UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The impact of electronic health record components on quality of patient care : a secondary data analysis Byard, Jillisa 2018

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

Item Metadata

Download

Media
24-ubc_2018_november_byard_jillisa.pdf [ 5.48MB ]
Metadata
JSON: 24-1.0372891.json
JSON-LD: 24-1.0372891-ld.json
RDF/XML (Pretty): 24-1.0372891-rdf.xml
RDF/JSON: 24-1.0372891-rdf.json
Turtle: 24-1.0372891-turtle.txt
N-Triples: 24-1.0372891-rdf-ntriples.txt
Original Record: 24-1.0372891-source.json
Full Text
24-1.0372891-fulltext.txt
Citation
24-1.0372891.ris

Full Text

THE IMPACT OF ELECTRONIC HEALTH RECORD COMPONENTS ON QUALITY OF PATIENT CARE: A SECONDARY DATA ANALYSIS  by  Jillisa Byard BScN, The University of Ottawa, 2012   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE  in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Nursing)    THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   October 2018  © Jillisa Byard, 2018 ii  Committee  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, a thesis/dissertation entitled:  The Impact of Electronic Health Record Components on Quality of Patient Care:  A Secondary Data Analysis  submitted by Jillisa Byard in partial fulfillment of the requirements for the degree of Masters of Science in Nursing  Examining Committee: Leanne Currie Supervisor  Angela Wolff Supervisory Committee Member  Glynda Rees Supervisory Committee Member Alison Phinney  Additional Examiner   iii  Abstract Aims:  The study examined the unique effects of electronic health record (EHR) components used by Canadian nurses in direct patient care on the quality of patient care, while controlling for individual nurses’ characteristics.  Specifically, the study aimed to measure how specific EHR components (i.e., computerized clinical decision support, computerized provider order entry, electronic medication administration record, nursing information systems) impact the quality of patient care.  Background:  The implementation of EHR systems and components in Canada are occurring at rapid pace with a goal to enhance quality, increase access and reduce costs. Research findings related to the impact of EHRs on quality of care remain inconclusive, as the majority of studies had small sample sizes and ambiguous descriptions of EHR components.  Methods: This cross-sectional secondary analysis drew upon data from 1031 direct patient care nurses from a Canadian survey conducted by Canada Health Infoway. Results:  One EHR component and three nurse characteristics were significantly associated with quality of patient care.  The clinical decision support tool was positively associated with quality of patient care.  Nursing experience was inversely associated to quality of patient care.  Experience documenting in EHRs and EHR satisfaction were positively associated with quality of patient care.  Conclusion: EHR component use, specifically clinical decision support systems, may impact quality of care. Nursing experience, experience documenting in EHRs and EHR satisfaction may also influence quality of care. Implications: Further research should be conducted to explore the impact of EHR use on nurses’ perception of quality of care.  Researchers should include adequate descriptions of their EHR iv  components and use psychometrically validated tools to increase the generalizability of their findings. Specifically, Infoway should continue to deploy their national nursing survey on health technology use to monitor EHR trends in Canada.     v  Lay Summary This study investigated how the use of electronic health record (EHR) components may impact nurses’ quality of care.  The results indicated that the use of clinical decision support was associated with higher quality of care.  While the use of nursing care plans and computer provider order entry may be associated with lower quality of care.  Furthermore, greater years of nursing experience was associated with lower quality of care, while experience documenting in EHRs and EHR satisfaction was associated with higher quality of care.  These results are consistent with some studies, but overall the evidence pertaining to EHR use and quality of care are inconsistent.  These results are important because it begins to explain how nurses’ characteristics and EHR use may impact quality of care in Canada.  This study can be used by Canada Health Infoway to improve their national survey on nurses’ health technology use.   vi  Preface  This thesis document is original, unpublished and independent work by the author Jillisa Byard.  This research study was conducted under the supervision of committee members Dr. Leanne Currie (Supervisor, Associate Professor, UBC School of Nursing), Dr. Angela Wolff (Committee Member, Clinical Assistant Professor, UBC School of Nursing) and Glynda Rees (Committee Member, Affiliate Instructor, UBC School of Nursing).  I conducted the literature review, cleaned and screened the raw data, and wrote chapters 1 to 5.  Design of the study and data analyses was conducted with collaboration with Dr. Leanne Currie and Dr. Angela Wolff.  All committee members reviewed and provided feedback of chapters 1 to 5.  The University of British Columbia research ethics board confirmed that formal ethics approval was not required for this study.   vii  Table of Contents Abstract ................................................................................................................................... iii Lay Summary ............................................................................................................................ v Preface ...................................................................................................................................... vi Table of Contents .................................................................................................................... vii List of Tables ............................................................................................................................ xi List of Figures ......................................................................................................................... xii Acknowledgements.................................................................................................................xiii Dedication ............................................................................................................................... xiv Chapter 1: Introduction ............................................................................................................ 1 1.1 Electronic Health Records in Canada ........................................................................... 3 1.2 Electronic Health Records and Quality of Patient Care ................................................ 5 1.3 Nurses’ Digital Health Technology Use: A Survey ...................................................... 6 1.4 Significance ................................................................................................................. 7 1.5 Purpose ........................................................................................................................ 7 1.6 Research Question ....................................................................................................... 8 1.7 Theoretical Framework ................................................................................................ 8 1.7.1 Rationale for Theoretical Model ........................................................................... 9 1.7.2 Operationalization of Study Variables ................................................................ 10 1.8 Summary ................................................................................................................... 11 Chapter 2: Literature Review ................................................................................................. 12 2.1 Introduction ............................................................................................................... 12 2.1.1 Identification of the Literature ............................................................................ 12 viii  2.2 Electronic Health Record Systems ............................................................................. 