition to the research focus, the implementation of PBMA at St. Joseph’s has also established the process as one of the streams of the annual Operating Plan. Key staff has now been trained in the PBMA process, the implementation structure has been developed and there has been some real life experience with proposal development, proposal rating and marginal analysis. Conclusion The PBMA process as it was implemented at St. Joseph’s Health Centre produced the expected results. A number of service reduction proposals were submitted and the PBMA recommendations regarding disinvestments were fully implemented. Just as important as these outcomes were the findings that most decision-makers who were interviewed for the evaluation were clear on what was acceptable as a disinvestment option and that only one of them faced challenges, in identifying those options, that were more than simple technical process issues. These findings support the key result from the exploratory research case studies: when there is clarity around acceptable service reductions, organizational pressures that would otherwise induce managers to limit their search for disinvestment options to redundant or unused services seem to be largely defused, which then seems to leverage the known PBMA success factors and therefore facilitate the search for genuine service reduction options, everything else being the same.. If this result is further validated through more case studies, a possible solution to a long-standing and serious challenge to the use of PBMA in health care will have been found; this could lead to a broader use of priority setting processes - and less reliance on historical spending patterns - to guide resource allocation decisions. 90 CHAPTER 5 CONCLUSION This project was about facilitating the identification of disinvestment options in the context of priority-setting activities in health care. The principal finding is that determination and communication of acceptable service reductions at the outset of the priority-setting process may improve the probability that the process will be successful by increasing the odds that realistic disinvestment options get on the table as part of that process.  How the research arrived at this finding and what these results contribute to the literature on priority-setting in health care are addressed in this concluding chapter. Why an interest in priority setting and, specifically, in disinvestments? The starting point for this research project was that, in health care, claims for resources typically exceed available resources (Evans, 2007). This is a reality that decision-makers must contend with on an ongoing basis, as evidenced by this recent statement from British Columbia’s Minister of Health Services: “The demand [for health care services] is seemingly endless” (Fowlie, 2010). When claims on resources exceed available resources, choices have to be made, explicitly or not. ‘Why an interest in priority-setting’ is really asking whether or not it matters if these choices are made explicitly . Historically, the choices have generally not been made explicitly (Ham, 1997; Bate et al., 2007). If one accepts the reasonable assumption that the lack of explicit decision-making on resource allocation is implicated in at least some of the considerable system inefficiencies that virtually every informed observer agrees plague the system, one must ask why formal, structured priority-setting processes are not used more. The answer to this question requires a historical review of priority-setting in health care and, ultimately, provides the reason for the specific interest in the issue of disinvestments. Priority setting in health care has progressed through a number of phases including, most notably, needs 91 assessment and economic evaluation to today’s sophisticated formal processes. Each step forward was designed to address shortcomings of the previous methodology in terms of responding to the needs of decision-makers. At this stage of the progression, the overarching issue with priority setting processes is: how can priority setting be made to fit better in the decision-making processes of health care organizations? The selected strategy to address this question was to continue with research on priority setting success factors, as these factors deal with the practical issues around the implementation of priority-setting processes. However, instead of examining success factors for priority- setting as a whole, it was decided to narrow the focus to one specific practical challenge to the implementation of priority-setting processes, and PBMA in particular. In this context, the issue of identification of and agreement on disinvestment options emerged as an obvious candidate. It is long-standing (Cohen, 1994) and significant (Mitton & Donaldson, 2004b) challenge to priority-setting, and moreover is a necessary ingredient to the success of priority-setting implementation. Key finding - What is the main contribution to knowledge? The first question asked with regards to the identification of disinvestment options was whether the practical challenges that this posed were related to a lack of knowledge on how to accomplish this task. Specifically, do decision-makers know36 how to release resources by means other than increasing efficiency (i.e., producing the same services with fewer resources)? A review of the literature on cost reduction or containment in health care found ten proven strategies to reduce or contain costs (i.e., reduce or contain the use of resources) other than increasing efficiency (chapter 2). Not all strategies fit all situations, but the results of this review demonstrated that the challenges posed by the identification of disinvestment options are not due to lack of knowledge on means to accomplish this. Such information is readily available in the health services literature, with multiple examples of real life applications.  36 Or if they do not know, is this knowledge readily available? 