14 2.2.1 Electronic Health Records and Quality of Patient Care ....................................... 17 2.2.2 Critique of Electronic Health Record Research .................................................. 18 2.3 Computerized Clinical Decision Support Systems and Quality of Patient Care .......... 19 2.3.1 Critique of Computerized Clinical Decision Support System Research............... 21 2.4 Computer Provider Order Entry and Medication Administration Technologies .......... 22 2.4.1 Impact on Quality of Patient Care ...................................................................... 24 2.4.2 Critique of CPOE and BCMA Technology Research .......................................... 26 2.5 Electronic Nursing Information Systems and Quality of Patient Care ......................... 27 2.5.1 Critique of Nursing Information System Research.............................................. 29 2.6 Nurses’ Individual Characteristics and Quality of Patient Care .................................. 30 2.6.1 Critique of Nurses’ Characteristics in Research .................................................. 33 2.7 Summary ................................................................................................................... 33 Chapter 3: Methods ................................................................................................................ 35 3.1 Introduction ............................................................................................................... 35 3.2 Summary of Study Variables ..................................................................................... 35 3.3 Research Question ..................................................................................................... 41 3.4 Research Conceptual Model ...................................................................................... 42 3.5 Study Design ............................................................................................................. 43 3.6 Data Source ............................................................................................................... 43 3.6.1 Questionnaire Design ......................................................................................... 44 3.6.2 Sampling Methods in Primary Study .................................................................. 45 3.6.3 Recruitment Methods in Primary Study .............................................................. 46 ix  3.6.4 Data Collection Methods in Primary Study ........................................................ 46 3.7 Data Extraction for Secondary Analysis ..................................................................... 47 3.7.1 Sampling Methods of Primary Data for Secondary Analysis .............................. 47 3.8 Data Analysis ............................................................................................................ 48 3.8.1 Data Screening and Recoding ............................................................................ 48 3.8.2 Descriptive Statistics .......................................................................................... 49 3.8.3 Bivariate Analysis .............................................................................................. 50 3.8.4 Regression Analysis ........................................................................................... 50 3.9 Ethical Considerations ............................................................................................... 52 Chapter 4: Results ................................................................................................................... 54 4.1 Study Sample Descriptives ........................................................................................ 54 4.2 Descriptives of Quality of Care and EHR Components .............................................. 57 4.3 Study Variable Correlations ....................................................................................... 60 4.4 Hierarchical Regression Findings of Study Variables ................................................. 63 4.4.1 EHR Components and Quality of Care ............................................................... 63 4.4.2 Nurses’ Characteristics and Quality of Care ....................................................... 65 4.5 Post-Hoc Descriptives of Nurses’ Characteristics ....................................................... 66 4.5.1 Post-Hoc Power Analysis ................................................................................... 71 4.6 Summary of Findings ................................................................................................ 72 Chapter 5: Discussion ............................................................................................................. 73 5.1 EHR Components and Quality of Care....................................................................... 73 5.1.1 Computerized Provider Order Entry and Quality of Care ................................... 74 5.1.2 Nursing Information Systems and Quality of Care ............................................. 75 x  5.1.3 Computerized Decision Support and Quality of Care ......................................... 76 5.2 Nurses’ Individual Characteristics and Quality of Care .............................................. 76 5.2.1 Nursing Experience and Quality of Care ............................................................ 77 5.2.2 Experience Using EHRs and Quality of Care ..................................................... 79 5.2.3 EHR Satisfaction and Quality of Care ................................................................ 82 5.3 Strengths and Limitations .......................................................................................... 85 5.3.1 Strengths ............................................................................................................ 85 5.3.2 Limitations......................................................................................................... 86 5.4 Implications ............................................................................................................... 88 5.4.1 Nursing Research ............................................................................................... 88 5.4.2 Nursing Education ............................................................................................. 92 5.4.3 Nursing Practice................................................................................................. 93 5.4.4 Nursing and Healthcare Policy ........................................................................... 95 5.5 Conclusion ................................................................................................................ 97 References ............................................................................................................................... 98 Appendices ............................................................................................................................ 112 Appendix A Primary Study Questionnaire Tool ................................................................... 112 Appendix B Full Model of Multiple Regression Findings for Quality of Patient Care .......... 131  xi  List of Tables  Table 3.1 Primary Study Variables: Nurses’ Characteristics and Quality of Care ....................... 37 Table 3.2 Primary Study Variables: EHR Components .............................................................. 38 Table 3.3 Variable Definitions: Nurses’ Characteristics and Quality of PatientCare................... 39 Table 3.4 Variable Definitions: EHR Components .................................................................... 40 Table 4.1 Study Sample and National Population ...................................................................... 55 Table 4.2 Highest Nursing Designation ..................................................................................... 55 Table 4.3 Direct Patient Care Nurses’ Domain of Practice ......................................................... 56 Table 4.4 Domains of Practice and Quality of Patient Care Scores ............................................ 57 Table 4.5 Quality of Patient Care Frequencies ........................................................................... 57 Table 4.6 EHR Component Use Frequencies and Quality of Patient Care .................................. 59 Table 4.7 Pearson’s Correlations of Study Variables ................................................................. 62 Table 4.8 Final Model of Regression Findings Predicting Quality of Care ................................. 64 Table 4.9 Nurses’ Characteristics and Quality of Patient Care ................................................... 69 Table 4.10 EHR Experience and Quality of Patient Care by Years of Nursing Experience......... 70 Table 4.11 EHR Satisfaction and Quality of Patient Care by Years of Nursing Experience ........ 71  xii  List of Figures  Figure 2.1 Literature Search Strategy Flowchart ........................................................................ 14 Figure 3.1  EHR Components and Quality of Care: Conceptual Model ...................................... 42    xiii  Acknowledgements First, I would like to thank my thesis supervisor, Dr. Leanne Currie.  I am forever grateful for your encouragement and motivation. Thank you for all of the immeasurable experiences you have provided me over the last year.  Your mentoring has been truly invaluable, and your passion for informatics and nursing is truly inspirational.  Next, I would like to thank Dr. Angela Wolff.  I truly appreciate all of the educational opportunities you have afforded me.  Thank you for all of your efforts, especially in supporting the research design and analyses of this work.  I would also like to thank Glynda Rees. Your expertise has been crucial throughout this journey.  I appreciate all of the informatics and educational insights that you added to this work, and for my own personal learning.  To Dr. Susan Dahinten, for teaching me quantitative statistical analyses.  I have enjoyed working with you all immensely, and cannot thank you enough. Finally, I would like to thank all of the nurses to whom I have had the pleasure of working with. You have all given me inspiration to complete this work.  xiv  Dedication I would like to dedicate this work to my parents, Dave and Lynn Byard, who have always believed in me and my academic pursuits. I am forever grateful and proud to be your daughter. Finally, to my partner, Mitchell Bosch, for your unwavering patience and adoration. This journey would not have been the same without you all by my side, every step of the way. With my whole heart, thank you.    1  Chapter 1: Introduction  The use of electronic health records (EHRs) is increasing in Canada (Canada Health Infoway [Infoway], 2018c, 2018d; Digital Health Canada, 2017).  An EHR is a longitudinal, secure, and integrated collection of a person’s encounters with the health care system that ideally should provide a comprehensive and complete digital view of a patient’s health history, as well as support other healthcare related functions (Infoway, 2018c).  When EHRs replace paper-based health records the delivery of patient care changes through the transformation of clinician workflows and communication channels (Coiera, 2015; Health Informatics Society of Australia, 2017; International Medical Informatics Association, 2007; Nelson & Staggers, 2017).   Several researchers have explored the impact of EHRs on healthcare delivery (Adler-Milstein, Everson & Lee, 2015; Ancker et al. 2015; Bright et al., 2012; Institute of Medicine [IOM], 2012; Kazley & Ozcan, 2008; Kruse & Beane, 2018; Yanamadala, Morrison, Curtin, McDonald, & Hernandez- Boussard, 2016); EHRs may improve healthcare quality by increasing communication, collaboration, and continuity of care (Gagnon et al., 2012; Kazley & Ozcan, 2008; Kossman & Scheidenhelm, 2008; Rouleau et al., 2017; Yanamadala et al., 2016); best-practice guideline adherence (Adler-Milstein et al., 2015; Ancker et al., 2015; Bright et al., 2012; Kazley & Ozcan, 2008; Yanamadala et al., 2016; Zadvinskis, Chipps, Yen, 2014); health information access (Adler-Milstein et al., 2015; Bright et al., 2012; Gagnon et al., 2012; Kazley & Ozcan, 2008; Rouleau et al., 2017; Yanamadala et al., 2016); cost efficiencies (Adler-Milstein et al., 2015; Ancker et al., 2015; Bright et al., 2012; Gagnon et al., 2012; Kossman & Scheidenhelm, 2008; Kruse & Beane, 2018); and patient outcomes (Adler-Milstein et al., 2015; Ancker et al., 2015; Bright et al., 2012; Gagnon et al., 2012; Kazley & Ozcan, 2008; Kossman & 2  Scheidenhelm, 2008; Kruse & Beane, 2018; Rouleau et al., 2017; Yanamadala et al., 2016; Zadvinskis et al., 2014).  However, poor implementation strategies (Gagnon et al., 2012; IOM, 2012), usability issues (Ancker et al., 2015; Gagnon et al., 2012; IOM, 2012; Kossman & Scheidenhelm, 2008; Rouleau et al., 2017; Young, Slebodnik, & Sands, 2010; Zadvinskis et al., 2014), clinician non-adoption (Gagnon et al., 2012; Zadvinskis et al., 2014), and lack of end-user training (Ancker et al., 2015; Gagnon et al., 2012) can impede clinical work and contribute to technology-mediated adverse events (IOM, 2012; Kossman & Scheidenhelm, 2008; Young et al., 2010; Zadvinskis et al., 2014).   In the context of nursing, many researchers have explored the impact of EHRs on nursing processes such as documentation and communication (De Sousa, Dal Sasso, & Barro, 2012; Pillemer et al., 2012; Rouleau et al., 2017; Urquhart et al., 2009; Kossman & Scheidenhelm, 2008; Kutney-Lee & Kelly, 2011; Zadvinskis et al., 2014).  However, findings related to the impact of EHRs on quality of patient care remain inconclusive; several reports suggest that EHRs can enhance the quality of patient care (Fowler, Sohler, & Zarillo, 2009; Hessels, Flynn, Cimiotti, Bakken, & Gershon, 2015; Kirkendall, Goldenhar, Simon, Wheeler, & Spooner, 2013; Kutney-Lee & Kelly, 2011; Rantz et al., 2009), and yet many studies have mixed or non-significant results (Bouyer-Ferrullo, Androwich, & Dykes, 2015; Daly, 2002; Dowding, Turley, Garrido, 2011; Fossum, Alexander, Ehnfors, & Ehrenberg, 2011; Furukawa, Raghu, & Shao, 2010; Kossman & Scheidenhelm 2008; Pillemer et al., 2012; Rood, Bosman, Spoel, Taylor, & Zandstra, 2005; White & Mungall, 1990; Young et al., 2010; Zadvinskis et al., 2014).  Furthermore, the majority of studies had small sample sizes, ambiguous descriptions of EHR components and were not longitudinal studies which is the recommended research design to investigate post-implementation outcomes (Bouyer-Ferrullo, et al., 2015; Daly, 2002; Dowding 3  et al., 2011; Fossum et al., 2011; Hessels et al., 2015; Rouleau et al., 2017; Urquhart et al., 2009; Rantz et al., 2009; White & Mungall, 1990).  EHRs are in the process of being implemented across Canada, creating relevancy in studying nurses’ perceptions of the impact of EHRs on patient care.  Multiple authors use different terms for computer systems in healthcare including: health information technology, clinical information systems, eHealth, digital health, information communication technology, electronic health records, and electronic medical records (Nelson & Staggers, 2017; Rouleau et al., 2017).  For the purpose of this thesis, the acronym EHR will be used to refer to an electronic health record system which includes various EHR components that are used by nurses who provide direct patient care.  EHR systems can be broken down into specific components, and “each component of an EHR incorporates unique functions and attributes that contribute to the integration of a comprehensive patient record” (Nelson & Staggers, 2017, p. 92).  The EHR components explored in this thesis include computerized clinical decision support systems, computerized provider order entry, barcode medication administration technologies, and nursing information systems.  These are described in detail in Chapter 2.    1.1 Electronic Health Records in Canada Infoway, which is funded through Health Canada, collaborates with Health Canada, provincial Ministries of Health and health authorities to support the replacement of paper-based systems with interoperable digital health solutions including EHRs (Infoway, 2018d).  For instance, Infoway funded up to 75% of provincial health information technology projects in order to support this national goal (Auditor General of Canada, 2010).  Infoway utilizes the 4  Healthcare Information and Management Systems Society (HIMSS) electronic medical record adoption model (EMRAM) as a framework for monitoring EHR uptake in Canadian healthcare institutions (Powers, 2009).  