92 A logical next step was then to conduct exploratory research focusing on the specific challenges posed by processes intended to identify disinvestment options. This exploratory research took the form of case studies of PBMA implementation in different settings. This thesis involved three observational case studies at the Vancouver Island Health Authority and one at Menno Place. PBMA in each of these cases was implemented following standard procedures, adjusted for the specific circumstances of each case. This exploratory research found key differences across the case studies in the observed results of the process in terms of disinvestment options generated as well as the impact on resource allocation decisions. Importantly, the observed differences could not be fully explained by variations in key success factors reported in the literature, such as leadership and organizational buy-in. However, one prominent feature present in the more successful case but absent in the less successful ones was the determination and communication, at the outset of the process, of the areas of the organization where service reductions would be acceptable. This observation indicated a possible connection between some form of clarity regarding areas of acceptable service reductions, the ability to identify meaningful disinvestment options (i.e., options other than more attempts to increase efficiency), and the extent to which the priority-setting process ultimately affects resource allocation decisions. The potential role of clarity regarding acceptable service reductions, as revealed in the exploratory research, was then tested at St. Joseph’s Health Centre (a large community teaching hospital in Toronto, Ontario). PBMA was implemented at St. Joseph’s following the same procedures as in the exploratory research cases, with the exception of the addition of the requirement for a determination and communication of acceptable service reductions at the outset of process implementation. The observed results in terms of disinvestment options submitted and impact of PBMA recommendations on resource allocation decisions were as anticipated (i.e. many service reduction proposals were made and the resource allocation decisions followed the process recommendations). Further, an evaluation of the process at St. Joseph’s found that the attitudes of decision-makers towards the identification of disinvestment options was generally as presumed. Only one of ten respondents talked of challenges in identifying 93 disinvestment options that were related to uncertainty regarding acceptable service reductions. In the exploratory research cases, such uncertainty was widespread, and was found to either lead managers to refuse to participate in the search for disinvestment options or to limit their search to disinvestments that do not directly cause service reductions, such as the elimination of unfilled positions. While, we can not establish a causal link between the procedural change implemented and the observed results at St. Joseph’s because these results could be due to other factors, such as the previous experience with disinvestment initiatives, these results support the key finding from this thesis research project which can be stated as follows: The case study work here has produced results that are consistent with (though certainly do not “prove”) the hypothesis that a determination and communication, at the outset of process implementation, of which services can be retrenched and which ones cannot, creates an environment for the priority-setting process that lessens the practical challenges posed by the requirement to submit disinvestment options, and increases the probability of a successful process. The determination and communication of acceptable service reductions is believed to have this effect because 1) it explicitly confirms that PBMA (or similar priority-setting process) is about making changes to services and that some services will be reduced (this establishes a linkage between the priority-setting process and subsequent decision-making), as opposed to being a process centered on ‘innovative’ ways to reduce costs with no ‘pain’ to anybody; and 2) it removes the uncertainty around possible disinvestment options; such uncertainty has been found, in the exploratory research, to lead to decision-makers adopting a ‘default’ definition of acceptable service reductions which restricts their search to services that are redundant or unused, or eliminates the search completely. The impact of a determination and communication of acceptable options for service disinvestments at the outset of process implementation is best illustrated by comparing the PBMA implementations at VIHA with those at Menno Place and St. 94 Joseph’s. At VIHA, no determination and/or communication of acceptable service reductions occurred at the outset of process implementation. Very few disinvestment options (other than increasing efficiency by, for example, reducing overtime hours or eliminating redundant services) were identified and essentially no reductions in services were implemented. A common theme in responses to questions about the lack of disinvestment options in that setting was that disinvestment options were ‘non-existent’ or an ‘impossibility.’ On the other hand, acceptable service reductions were discussed, determined, and communicated at Menno Place and at St. Joseph’s. The discussion and determination of acceptable service reductions at Menno Place emerged on its own in the planning stages of the process, while at St. Joseph’s, it was established as a formal step of the process. In both organizations, PBMA led to meaningful service reduction proposals and to the implementation of reductions in services37. The determination and communication of acceptable service reductions appears, limitations discussed below notwithstanding, to have created an environment where a priority-setting process ‘with teeth’ has a role to play in decision-making and where decision-makers are encouraged to participate. The greatest challenges to the identification of disinvestment options in these two ‘successful’ organizations came in the form of discussions around technical process issues as opposed to “the impossibility” of finding service cuts in an “under funded” environment. For example, vigorous debate at both Menno Place and St. Joseph’s centered on the nature of the criteria and the specific rating scales for each criterion. At VIHA, the debate tended to focus on what department(s) should be exempted from the process (e.g. support services would argue that because they have historically suffered the greatest budget cuts, they should be exempted from the requirement to submit disinvestment options). How does this key finding fit with what we knew? Previous research has demonstrated that decision makers in health care want help with resource allocation decisions and that, in theory, the characteristics and procedures  37 The few service reductions implemented at VIHA were for services that had already been replaced, in practice, by new services. 