The EMRAM model is comprised of eight stages ranging from level zero, which indicates that the organization has not installed any EHR components, up to level seven (Healthcare Information and Management Systems Analytics [HIMSS], 2018a).  EMRAM level seven indicates that there is a complete electronic medical record; that health information can be shared via standardized electronic transactions with external authorized health providers; and that the EHR system can analyze patterns of clinical data to improve health care efficiencies and patient safety (HIMSS, 2018a).  The implementation of EHR systems and components in Canada are occurring at rapid pace with a goal to enhance quality, increase access and reduce health systems costs (Infoway, 2018d).  As described above, the current evidence towards these goals remain inconclusive.  During this healthcare evolution, it is imperative that there is adequate evidence supporting nurses’ use of EHR components to provide quality care.  Infoway and the Canadian Nurses Association (CNA) continue to encourage EHR implementation and sustainability within the Canadian healthcare system to provide safe and patient centered care (CNA, 2017; Infoway, 2018d).  In a joint position statement from the Canadian Nursing Informatics Association and the CNA (2006; 2017), the groups suggest that EHRs provide nurses with access to timely, evidence-based information, which may result in safer patient care and better health outcomes.  According to a set of professional standards from the College of Registered Nurses of British Columbia (2017), nurses are required to document and report their nursing processes and client outcomes in a timely and appropriate manner.  There are nearly 400,000 regulated nurses in Canada, making nurses the largest end-users of EHRs (Canadian Institute for Health Information [CIHI], 2016a).  Therefore, it is crucial to 5  examine how nurses use EHRs in their practice so that healthcare funding agencies, healthcare leaders and clinicians understand potential benefits of, consequences of, and required supports for EHR adoption.  1.2 Electronic Health Records and Quality of Patient Care  Quality of patient care has many dimensions in nursing, including error reduction and patient safety, which can be directly affected by EHR components (De Sousa et al., 2012; Kazley & Ozkan, 2008; Kirkendall et al., 2013; Kossman & Scheidenhelm, 2008; Pillemer et al., 2012; Walker-Czyz, 2016).  In the nursing literature, the term quality of patient care has various meanings.  For example, one study conceptualized nurses’ assessed quality of care as both nursing processes like unfinished clinical tasks and nurse-sensitive patient outcomes (Wong, Laschinger, & Cummings, 2010).  In another study, nurses’ assessed quality of care included both quality-related effects such as patient satisfaction and nurse-sensitive measures such as fall incidents (Purdy, Laschinger, Finegan, Kerr, & Olivera, 2010).  Largely, quality of patient care can encompass both nurses’ perceptions and objective nurse-related patient outcomes (Kutney-Lee & Kelly, 2011; Lin, Chiou, Chen, & Yang, 2016; Purdy et al., 2010; Stalpers, Linden, Kaljouw, & Schuurmans, 2016; Wong et al., 2010).  For the purpose of this thesis, studies measuring objective nurse-sensitive patient outcomes will be considered part of the quality of patient care concept because nurses’ assessed quality of patient care often incorporates both subjective and objective measures of patient care (Kutney-Lee & Kelly, 2011; Lin et al., 2016; Stalpers et al., 2016).   Irurita (1999) suggests that the quality of nursing patient care is dependent on “contextual and intervening conditions pertaining to the broader environment, the organization and personal 6  factors of the nurse and patient” (Irurita, 1999).  As such, several factors have been examined related to quality of patient care including nurses’ years of experience, previous EHR experience, and EHR satisfaction.  Many researcher have studied how quality of patient care may be influenced by nurses’ demographics (Andre, Ringdal, Loge, Rannestad, Lareum, & Kaasa, 2008; Chow, Chin, Lee, Leung, & Tang, 2011; Daly et al., 2002; Eley, Fallon, Soar, Buikstra, & Hegney, 2008; Fowler et al., 2009; Kirkendall, 2013; Kossman et al., 2008; Moreland, Gallagher, Bena, Morrison, & Albert, 2012; Pillemer et al., 2012; Walker-Czyz, 2016; Zadvinskis et al., 2014).  Individual nurses’ characteristics such as nurses’ age and years of nursing experience (Chow et al., 2011; Eley et al., 2008; Kirkendall, 2013; Kossman et al., 2008; Lin et al., 2016), previous EHR experience (Chow et al., 2011; Eley et al., 2008; Kossman et al., 2008; Lin et al., 2016), and EHR satisfaction (Chow et al., 2011; Fowler et al., 2009; Holtz & Krein, 2011; Kirkendall, 2013; Lin et al., 2016; Walker-Czyz, 2016) have shown mixed or positive effects on quality of care and should be controlled during statistical analysis.  Further inquiry of how these factors influence EHR components and quality of patient care must be considered as they may affect patient outcomes.  The literature on quality of patient care and EHRs is further synthesized in Chapter 2 of this document.  1.3 Nurses’ Digital Health Technology Use: A Survey In 2014, Infoway conducted the “Nurses Digital Health Technology Use Survey” to explore nurses’ digital health technology use across all practice settings in Canada (Infoway, 2017a, 2017b).  In 2017, a second survey was conducted with slight changes to the survey questions to capture present EHR trends and more precisely capture nursing roles (Infoway, 2017b).  The 2017 Infoway survey questions captured nurses’ perceptions of EHR use in their 7  place of work and was completed by over two thousand nurses across Canada.  In this thesis study, a secondary analysis of the 2017 Infoway survey was performed to examine how specific EHR components influence nurses’ assessed quality of patient care.  The methods used in the original 2017 survey are explained in detail in Section 3.6 of this document.  1.4 Significance EHR adoption research has largely focused on physicians, despite nurses being the largest body of healthcare workers and EHR end-users (Yanamadela et al., 2016).  Many studies did not include rich descriptions of the specific EHR components studied, which may have contributed to the mixed results of the impact of EHRs on quality of care (Rouleau et al., 2017).  By conducting a secondary data analysis, this thesis will increase the body of knowledge surrounding EHR components and quality of patient care.  1.5 Purpose  Nurses have historically played an important role in ensuring patient safety and are the largest group of users of EHRs in Canadian healthcare settings (CIHI, 2016a; Kossman & Scheidenhelm, 2008).  The purpose of this study was to examine the unique effects of EHR components used by Canadian nurses in direct patient care roles on the quality of patient care, while controlling for individual nurses’ characteristics.  Specifically, the study aimed to measure how specific EHR components (i.e., computerized clinical decision support, computerized provider order entry, electronic medication administration record, nursing information systems) impact the quality of patient care.    8  1.6 Research Question This thesis addresses one research question:  Research Question 1: In direct patient care nursing, what are the relationships between EHR components (CDSS, CPOE, eMAR, NIS) and quality of patient care when controlling for nurses’ characteristics (nursing experience, experience viewing EHRs, experience documenting EHRs, EHR satisfaction)?  Upon examination of the results from the research question, a subsequent question was added. Subsequent Question: What is the association between quality of patient care and nurses’ characteristics (nursing experience, experience viewing EHRs, experience documenting EHRs, EHR satisfaction)?   1.7 Theoretical Framework When quantitative research studies are embedded within a theoretical framework, the findings are likely to have greater significance (Polit & Beck, 2017).  Rogers’ Diffusion of Innovations (DoI) (1962, 2003) theory was used in this study because it enables the framing of how innovations are adopted and how the consequences of an innovation may affect the social system.  In the past, nursing researchers have used Rogers’ DoI to examine the adoption of innovations in healthcare; such as new organizational standards of care and technologies (Lee, Lee, Lin, & Chang 2004; Ward, 2013).   Rogers (2003) posited that the individual’s decision to adopt or reject an innovation depends on receiver variables (individual and socioeconomic characteristics), social system variables, and perceived characteristics of innovation (relative advantage, compatibility, 9  complexity, trialability and observability).  Relative advantage is the degree to which an innovation is perceived as better than the current idea or practice; the greater the perceived relative advantage of an innovation, the more rapid its rate of adoption (Rogers, 2003).  If nurses perceive that their nursing care will improve through the use of EHRs, the more likely they are to use the EHR components to their full capabilities.  Consequences are viewed as changes that occur to an individual or social system as a result of the adoption or rejection of an innovation (Rogers, 2003).  Studying the consequences of an innovation is complex and should capture the perceived functionality in the user’s culture (Rogers, 2003).  In this study, quality of care can be viewed as a consequence of EHR use.  1.7.1 Rationale for Theoretical Model DoI is a useful framework to help explain the adoption of technology in nursing and may improve the likelihood of the innovation’s acceptance by guiding implementation methods (Doyle, Garrett, & Currie, 2014; Lee et al., 2004; Ward, 2013).  EHRs have been adopted in many settings in Canada, but the consequences of EHR implementation, and particularly in the nursing context, have yet to be analyzed using the DoI model.  By using Rogers’ DoI model as the framework of analysis, the outcomes of direct patient care nurses using EHR components as the innovation within their social system was examined.  As nurses’ perceptions of patient care quality are related to patient safety and nurse sensitive outcomes, it is important that the consequences of nurses utilizing EHRs are explored.  10  1.7.2 Operationalization of Study Variables Due to data limitations in the primary data set, this secondary analysis does not examine the applicability of Rogers’ DoI model as a whole, rather it utilizes DoI to analyze innovation consequences and to aid in conceptual clarity of the research variables.  The perceived relative advantage of the impact of EHR components on the quality of patient care is explored in this study.  Nurses are more likely to accept the innovation of an EHR system and components into their practice if they believe that their patient care will improve (Kaye, 2017).  Oppositely, nurses may reject EHRs if they perceive that their quality of patient care is at risk.  When the adoption of an innovation is not by choice, (e.g., the hospital chooses to adopt an EHR system that the nurses must use), rejecting this innovation may be demonstrated by using the EHR system or components in a manner that was not planned (such as performing work-arounds) (Rogers, 2003).   The quality of patient care is the consequence that occurs as a result of the adoption of EHRs.  According to Rogers (2003), depending on the individual and social system that adopts the innovation, the consequences can be direct or indirect; desirable or undesirable; and anticipated or unanticipated.  The consequences are viewed as the innovation outcomes, and therefore quality of care can be affected by EHRs in a negative, positive or neutral way.  The nurses’ characteristics are constructed as the individual receiver variables that influence the adoption of innovations.  It is important to examine individual factors because they are critical in influencing technology adoption in healthcare (Ward, 2013).  Therefore, nurses’ individual characteristics were statistically controlled when analyzing the relationship of EHR components on quality of patient care.  There could be several reasons why EHR component use impacts quality of patient care.  The individual characteristics of nurses may impact the relationship 11  between EHR component use and quality of patient care in such a way that is provides insight into the mechanisms as to how these outcomes occur.  1.8 Summary EHR implementation in healthcare settings is increasing in Canada.  In the nursing literature, the impacts of EHRs on quality of patient care largely remain inconclusive, and many studies have methodological weaknesses.  The purpose of this study was to examine the unique effects of EHR components used by Canadian nurses in direct patient care roles on the quality of patient care, while controlling for individual nurses’ characteristics.  The influence of EHR component use on quality of patient care is explained in part through the effects of nurses’ characteristics (nursing experience, experience viewing EHRs, experience documenting EHRs, EHR satisfaction).  These effects are explained within the DoI framework regarding the perceived relative advantage and consequences of EHRs in nursing, addressing an important gap in the literature.  The following chapter presents a review of the literature that is relevant to this topic.   12  Chapter 2: Literature Review  2.1 Introduction In this chapter a review of the literature is described to provide background to the use of EHR systems and components in nursing.  First, the literature search strategy is described. Second, EHR components are explained, and the evidence related to the relationship between EHR components and quality of patient care are synthesized.  The EHR components explored in this review included the following: computerized clinical decision support systems, computerized provider order entry, medication administration technologies, and electronic nursing information systems.  The relationships between nurses’ individual characteristics and quality of patient care are discussed.  Finally, the chapter concludes with current research limitations and gaps in the evidence which this study sought to fulfill.  2.1.1 Identification of the Literature A search of the literature up to March 15th, 2018 was conducted to find relevant information using three databases: PubMed, Cumulative Index of Nursing and Allied Health Literature (CINAHL), and Engineering Village.  Special attention was paid to informatics journals such as Computers, Informatics, Nursing, the Journal of the American Medical Informatics Association, and the International Journal of Medical Informatics by using the search terms in their journal webpage search box.  All searches were limited to the English language.  To retrieve a focused search yield, only subject headings (MeSH terms for PubMed and CINAHL terms for CINAHL) were included when possible in the search strategy: “nurses”; “electronic health record”; “health information systems”; “clinical information systems”; 13  “decision support systems, clinical”; “electronic order entry”; “nursing care plans, computerized”; “health information interoperability”; "electronic data interchange”; “quality of health care”; “quality of nursing care”.  All possible combinations of search terms were used.  Grey literature was reviewed from Infoway, Digital Health Canada, CNA, HIMSS, Health Informatics Society of Australia, and the International Medical Informatics Association to add perspectives from these organizations. The inclusion criteria were articles that used qualitative, quantitative or mixed method designs, as well as reviews and dissertations.  EHR systems and components had to possess patient specific software programs.  For the purpose of this review quality of patient care included nurse-sensitive patient outcomes and nurses’ assessed quality of patient care.  Only articles that took place in direct patient care settings and involved nurses were included. Therefore, articles that were interprofessional in nature were only included if they separated nursing results from other disciplines.  Articles that only measured nursing processes such as documentation time were excluded.  Relevant journal articles were first screened by title and abstract, then by reading the full text.  Additional studies were included through examining the included articles’ references.  