95 of the PBMA process meet the demands or wishes of decision makers (Dionne et al., 2008; Bate et al, 2007). None of this was challenged in any of the priority-setting applications included in this project. In the case of VIHA, for example, senior management explicitly confirmed the need for a process such as PBMA and continued supporting the PBMA process even in the face of limited success or impact (i.e. as one senior executive at VIHA stated, “the fundamentals of PBMA are incredibly sound”). Despite this need and support for formal structured priority-setting processes, use of such processes is still limited due to the practical challenges posed by their implementation (Bate et al., 2007; Peacock et al., 2009). Accordingly, much of the recent research on priority-setting processes has focused on the practical challenges to priority- setting implementation and this has resulted in the identification of crucial success factors. The key finding of this research project builds on these recent findings. This project’s key finding can be seen as a step that allows the known success factors to come to bear directly on a specific and important practical challenge to PBMA implementation. What are the limitations of this research project? The key finding from this project emerged from four exploratory case studies and one case study where the impact of a procedural change related to this finding was studied. Each of these cases was selected for convenience, based on two eligibility criteria: having priority setting issues and an interest in structured priority setting process to address these issues. Research in all cases followed an action research design, which has proven a particularly powerful way to influence practice (Meyer, 2000; Susman & Evered, 1978). However, this method has some intrinsic limitations, primarily that no causal relationship can be established on the basis of the findings and that these findings cannot be generalized to other settings. These limitations are related to sample size and composition, the absence of comparators, and the lack of control over many possible confounders. 96 Despite these limitations, the key finding is meaningful because of its grounding in real life settings. In complex policy issues, such as the practical challenges to priority- setting implementation, an experimental research design may not provide the depth of insight necessary to understand the dynamics that led to the observed results (Pine, 2009). With so many possible influences on the observed results, as is the case with priority- setting implementation, a detailed ‘local’ understanding of how the results were achieved provides valuable information for assessing whether or not the studied change is likely to produce similar results in a different setting. Where do we go from here? The logical next steps in the new avenue of research opened by this project— focusing on specific practical challenges to PBMA implementation—are first to address the limitations of this project and second to expand the research to other known specific practical challenges. Addressing the limitations identified above can be done through replication of the work done at St. Joseph’s. If similar results are observed in multiple different settings with variation in their other critical success factors, in terms of both actual disinvestment options submitted and disinvestment initiatives implemented as well as with respect to the type of challenges faced by decision-makers during the process, then the initial conclusions will be more robust. Replication will also provide opportunities to refine the logistical implications of the change in procedure. For example, the St. Joseph’s case study raised two technical issues related to the inclusion of a requirement for a determination and communication of acceptable service reductions in the PBMA process. First, the role to be played by support service managers should be explicitly defined at the same time as the acceptable service reductions. Without such clarification, support service managers at St. Joseph’s limited their search for disinvestment options to support services that did not directly affect clinical activities. This had two effects: it made their search very difficult and limited input to the process of identifying disinvestment options by essentially excluding these managers. The second issue was the need to be very 97 explicit in the communication of acceptable service reductions. Some services have traditionally been exempted from service reductions and, for at least one manager at St. Joseph’s, they were still excluded despite the message that all services were eligible for service reduction proposals. Other specific practical challenges to the implementation of PBMA that may benefit from an approach similar to that taken in this research project are the handling of support services (as opposed to the role of support service managers in identifying disinvestment options which is a technical issue) and the rating of proposals of very different values. With regard to support services, the main challenges are 1) whether support services should be treated the same way as clinical services or if the process should focus on clinical services, with each proposal for change to clinical services having the support service impact included; and 2) if support services and clinical services are treated equally, can one set of criteria assess fairly possible changes to both types of services? As for the rating of proposals of very different values, the challenge lies in differentiating size of impact from nature of impact. A $500,000 disinvestment will most likely have a much larger impact on the organization’s ability to meet its objectives than a $50,000 disinvestment. But how much of this is due to the sheer size of the disinvestment? An effective process must be able to accurately assess a loss of benefits (or ability to meet the organization’s objectives) per dollar of resources saved. At this point, we can conclude that the results of the PBMA process, as implemented at St. Joseph’s, in terms of disinvestment options identified and acted upon, were as anticipated based on the observations and findings from the exploratory research. We can also say that the PBMA process arrived at these results in the manner that was expected, i.e. the search for disinvestment options for most managers interviewed in the evaluation was not centered on services that are ‘non-core’ or unused. We also have one case of successful knowledge transfer as the Executive Vice-President and Chief Financial Officer fully endorsed continued use of the PBMA process at St. Joseph’s. 98 The initial literature review, the exploratory research, and the case study at St. Joseph’s have surfaced a potential solution, or a piece of a solution, to the practical challenges posed by the identification of disinvestment options: when clarity exists with regard to acceptable service reductions, organizational pressures that would otherwise induce managers to limit their search for disinvestment options to redundant or unused services are largely defused.  