A total of 27 articles were included in the literature review (see Figure 2.1 for search strategy flowchart).  This included 9 experimental studies, 8 non-experimental studies, 2 qualitative studies, and 8 reviews.  The majority of studies were from the US (n = 18), Norway (n = 2), Netherlands (n = 2) and Brazil, Canada, United Kingdom, Taiwan, and Australia had one study each.   14  Figure 2.1 Literature Search Strategy Flowchart   2.2 Electronic Health Record Systems EHRs can be broadly defined as a secure, longitudinal electronic record of patient health information (Coiera, 2015; HIMSS, 2018b; Infoway, 2018c).  A patient’s medical history, medications, diagnostic reports, and clinician progress notes produced by various healthcare Records from electronic database search: 206CINAHL: 81PUBMED: 29Engineering Village: 96Duplicates removed (n = 7)Not English language (n = 3)Screened articles by title & abstract (n = 196)Articles excluded (n = 124)Population not regulated nurses (n = 19)Did not identify EHR or quality of patient care (n = 81)Not in direct patient care settings (n = 24)Full text articles assessed for eligibility (n = 72) Articles excluded (n = 59)Population not regulated nurses (n = 15)Did not identify EHR or quality of patient care (n = 39)Not in direct patient care settings (n = 5 )Articles meeting inclusion criteria (n = 13) Articles added via reference list checking & grey literature  (n = 14)Articles included in Literature Review (n = 27)15  encounters are examples of information that can be found in EHRs (HIMSS, 2018; Infoway, 2018c).  EHR adoption is being encouraged by health organizations in hopes to improve health care access, efficiency and patient care (Infoway, 2018d).  There are many different EHR system configurations; from basic administration components, to systems that incorporate components with data analytics for evidence-based decision support and real-time quality outcome reporting (Linnen, 2016; Rouleau et al., 2017).   Healthcare institutions have been encouraged to adopt EHRs with the goal to streamline health access, and to improve quality of care (Dowding, Turley, & Garrido, 2012; Furukawa, Raghu, & Shao, 2010; Pillemer et al., 2012; Walker-Czyz, 2016).  Currently, EHR implementation research has concentrated on physicians, despite nurses being the largest group of end-users of EHRs (Dowding et al., 2012; Kirkendall et al., 2013; Walker-Czyz, 2016).  In nursing, researchers have focused on work processes such as time management, documentation quality, organizational efficiencies and costs (Dowding et al., 2012; Kossman & Scheidenhelm, 2008).  Subsequently, the literature on the impact of EHRs on patient care and nursing remains limited (Dowding et al., 2012; Furukawa et al., 2010). The influence of EHRs on the quality of patient care across healthcare settings have shown positive or mixed results (Dowding et al., 2012; Kirkendall et al. 2013; Kossman & Scheidenhelm, 2008; Pillemer et al., 2012; Rantz et al., 2009; Walker-Czyz, 2016).  In the study by Kirkendall et al. (2013) quality of patient care and patient safety increased after advancing from a partial EHR to a complete EHR system.  The perceived increase in patient safety and quality of care contributed towards the nurses’ overall positive attitude regarding the EHR.  Similarly, Kossman & Scheidenhelm (2008) analyzed medical, surgical and critical care nurses’ thoughts on the impact of EHRs on patient care.  The study revealed that nurses believed EHRs 16  provided safer patient care, but decreased the quality of patient care (Kossman & Scheidenhelm, 2008).  The findings by Kossman & Scheidenhelm (2008) indicated that nurses view electronic documentation as safer for patients, but also disruptive towards the patient-provider relationship.  This contradictory result should be further explored, as nurses may not use EHRs as designed if nurses think that patient care is reduced (Young et al., 2010).  The concept of quality of patient care varied between studies, as patient safety was thought to be separate from quality of care in one study (Kossman & Scheidenhelm, 2008), but was indicative of quality of care in the other (Kirkendall et al., 2013).  The mixed perceptions of EHRs and patient care quality implies that further research is required in order to discover which clinical contexts affect nurses’ assessed quality of care with respect to EHR systems.  In the nursing literature, EHR use and patient outcome research is scarce.  In long-term care settings, Pillemer et al. (2012) and Rantz et al. (2009) measured nurse-sensitive resident outcomes after EHR implementation which demonstrated inconsistent findings.  While residents’ activities of daily living and behavioural symptoms improved in the study by Rantz et al. (2009), activities of daily living were not significantly affected in the Pillemer et al. (2012) study, and behavioural symptoms were negatively affected.  Interestingly, the same sample of residents who exhibited worsened behavioural symptoms post EHR implementation in the Pillemer et al. (2012) study self-reported positive attitudes towards EHR in a questionnaire.  Multiple authors have suggested that the use of an EHR reduced pressure ulcer incidents across various healthcare settings (Dowding et al., 2012; Rantz et al., 2009; Walker-Czyz, 2016).  The incidents of patient falls were unaffected by EHR use in the studies by Dowding et al. (2012) and Pillermer et al. (2012), but significantly decreased post EHR implementation in the Walker-Czyz (2016) study.  Furthermore, EHR implementation decreased the incidents of nurse-sensitive hospital acquired 17  infections in two studies (Dowding et al., 2012; Walker-Czyz, 2016).  While some studies suggest patient care improvements after EHR implementation, there remain inconsistencies in the measurements and results of patient outcomes in others. Patient mortality following EHR implementation was non-significant in the study by Pillemer et al. (2012) but was found to be significant depending on the EMRAM implementation level in another study (Furukawa et al., 2010).  Specifically, compared to level 0 EMRAM, level 1 had no effect on patient mortality, but level 2 had a 17% lower mortality rate, followed by a 3-4% lower mortality rate in level 3 EMRAM systems (Furukawa et al., 2010).  With EHR adoption rapidly growing worldwide, it is important to acknowledge how different EHR systems and adoption levels may affect patient outcomes.  Despite similar findings of pressure ulcer and hospital acquired infection incidents decreasing in some studies, the current evidence of EHRs’ effect on patient outcomes and nurse-sensitive quality indicators were conflicting at times and warrant further investigation.   2.2.1 Electronic Health Records and Quality of Patient Care In the studies included in this review there were various quality of care definitions, but generally quality of patient care encompassed both subjective clinical standards and objective patient outcomes (Kutney-Lee & Kelly, 2011; Lin et al., 2016; Stalpers et al., 2016).  Findings from the study by Stalpers, Linden, Kaljouw, and Schuurmans (2016) indicated a significant positive association between objectively measured nurse-sensitive patient indicators and subjectively measured nurses’ perception of care quality.  Kutney-Lee & Kelly (2011) found that nurses in hospitals with fully implemented EHRs were significantly less likely to report medication errors and poor quality of care than nurses in hospitals without fully implemented 18  EHRs after controlling for hospital characteristics.  Nurse-sensitive indicators are defined as empirical evidence of patient outcomes that are based on and linked to nurses’ practice and processes (Stalpers, Linden, Kaljouw, & Schuurmans, 2016).  Nurse-sensitive indicators provide a quantitative basis to monitor and evaluate nursing care (Stalpers et al., 2016).  Nurse-reported measures are used to determine nurses’ perceptions and are referred to as subjective quality measures (Stalpers et al., 2016).  In the study by Lin, CHiou, Chen and Yang (2016), quality of nursing care referred to the nurses' evaluation of high-quality care in addition to objective assessments of patient care in their setting of work.  