This then seems to leverage the impact of the known PBMA success factors and facilitate the search for genuine service reduction options. In addition, this clarity seems to place the priority-setting activities within the realm of the decision- making process by making service reduction proposals an acceptable management option within the defined parameters which serves to further leverage the existing success factors by enhancing the relevance of the resulting recommendations. PBMA has long held considerable theoretical appeal for decision-makers, but successful implementation has been challenging because it creates ‘winners’ and ‘losers.’ The current work has attempted to move the practice of priority-setting forward so that ‘theoretical appeal’ will take on more ‘practical appeal’ and lead to more ‘successful implementation.’ 99 BIBLIOGRAPHY Abourjaily, P., Gouveia, W.A., Selker, H.P., & Zucker, D.R. (2005). Evaluating the nondrug costs of formulary coverage restrictions. Manag Care., 14(8), 50-7, 62. Alberta Bone & Joint Health Institute (2008). Alberta Hip & Knee Replacement Project. Retrieved from http: //albertaboneandjoint.com/projects/arthroplasty/faqs.asp Allen, J.W., Hahm, T.X., & Polk, H.C. Jr. (2003). Surgeon-led initiatives cut costs and enhance the quality of endoscopic and laparoscopic procedures. JSLS, 7(3), 243-7. Antonucci, Y.L., & Bender, A.D. (1998). The quality and cost of medical care. The potential for information technology to meet the challenge. Med Group Manage J., 45(3), 12-4, 16-9. Babazono, A., Miyazaki, M., Une, H., Yamamoto, E., Tsuda, T., & Mino, Y. (2004). A study on a reduction in visits to physicians after introduction of 30% co-payments in the employee health insurance in Japan. Ind Health, 42(1), 50-6. Bate, A., Donaldson, C., & Murtagh, M. J. (2007). Managing to manage health care resources in the English NHS? What can health economics teach? What can health economics learn? Health Policy, 84, 249-261. Bates, D.W., Pappius, E.M., Kuperman, G.J., Sittig, D., Burstin, H., Fairchild, D… Teich, J.M. (1998). Measuring and improving quality using information systems. Medinfo, 9 Pt 2, 814-8. 100 Bellanger, M.M., &  Mossé, P.R. (2005). The search for the Holy Grail: combining decentralised planning and contracting mechanisms in the French health care system. Health Econ., 14(Suppl 1), S119-32. Berg, G.D., Thomas, E., Silverstein, S. Neel, C.L., & Mireles, M. (2004). Reducing medical service utilization by encouraging vaccines: randomized controlled trial. Am J Prev Med., 27(4), 284-8. Bernfort, L. (2003). Decisions on inclusion in the Swedish basic health care package-- roles of cost-effectiveness and need. Health Care Analysis, 11(4), 301-308. Birkmeyer, J.D., Skinner, J.S., & Wennberg, D.E. (2002). Will volume-based referral strategies reduce costs or just save lives? Health Aff (Millwood), 21(5), 234-41. Bloom, J.R., Hu, T.W., Wallace, N., Cuffel, B., Hausman, J.W., Sheu, M.L., & Scheffler, R. (2002). Mental health costs and access under alternative capitation systems in Colorado. Health Serv Res., 37(2), 315-40. Bodrock, J.A., & Mion, L.C. (2008). Pay for performance in hospitals: implications for nurses and nursing care. Qual Manag Health Care, 17(2), 102-11. Bottomley J.M., & Raymond, F.D. (2007).  Pharmaco-economic issues for diabetes therapy. Best Pract Res Clin Endocrinol Metab., 21(4), 657-85 Braun,V. & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3, 77-101 Burns, S.M., Earven, S., Fisher, C. Lewis, R., Merrell, P., Schubart, J.R., & Bleck, T.P. (2003). Implementation of an institutional program to improve clinical and financial outcomes of mechanically ventilated patients: one-year outcomes and lessons learned. Crit Care Med., 31(12), 2752-63. 101 Canadian Institute for Health Information (2007) Canada’s Health Care Providers, Ottawa, On: CIHI. Chernew, M.E., Shah, M.R., Wegh, A., Rosenberg, S.N., Juster, I.A., Rosen, A.B.,…Fendrick, A.M. (2008). Impact of decreasing copayments on medication adherence within a disease management environment. Health Aff (Millwood), 27(1),103-12. Cohen, D. (1994). Marginal analysis in practice: an alternative to needs assessment for contracting health care. British Medical Journal, 309, 781-784. Cohen, D. R. (1995). Messages from Mid Glamorgan: a multi-programme experiment with marginal analysis. Health Policy, 33(2), 147-155. Daniels, N., & Sabin, J. (2002). Setting Limits Fairly: Can We Learn to Share Medical Resources? Oxford: Oxford University Press. Davis, R.N. (2004). No more chances for supply-chain savings? Look again! Healthc Financ Manage., 58(1), 68-75. Devlin, N., & Parkin, D. (2004). Does NICE have a cost-effectiveness threshold and what other factors influence its decisions? A binary choice analysis. Health Economics, 13, 437-452. . Dionne, F., Mitton, C., Smith, N. & Donaldson, C. (2008). Decision maker views on priority setting in the Vancouver Island Health Authority. Cost Effectiveness and Resource Allocation, 6, 13. Domaszewicz, A. (2006). Consumerism's sea change: how it will affect your company in the coming years. Benefits Q., 22(2), 41-5. 102 Donaldson, C., & Mooney, G. (1991). Needs assessment, priority setting, and contracts for health care: an economic view. British Medical Journal, 303, 1529-1530. Donaldson, C. (2001) Priority Setting within Regional Funding Envelopes: The Use of Program Budgeting and Marginal Analysis. CHSRF Research Report. Retrieved from http://chsrf.ca/final_research/ogc/donaldson_e.php Elshaug, A. G., Hiller, J. E., & Moss, J. R. (2008). Exploring policy-makers’ perspectives on disinvestment from ineffective health care practices. International Journal of Technology Assessment in Health Care, 24(1), 1–9. Evans, R. (2007). “Extravagant Americans, Healthier Canadians: The Bottom Line in North American Health Care.” In: D. M. Thomas (ed.), Canada and the United States: Differences that Count. Peterborough, ON: Broadview Press. EVIDEM (2009) Evidence and Value: Impact on decision Making. Retrieved from http: //evidem.org on April 23, 2009. Fendrick, A.M., & Chernew, M.E. (2006). Value-based insurance design: aligning incentives to bridge the divide between quality improvement and cost containment. Am J Manag Care,12 Spec no., SP5-SP10. Fowlie, J. (2010, March 6) Health innovation starts at home, Falcon says. The Vancouver Sun.  Retrieved from http://www.vancouversun.com/health/Health+innovation+starts+home+Falcon+sa ys/2649430/story.html Frakes, J.T. (2002). Outpatient endoscopy. The case for the ambulatory surgery center.            Gastrointest Endosc Clin N Am., 12(2), 215-27. 103 France, N.C., Francis, G.A., & Lawrence, S.R.(2003). The costs of New Zealand pathology. Health Policy, 64(2), 131-41. France, J., Keen, C., & Bowyer, S. (2008). Communicating risk to emergency department patients with chest pain. Emerg Med J., 25(5), 276-8. Frezza, E.E,, Girnys, R,P., Silich, R.J., Coppa, G.F. (2000). Commentary: quality of care and cost containment are the hospital-based ambulatory surgery challenges for the future. Am J Med Qual., 15(3), 114-8. Gallagher, T.H., St Peter, R.F., Chesney, M., & Lo, B. (2001).  Patients' attitudes toward cost control bonuses for managed care physicians. Health Aff (Millwood), 20(2), 186-92. Gebara, R., Najarian, K., Wagner, J.M., & Narla V. (2004). Five steps to cost savings: adjusting ventilator humidification protocols. Biomed Instrum Technol., 38(2), 147-54. Gerard, K. (1992). Cost-utility in practice: A policy maker’s guide to the state of the art. Health Policy, 21, 249-279. Gibson, J. Mitton,C. Martin, D. Donaldson,C. & Singer, P. (2006). Ethics and economics: does programme budgeting and marginal analysis contribute to fair priority setting? Journal of Health Services Research and Policy, 11, 32-37. Giesbrecht, N., & Haydon, E. (2006). Community-based interventions and alcohol, tobacco and other drugs: foci, outcomes and implications. Drug Alcohol Rev., 25(6), 633-46. Ginsburg, P.B. 2004. “Controlling Health Care Costs” N Engl J Med., 351,1591-3. 104 Goddard, M., Mannion, R., & Ferguson, B. (1997). Contracting in the UK NHS: Purpose, Process and Policy. Discussion Paper 156. Centre for Health Economics, York Health Economics Consortium, NHS Centre for Reviews and Dissemination. Goroll, A.H., Berenson, R.A., Schoenbaum, S.C., & Gardner, L.B. (2007). Fundamental reform of payment for adult primary care: comprehensive payment for comprehensive care. J Gen Intern Med., 22(3), 410-5. Gugushvili, A. (2007). The advantages and disadvantages of needs-based resource allocation in integrated health systems and market systems of health care provider reimbursement. MPRA Paper No. 3354. Retrieved from http://mpra.uni- muenchen.de/3354/ Ham, C. (1997). Priority setting in health care: learning from international experience. Health Policy, 42(1), 49-66. Ham, C. (2008). World class commissioning: a health policy chimera? J Health Serv ResPolicy, 13(2), 116-21. Hanley, N., & Spash, C. L. (1993). Cost-Benefit Analysis and the Environment. Cheltendam, UK: Edward Elgar Publishing Ltd. Hauck, K., Smith, P. C., & Goddard, M. (2004). The economics of priority setting fore health care: a literature review”. The International Bank for Reconstruction and Development / The World Bank, Washington, DC. Health Innovation Forum. Operating Room Efficiency. Case study. Richmond Hospital staggers surgeries in dedicated hip and knee unit. Retrieved from http://www.healthinnovationforum.org/2008/apr/01/operating-room-efficiency/ 105 Healy, W.L., & Iorio, R. (2007). Implant selection and cost for total joint arthroplasty: conflict between surgeons and hospitals. Clin Orthop Relat Res. 457, 57-63. Hopkins, J.R. (1999). Financial incentives for ambulatory care performance improvement. Jt Comm J Qual Improv., 25(5), 223-38. Hudson, M.A. (2005). Texas passes first law for safe patient handling in America: landmark legislation protects health-care workers and patients from injury related to manual patient lifting. J Long Term Eff Med Implants, 15(5), 559-66. Jackson, S., Sistrunk, H., & Staman, K.(2003). Improved patient care and reduced costs: results of Baptist Health Systems' acute coronary syndromes project. J Cardiovasc Manag., 14(4), 17-20. Jhaveri, M., Seal, B., Pollack, M., & Wertz, D. (2007). Will insomnia treatments produce overall cost savings to commercial managed-care plans? A predictive analysis in the United States. Curr Med Res Opin., 23(6),1431-43. Keating, N.L., Landrum, M.B., Landon, B.E., Ayanian, J.Z., Borbas, C., & Guadagnoli, E. J. (2007). The influence of cost containment strategies and physicians' financial arrangements on patients' trust and satisfaction. Ambul Care Manage., 30(2), 92- 104. Ketcham, J.D., & Furukawa, M.F (2008). Hospital-physician gainsharing in cardiology. Health Aff (Millwood), 27(3), 803-12. Langton, J., & Crampton, P. (2008). Capitation funding of primary health organisations in New Zealand: are enrolled populations being funded according to need? N Z Med J., 121(1272), 47-58. 106 Larmer, P.J., Tillson, T.M., Scown, F.M., Grant, P.M., & Exton, J. (2008). Evidence- based recommendations for hand hygiene for health care workers in New Zealand. N Z Med J., 121(1272), 69-81. Levy, R.N. (2000). Does cost containment create conflict in the care of the elderly patient? Clin Orthop Relat Res., 378, 66-70. Lewandrowski, K. (2003). Managing utilization of new diagnostic tests. Clin Leadersh Manag Rev., 17(6), 318-24. Lingard, L., Regehr, G., Orser, B., Reznick, R., Baker, G.R., Doran, D.,…Whyte, S. (2008). Evaluation of a preoperative checklist and team briefing among surgeons, nurses, and anesthesiologists to reduce failures in communication. Arch Surg., 143(1), 12-7, discussion 18. Long, S.K., Coughlin, T.A., & King, J. (2005). Capitated Medicaid managed care in a rural area: the impact of Minnesota's PMAP program. J Rural Health., 21(1), 12- 20. Mackintosh, M. (1993). Economic Behaviour and the Contracting Outcome under the NHS Reforms. Theory and Example of Community Nursing. Accounting, Auditing & Accountability Journal, 6(3), 133-155. Marchibroda, J.M. (2008). The impact of health information technology on collaborative chronic care management. J Manag Care Pharm., 14(2 Suppl), S3-11. Marmor T, J. Oberlander and J. White. 2009. “The Obama administration's options for health care cost control: hope versus reality.” Ann Intern Med., 150(7), 485-9. Epub 2009 Mar 2. 107 Martin, J. (2007). Waking the SLEEPERS: Using Know4go to Incorporate Social, Legal, Ethical and Political Dimensions into Health technology Decision Making. Power Point presentation retrieved from http:              //www.cadth.ca/media/Symposium07/powerpoints/cadth_sym2007_monday- session2_Janet-Martin.pdf Massy-Westropp, M., Giles, L.C., Law, D., Phillips, P.A., & Crotty, M. (2005). Connecting hospital and community care: the acceptability of a regional data linkage scheme. Aust Health Rev., 29(1), 12-6. McGregor, M. (2003). Cost–utility analysis: Use QALYs only with great caution. Canadian Medical Association Journal, 168 (4). Melia, J., Pendry, L., Eiser, J.R., Harland, C., & Moss, S. (2000). Evaluation of primary prevention initiatives for skin cancer: a review from a UK perspective. Br J Dermatol., 143(4), 701-8. Mercer, K. (2008). Facilitating organizational mergers: amalgamation of community care access centres. J Health Serv Res Policy, 13 Suppl 1, 46-51. Mirtskhulava, V., Kempker, R., Shields, K.L., Leonard, M.K., Tsertsvadze, T., Del Rio, C., …Blumberg, H.M. (2008).  