Furthermore, Lin et al. (2016) stated that patient safety referred to nurses' evaluation of nursing care not having a negative influence on patient health, assessed overall by the results of predicted patient outcomes.  For the purpose of this study, quality of patient care is a nurses’ self-assessment of quality of care that encompasses both subjective and objective measures of patient care and safety (Kutney-Lee & Kelly, 2011; Lin et al., 2016; Stalpers et al., 2016).    2.2.2 Critique of Electronic Health Record Research The current literature suggests there is inconclusive evidence regarding the impact of EHRs on nurses’ assessed quality of care and nurse-sensitive quality indicators.  Specific nurse-sensitive quality indicators in the literature varied significantly, even in similar healthcare settings.  The poor conceptualization of what constitutes quality of patient care poses difficulties in synthesizing patient outcomes in the literature (Rouleau et al., 2017).  Numerous articles focused on nursing processes such as documentation and workflow, as opposed to nursing outcomes (De Sousa, Dal Sasso, & Barra, 2012; Rouleau et al., 2017).  The assumption in the literature that efficient nursing care processes lead to better patient care had weak empirical 19  evidence (De Sousa et al., 2012; Rouleau et al., 2017).  Furthermore, nursing research was conducted in various healthcare settings and organizational contexts, which contributed to that lack of ability to generalize the findings.  The measurement tools used to obtain patient outcomes and nurse-sensitive quality indicators had strong psychometric properties, but the included studies used multiple measurement tools.  Lastly, the heterogeneity of EHR systems are one of many factors that contribute to the inconclusive evidence pertaining to the impact of EHRs on quality of care (Rouleau et al., 2017).  Subsequently, the quantity of research evidence regarding EHRs and quality of patient care to investigate remain low.  Alternatively, researchers are beginning to analyze specific EHR components as opposed to broader EHR systems.  These specific EHR components will be discussed in the next sections.  2.3 Computerized Clinical Decision Support Systems and Quality of Patient Care A computerized clinical decision support system (CDSS) provides clinicians with clinical knowledge and patient-specific information to enhance patient care through integrating best practice guidelines and prompts to assist with timely clinical action (Bouyer-Ferullo, Androwich, & Dykes, 2015; Fossum, Alexander, Ehnfors, & Ehrenberg, 2011).  Linnen (2016) argues that by adding big data analytic technologies to CDSSs, patient safety and nurse-sensitive care outcomes can be further improved.  Despite the promise of predicting nurse-sensitive patient outcomes, big data analytics in CDSS appear to be scant in the nursing literature (Linnen, 2016). Rouleau et al. (2017) found that many CDSS reports did not describe features or components of the technology, nor contexts of use.  The majority of CDSS studies focused on nursing care processes, access to information, time efficiency, clinical knowledge and critical thinking skills (Rouleau et al., 2017).  In the literature, using CDSSs to improve guideline 20  adherence such as vaccinations and diagnostic test ordering have shown positive results (Fossum et al., 2011).  However, the direct impact CDSSs had on patient care through guideline adherence and alert responsiveness largely remained unclear (Fossum et al., 2011; Rood et al., 2005).   CDSSs should provide nurses with individualized, evidence-based recommendations when and where the decision-making processes occur (Fossum et al., 2011).  Current studies have suggested that the impact of CDSSs on patient outcomes vary (Fossum et al., 2011; Rouleau et al., 2017).  In the study by Rood, Bosman, Spoel, Taylor, and Zandstra (2005), late blood glucose testing occurred 42% of the time when using a paper-based guideline compared to 28% of the time using a computerized guideline.  However, the patients’ blood glucose levels that fell within prescribed blood glucose target range remained the same (53% versus 54%) between the paper-based guidelines and CDSS (Rood et al., 2005).  This suggests that although there was an increase in guideline adherence with the CDSS, there was no significant effect on the patients’ clinical outcome.  Similarly, the studies by White and Mungall (1991) and Bouyer-Ferrullo et al. (2015) indicated that there was no significant difference in patient outcomes with CDSS use.  When comparing computerized guideline warfarin dosing versus a clinical nurse specialist, White and Mungall (1991) concluded that both the final prothrombin time (PTT) and target PTT were managed with comparable accuracy, despite the CDSS using mathematically standardized dosing regimens.  However, the researchers suggested that non-significant findings may have resulted from a crossover effect (White & Mungall, 1991).  In the study by Bouyer-Ferrullo et al. (2015), the researchers implemented a peripheral nerve injury specific CDSS in an operating room setting with expectations of decreasing nerve injuries.  Although the CDSS increased nursing 21  documentation and clinical knowledge in a significant way, there was no significant change in peripheral nerve injury incidences one year post-implementation (Bouyer-Ferrullo et al., 2015).  Fossum et al. (2011) also found an increase in documentation and guideline adherence for both pressure ulcers and nutrition following CDSS implementation.  However, the CDSS only significantly decreased the incidences of malnourishment, as pressure ulcer incidents remained non-significant (Fossum et al., 2011).  As mentioned above, the literature suggests that CDSSs can positively impact clinicians’ documentation and guideline adherence, but the impact of CDSSs on quality of patient care remained inconsistent among studies.  Although the study by White and Mungall (1991) is over twenty-five years old, it was one of the first studies to examine nurses’ use of CDSS.  Absent from the literature was nurses’ perspectives about the impact of CDSS components on patient care.  2.3.1 Critique of Computerized Clinical Decision Support System Research Although researchers at times hypothesizes that increasing guideline adherence will translate to improved patient outcomes, the empirical evidence to support this is limited.  This may relate to the fact that patient outcomes are not well-defined in the literature, and the majority of CDSS research in nursing measure nursing processes, not outcomes (Rouleau et al., 2017).  Most of the included studies that measured the effect of CDSSs on patient outcomes had inadequate power due to small sample sizes (Bouyer-Ferrullo et al., 2015; Fossum et al., 2011; White and Mungall, 1991), and were not longitudinal study designs (Bouyer-Ferrullo et al., 2015; Fossum et al., 2011).  Measuring results too soon after CDSS implementation may not capture full adoption contexts.  Although the literature favours the use of CDSS, the current evidence of the effects of CDSSs on quality of patient care remains inconclusive.  The current 22  body of literature did not include studies analyzing nurses’ assessed quality of care which could add a much needed perspective on CDSS and nurse-sensitive patient outcomes.  Furthermore, the studies by White and Mungall (1991) and Rood et al. (2005) are over 10 years old, and may not capture the current health technologies and clinical contexts of the present day.  Nursing research would benefit from a shift in perspective of measuring nursing processes to measuring nurse-sensitive quality indicators and nurses’ perceptions of patient care in order to increase the overall knowledge on CDSS components and quality of patient care.  2.4 Computer Provider Order Entry and Medication Administration Technologies Studies examined for the purpose of this thesis focused on computerized provider order entry (CPOE) that involved online ordering of medications, diagnostic tests, and patient care; where the prescriber is identifiable to others responsible for carrying out the orders (Nelson & Staggers, 2017; Stokowski, 2001).  These orders are not only legible, but have prompts to fill out necessary information, such as medication dosing.  