Prevalence and risk factors for latent tuberculosis infection among health care workers in Georgia. Int J Tuberc Lung Dis., 12(5), 513-519. Mitton, C. & Donaldson, C. (2003) Setting priorities and allocating resources in health regions: lessons from a project evaluating program budgeting and marginal analysis (PBMA). Health Policy, 64 (3), 335-348. Mitton, C., & Donaldson, C. (2004a). Health care priority setting: principles, practice and challenges.  Cost Effectiveness and resource Allocation, 2, 3-11. 108 Mitton, C., & Donaldson, C. (2004b). Priority setting Toolkit: A guide to the use of economics in health care decision making. London: BMJ Publishing Group. Mitton, C., Dionne, F., Peacock, P. & Sheps, S. (2006) Quality and Cost in Healthcare- A Relationship Worth Examining. Appl Health Econ Health Policy, 5 (4), 201- 208. Morgan, A., & Chow, S. (2007). The economic impact of implementing an ergonomic plan. Nurs Econ., 25(3), 150-6. Moore, J. (1998). Choosing the right ladder: creating a new model for healthcare. J Health Soc Policy, 10(1), 1-11. Neumann, P. J. (2005). Using cost-effectiveness analysis to improve health care: opportunities and barriers. New York, NY: Oxford University Press. O'Brien, B. J., Drummond, M. F., Labelle, R. J., & Willan, A. (1994). In search of power and significance: Issues in the design and analysis of stochastic cost-effectiveness studies in health care. Medical Care, 32(2), 150-163. O'Connor, A.M., Llewellyn-Thomas, H.A., & Flood, A.B (2004). Modifying unwarranted variations in health care: shared decision making using patient decision aids. Health Aff (Millwood), Suppl Web Exclusives, VAR63-72. Olds, D.L., Sadler, L., & Kitzman, H. (2007). Programs for parents of infants and toddlers: recent evidence from randomized trials. J Child Psychol Psychiatry, 48(3-4), 355-91. 109 Otis, J.D., Macdonald, A., & Dobscha, S.K. (2006). Integration and coordination of pain management in primary care. J Clin Psychol., 62(11), 1333-43. Peacock, S., & Edwards, D. (1997). Setting Priorities in South Australian Community Health I: The Mental health Program Budget. Program Evaluation, Research Report 13. Centre for Health Program Evaluation, Monash University, Australia. Peacock, S., Ruta, D., Mitton, C., Donaldson, C., Bate, A. & Murtagh, M. (2006). Using economics to set pragmatic and ethical priorities. BMJ, 332, 482-485. Peacock, S., Richardson, J.R., Carter R. & Edwards D. (2007) Priority setting in health care using multi-attribute utility theory and programme budgeting and marginal analysis (PBMA). Soc Sci Med. 64(4), 897-910. Peacock, S., Mitton, C., Bate, A., McCoy, B. & Donaldson, C. (2009). Overcoming barriers to priority setting using interdisciplinary methods. Health Policy, 92 (2), 124-132. Pearson, S. & Littlejohns, P. (2007). Reallocating resources: how should the National Institute for Health and Clinical Excellence guide disinvestment efforts in the National Health Service? Journal of Health Services Research & Policy, 12, 160- 165 Petersen, L.A., Woodard, L.D., Urech, T., Daw, C., & Sookanan, S. (2006). Does pay- for-performance improve the quality of health care? Ann Intern Med., 145(4), 265-72. Petrou, S. (1998). Health needs assessment is not required for priority setting. British Medical Journal, 317, 1154. 110 Phillips, C., & Thompson, G. (2009). What is a QALY? What is…? Series. Retrieved from http://www.evidence-based-medicine.co.uk/ebmfiles/WhatisaQALY Pine, G. J. (2009). Teacher Action Research: Building Knowledge Democracies. Thousand Oaks, California: Sage Publications Inc. Possamai, M.A. (2007). SARS and health worker safety: lessons for influenza pandemic planning and response. Healthc Pap., 8(1), 18-28. Rachlis, M. (2005). Prescription for Excellence: how innovation is saving Canada’s health care system. Toronto, ON: Harper Perrenial Canada. Ramer, S.L. (2005). “Site-ation pearl growing: methods and librarianship history and theory.” J Med Libr Assoc., 93(3), 397–400. Ray, W.A., Daugherty, J.R., & Meador, K.G. (2003). Effect of a mental health "carve- out" program on the continuity of antipsychotic therapy. N Engl J Med., 348(19),1885-94. Regopoulos, L.E., & Trude, S. (2004). Employers shift rising health care costs to workers: no long-term solution in sight. Issue Brief Cent Stud Health Syst Change, (83), 1-4. Reid, R.A. (1998). Using the tail to wag the dog: a consumer's perspective on management of the cost of cancer. Cancer, 82(10 Suppl), 2022-5. Reschovsky, J.D., Hadley, J., & Landon, B.E. (2006). Effects of compensation methods and physician group structure on physicians' perceived incentives to alter services to patients. Health Serv Res., 41(4 Pt 1), 1200-20. 111 Rice, N., & Smith, P.C. (2000). Capitation and risk adjustment in health care. Health Care Manag Sci., 3(2), 73-5. Richardson, J. (1997). Critique and Some Recent Contributions to the Theory of Cost Utility Analysis. Working Paper 77. Melbourne, Australia: Monash University. Ruser, C.B., Sanders, L., Brescia, G.R., Talbot, M., Hartman, K., Vivieros, K., & Bravata, D.M. (2005). Identification and management of overweight and obesity by internal medicine residents. J Gen Intern Med., 20(12),1139-41. Sabik, L. M. &. Lie, R. K. (2008). Priority setting in health care: Lessons from the experiences of eight countries. International Journal for Equity in Health, 7, 4. Schwartz-Shea, P. (2006). Judging Quality: Evaluative Criteria and Epistemic Communities. In Yanow, D. & P. Schwartz-Shea, Interpretation and Method, Empirical Research Methods and the Interpretative Turn (89-114). New-York, NY: M.E. Sharpe Inc. Sharman, Z. (2007).  Remembering the basics: Administrative technology and nursing care in a hospital emergency department. International Journal of Medical Informatics, 76s, s222-s228. Schumock, G.T., Michaud, J., & Guenette, A.J. (1999). Re-engineering: an opportunity to advance clinical practice in a community hospital. Am J Health Syst Pharm., 56(19), 1945-9. Shrank, W., Ettner, S.L., Slavin, P.H., & Kaplan, H.J. (2005). Effect of physician reimbursement methodology on the rate and cost of cataract surgery. Arch Ophthalmol., 123(12), 1733-8. 112 Sime, S.L. (2005). Strengthening the service continuum between transfusion providers and suppliers: enhancing the blood services network. Transfusion. 45(4 Suppl), 206S-23S. Simon, H. A. (1990). Bounded Rationality. In J. Eatwell, M. Milgate, & P. Newman (Eds.), Utility and Probability (15-18). New York, NY: The Macmillan Press Limited. Smith, A.L. (2007).. Merging P4P and disease management: how do you know which one is working? J Manag Care Pharm., 13(2 Suppl B),S7-10. Soumerai, S. (2003). Unintended outcomes of medicaid drug cost-containment policies on the chronically mentally ill. J Clin Psychiatry, 64 Suppl 17, 19-22. Speroff, T., Miles, P., & Mathews, B. (1998). Improving health care, Part 5: Applying the Dartmouth clinical improvement model to community health. Jt Comm J Qual Improv., 24(12), 679-703. Stevens, A., & Gillam, S. (1998). Health needs assessment; Needs assessment: from theory to practice. British Medical Journal, 316, 1448-1452. Susman, G.I. (1983). "Action Research: A Sociotechnical systems perspective," in G. Morgan (ed.), Beyond Method: Strategies for Social Science Research (95-113). London: Sage Publications. Tan, C.C. (2002). National disease management plans for key chronic non-communicable diseases in Singapore. Ann Acad Med Singapore, 31(4), 415-8. Teng, F., Mitton, C. & MacKenzie, J. (2007). Priority setting in the provincial health services authority: survey of key decision makers. BMC Health Serv Res., 7: 84. 113 Torrance, G. W. (1997). Preferences for health outcomes and cost-utility analysis. American Journal of Managed Care, May 3 Suppl, S8-20. Urquhart, B., Mitton, C., &  Peacock, S. (2008). Introducing priority setting and resource allocation in home and community care programs. Journal of Health Services Research and Policy, 13 Suppl 1, 41-45. Viney, R., Haas, M., & Mooney, G. (1995). Program budgeting and marginal analysis: A guide to resource allocation. New South Wales Public Health Bulletin, 6(4), 29- 32. Vonlanthen, R., Slankamenac, K., & Ernst, C. (2010). Qualitative and Economic Aspects of Liver Surgery. In S. Breitenstein, J. Belghiti, R. Chari, J. Llovet, C.-M.., Lo, M. Morse,…J.-N. Vauthey (Eds), Malignant Liver Tumors, Current and Emerging Therapies (3rd ed.). (531-538). Hoboken, NJ: Blackwell Publishing Ltd. Weinstock, M. (2006). Cutting both ways. Hosp Health Netw., 80(7),54-6, 58, 2. Wickizer, T.M., & Lessler, D. (2002). Utilization management: issues, effects, and future prospects. Annu Rev Public Health., 23, 233-54. Williams, J. (2007). Dollars and sense engaging physicians in supply-cost control. Healthc Financ Manage., 61(4), 62-8. Williams, J. (2008). A team approach to cost containment. Healthc Financ Manage., 62(4), 50-6, 58. Wilson, E. C. F., Peacock, S. J., & Ruta, D. (2009). Priority setting in practice: What is the best way to compare costs and benefits? Health Economics, 18, 467-478. 114 Wright, J., Williams, R., & Wilkinson, R. (1998). Health needs assessment. Development and importance of health needs assessment. British Medical Journal, 316, 1310- 1313. Wildavsky, A. (1977). Doing Better and Feeling Worse: The Political Pathology of Health Policy. In J. H. Knowles (Ed.), Doing better and feeling worse: Health in the United States (p. 105). New York: Norton. Yin, R. K. (2008). Case Study Research: Design and Methods. (4th ed.). Thousand Oaks, California: Sage Publications Inc. Yokoyama, K., Yang, W., Preblick, R., & Frech-Tamas, F. (2007). Effects of a step- therapy program for angiotensin receptor blockers on antihypertensive medication utilization patterns and cost of drug therapy. J Manag Care Pharm., 13(3), 235- 44. Young, G.J., Meterko, M., White, B., Bokhour, B.G., Sautter, K.M., Berlowitz, D., & Burgess, J.F. Jr. (2007). Physician attitudes toward pay-for-quality programs: perspectives from the front line. Med Care Res Rev., 64(3), 331-43. 115 APPENDIX A   The seven steps of the PBMA process (from Peacock et al., 2006) PREPARATION PHASE 1) Determine the aim and scope of the priority setting exercise—Will the analysis examine changes in services within a given programme or between programmes? 2) Compile a programme budget—The resources and costs of programmes combined with activity information 3) Form a marginal analysis advisory panel—The panel should include key stakeholders (managers, clinicians, consumers, etc) in the priority setting process 4) Determine locally relevant decision making criteria—The advisory panel determines local priorities (maximising benefits, improving access and equity, reducing waiting times, etc) with reference to national, regional, and local objectives IMPLEMENTATION PHASE 5) Identify where services could grow and where resources could be released through improved efficiency or scaling back or stopping some services—The panel uses the programme budget along with information on decision making objectives, evidence on benefits from service, changes in local health care needs, and policy guidance to highlight options for investment and disinvestment 6) Evaluate investments and disinvestments—Evaluate the costs and benefits for each option and make recommendations for change 7) Validate results and reallocate resources—Re-examine and validate evidence and judgments used in the process and reallocate resources according to cost-benefit ratios and other decision making criteria 116 APPENDIX B  VIHA’s organizational structure 1- Integrated Health Services HIGH INTENSITY AND REHABILITATION SERVICES Emergency Medicine Trauma Care Heart Health & Cardiac Services Adult Intensive Care Neurosciences Rehabilitation Services Access and Patient Flow PHARMACY, DIAGNOSTIC AND SURGICAL SERVICES Laboratory Medicine Medical Imaging Anaesthesia and Surgical Services Post Surgery Care Programs Pharmacy PRIMARY HEALTH CARE, POPULATION AND FAMILY HEALTH Primary Health Care & Chronic Disease Management Comox Valley Nursing Centre Health Point Care Centre Medical Health Officer/Public Health Community Care Facilities Licensing Health Protection and Environmental Services Public and Population Health Observatory Child, Youth and Family Health Prevention Services Healthy Children Aboriginal Health Rural Health CONTINUING HEALTH SERVICES Mental Health and Addiction Services Seniors Health Advance Directives Hillside Seniors Health Centre Spiritual Care 117 End of Life Care Victoria Hospice Residential Services Home and Community Care Assisted Living MEDICINE AND COMMUNITY HOSPITALS Medicine Community Hospitals/Health Centres Staff Scheduling QUALITY AND PATIENT SAFETY Infection Prevention and Control Quality Assurance and Process Improvement Continuing Professional Development for Physicians (previously Continuing Medical Education) University Liaisons Patient Care Quality Office (previously Client Relations) Information and Privacy Clinical Ethics System Quality and Patient Safety RESEARCH AND ACADEMIC DEVELOPMENT MEDICAL AFFAIRS Physician Compensation 2- Corporate and Strategic Services HUMAN RESOURCES AND ORGANIZATIONAL DEVELOPMENT Work Design & Compensation Employment Services HR Data & Benefits Management Consulting Services Wellness & Safety Learning & Development Organizational Development Workforce Planning HR Systems & Reporting Conflict Management Multimedia Services Library Services Value and Recognition 118 OPERATIONS & SUPPORT SERVICES Facilities Operations Materiel Management Business Development) Protection Services General Support Operational Services Health Emergency Management PROJECT MANAGEMENT OFFICE RJH Patient Care Centre PLANNING Strategic Planning Health Service Planning and Development Capital Planning Transformational Change Service & Program Integration Community Consultation & Coordination re: Service Delivery FINANCE Finance Performance Monitoring & Health Information Risk Management Decision Support Contract Management & Policy Development COMMUNICATIONS Issues Management Internal Communications Media Relations Community Relations Government Relations Web Content Management Visual Identity Standards INFORMATION MANAGEMENT / INFORMATION TECHNOLOGY Technology Management Application Management Client Services Management Project Delivery/Management Health Record Services Clinical Informatics/Telehealth 119 APPENDIX C  Examples of approved disinvestments at Menno Place Food Services Worker Retirement at Menno Home: Food Services Worker, presently 0.5 FTE, is anticipating retirement and will not be replaced. This would help to address the high number of FTE’s in the kitchen in comparison to industry standards.  Release $18,000 Recreation Coordinator and Recreation/Volunteer Manager: The Recreation Coordinator at Menno Home has resigned. A Recreation/Volunteer Manager will be hired and shared between Menno Home, Menno Hospital and Menno Housing. This change will provide a service that has been lacking at Menno Hospital and Menno Housing and take advantage of economies of scale. Net release $7,760 Staff Schedule/Full-Time Positions at Menno Home: Efforts will be made to maximize full-time positions by changing schedules. This would eliminate some of the part-time positions and lead to an associated savings in benefits. Release $40,000 Resident Chargeables at Menno Home: Charge the residents a nominal fee (i.e., $8) to cover personal hygiene products, dental hygiene treatment and cable. In comparison, Menno Hospital charges $18 monthly for personal hygiene products and socks and undershirts. Revenue $18,816 CRN Positions at Menno Home and Hospital and Professional Practice Leader: Eliminate Clinical Resource Nurse (CRN) positions at the Home (1.0 FTE) and at the Hospital (0.5 FTE) and create the position of a Professional Practice Leader to be shared by the Home and Hospital. This position will replace the services provided by the CRN’s and provide management leadership to contracted services. Net release $46,703 Food Services Coordinator Reduction in Hours at Menno Home: Reduce the Food Services Coordinator’s position from full-time to part-time (3 days per week). The coordinator can also pick up casual shifts as desired. This would help to address the high number of FTE’s in the kitchen in comparison to industry standards. Release $16,380 Resident Care Aid and Licensed Practical Nurse Positions at Menno Home: Release 6.0 FTE Resident Care Aid positions and increase Licensed Practical Nurse positions by 4.0 FTE.  These changes will more closely align the staffing mix with best practices in long-term care. Net Release: $83,392 120 Resident Care Aid Hours at Menno Hospital: Reduce 2 Resident Care Aid hours on day and evening shift per unit, for a total of 12 hours per day. These hours would be eliminated during the period in the daily schedule that has lower care demands. These changes will more closely align the staffing mix with best practices in long-term care. Release $91,761 Recreation Aide Hours at Menno Home: Recreation Aid hours equivalent to 0.8 FTE will be eliminated. Current Recreation Aid staff to resident ratio is significantly in excess of industry standard.  Resident Care Aids will be cross-trained so that they are able to take on aspects of recreational activities in the evening and on weekends. Release $41,179 Registered Nurse Hours at Menno Home: Registered Nurse hours on day shift will be reduced by 2 hours per day that is the equivalent of 0.4 FTE. These changes will more closely align the staffing mix with best practices in long-term care. Release $26,039 Relief Hours for Unit Clerks at Menno Home: There will be no relief for vacation time. There will be sick relief for one unit clerk when both unit clerks are away. Release $8,358 Temporary Resident Care Coordinator Position at Menno Home: One Resident Care Coordinator position at the Home will be eliminated. This position is currently vacant and will not be filled. This will bring Menno Home more in line with industry standards.  Release $97,297 121 APPENDIX D   Interview schedule for evaluation interviews at St. Joseph’s Health Centre DESCRIPTIVE Describe your involvement in the PBMA process? How would you describe your level of understanding of the process? How was the PBMA training and communication? How would you describe the role that PBMA played in the budgeting process? EVALUATIVE What are the strengths of the process in your view? Weaknesses? Was it fair? Do you think the allocation recommendations that resulted from the process will lead to better alignment with SJHC objectives? i.e., are they the ‘right’ decisions? Was there appropriate involvement in the process from various stakeholders? PROCESS DETAILS The PBMA recommendations involve some program cutbacks, do you have any comment, overall, on the way the process came to these specific recommendations? Were you involved in the development of options for resource releases? If yes: How challenging was it to come up with resource release proposal (s)? If challenging:  Why was it difficult? What were the main issues, concerns? (if it does not come up, we will ask specifically about the understanding regarding what was eligible for resource release proposals) If not challenging:  How did you approach the task of identifying resource release possibilities? What expenses were eligible for resource release proposals? 122 If no: Were you involved in the rating of resource release proposals? Yes: How would you describe the meaning of the resource release rating scores? Do you believe that these scores are a fair representation of the impact of the proposed changes? No: As an outsider to the process that led to the recommendations on resource releases, do you have any specific concerns with the actual recommendations? GOING FORWARD Does PBMA have a place at SJHC going forward? Do you have any suggestions for improvement?

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0071461/manifest

Comment

Related Items