CPOE can be integrated into existing EHR systems to align with other components such as CDSS and barcode medication administration technologies (Nelson & Staggers, 2017).  CPOE research is mostly concerned with investigating physician adoption due to their relatively large proportion of prescribing responsibility (Weir, Staggers, & Laukert, 2012).  For example, Stokowski (2001) referred to how the combination of CPOE and CDSS decreased prescribing errors up to 80% in one study, but did not mention any nurse related results.   The systematic review by Weir, Staggers, and Laukert (2012) noted that despite qualitative literature implying that CPOE can considerably influence multiple nursing processes such as communication and workflow, nursing research has tended to focus on medication 23  administration errors.  This is because the processes of medication administration are complex, creating multiple opportunities for medication errors to occur (Young et al., 2010).  A medication error is described as a preventable event that may cause incorrect medication administration and harm to a patient when a healthcare provider is responsible for their care (Fowler, Sohler, & Zarillo, 2009; Stokowski, 2001).  One study found that approximately 38% of medication errors were from medication administration errors, and only 2% of potential medication administration errors were caught before the drug was administered to the patient (Young et al., 2010).  As medication administration errors have profound patient consequences, the nursing literature has focused on this area when studying the effects of CPOE and barcode medication administration (BCMA) components on quality of patient care. BCMA technologies have been studied singularly or in conjunction with other EHR components in the nursing literature (Nelson & Staggers, 2017; Weir et al., 2012).  The BCMA administration process includes scanning a patient’s armband, the nurse’s identification badge, and the patient’s specific medications to verify that the drug for administration matches the actual drug order, which is then automatically signed in an electronic medication administration record (Fowler et al., 2009; Stokowski, 2001).  Similar to CPOE, the current literature about BCMA technologies emphasized medication administration error prevention because the main functions of barcode technologies are to accurately confirm steps in the medication administration processes (Young et al., 2010).  BCMA technologies were created to ensure that the medication administration system was closed-loop, which can decrease potential medication administration errors relating to the five medication verification procedures; right patient, right drug, right dose, right route, and right time (Fowler et al. 2009; Stokowski, 2001; Young et al., 2010). 24  Limited but growing research in nursing has acknowledged that BCMA technologies have usability issues and poor implementation strategies which can be major contributors to workarounds and system overrides, which can further contribute to medication administration errors (Fowler et al., 2009; Young et al., 2010).  In the literature, CPOE and BCMA functions are favourable in decreasing medication errors, leading to increased patient safety and improved patient outcomes (Fowler et al., 2009; Zadvinskis et al., 2014).    2.4.1 Impact on Quality of Patient Care In a phenomenological study by Zadvinskis, Chipps, and Yen (2014), nurses described how BCMA technologies increased the quality of patient care through medication administration error reduction.  Additionally, increased documentation efficiency and improved patient satisfaction also contributed to improved quality of care (Zadvinskis et al., 2014).  The BCMA technologies increased patients’ sense of safety through being involved in the identification armband scanning process (Zadvinskis et al., 2014).  However, nurses were also concerned that spending less time at the bedside, sleep disruptions from scanning the armband at night, and pushing the computer workstations to the bedside during family visits decreased their quality of patient care (Zadvinskis et al., 2014).  The nurses’ perception of quality of care were mixed, despite a perceived decrease in medication administration errors following the BCMA administration technology implementation.  To note, the study was conducted four months after the BCMA technology was implemented, which may not have been long enough for staff nurses to get fully acquainted to the changes in their workflow.  Fowler, Sohler, and Zarillo (2009) reported that nurses’ satisfaction increased due to their perception of patient safety improving after BCMA technology implementation.  This increase in 25  satisfaction correlated to the number of medication administration errors decreasing, suggesting that nurses’ assessed patient safety and medication administration error rates had a positive correlation (Fowler et al., 2009).  Nurses’ assessed quality of patient care should be further investigated as it may be closely related to medication administration errors and patient outcomes.  The most common patient outcome measured in BCMA technology research were nursing medication administration errors.  Studies used self-reported adverse event reports or BCMA system logs to capture medication administration error incidents (Fowler et al., 2009; Young et al., 2010).  Fowler et al. (2009) measured category C error incidents before and after BCMA technology implementation.  Category C medication administration errors are those that occurred but did not cause any harm to the patient (Fowler et al., 2009).  At two and a half months post-implementation, there was a significant decrease in multiple medication administration error stages including: documentation errors, wrong medication, wrong dose, wrong patient, wrong route, and wrong timing (Fowler et al., 2009).  There was an increase from 1 to 4 omissions of giving medications, which the researches largely associated with patients’ health conditions changing, or patients being away for procedures (Fowler et al., 2009).  It is unknown if more serious category errors were affected by this BCMA technology system.   Similarly, a longitudinal study implied that a BCMA technologies reduced medication administration errors by 71% over a 3-year period, but it was unclear which type of medication administration errors were measured (Stokowski, 2001).   The systematic review by Young et al. (2010) described the empirical findings of six quantitative studies; one study conducted in an intensive care unit reported that the medication administration error rates were reduced from 19.7% to 8.7% after BCMA implementation (P = < 26  0.001), and the second study conducted on a surgical ward reported that the medication administration errors decreased from 8.6% to 4.4% (p = < 0.01) (2010).  However, the third study showed non-significant results, and the fourth study indicated mixed results following BCMA technology use (2010).  The fifth study that was conducted in a neonatal intensive care unit exhibited an increase in medication administration errors from 69.5 to 79.9 per 1000 doses after BCMA implementation (p = < 0.001) and a sixth study did not report overall medication administration error rates (2010).  The review further analyzed medication administration error incidences by classifying them into the five medication verification rights, which similarly revealed mixed results (2010).  Comparable to the Fowler et al. (2009) study, the most common type of medication administration errors were wrong dose and wrong time, which researchers concluded were partially due to nurses adapting to patient situations and while deviating from specified order, may have improved patient care (Young et al., 2010).  Overall, the mixed results could be attributed to three out of six studies showing insufficient power, among other methodological weaknesses (Young et al., 2010).  Although some studies reported positive results, others revealed mixed findings of BCMA technologies, which warrants further evaluation of the impact on BCMA technologies on quality of patient care.  2.4.2  Critique of CPOE and BCMA Technology Research In the literature, CPOE has been studied with or without other EHR components, creating difficulties in generalizing the quality of care and patient outcome findings (Weir et al., 2012).  The review by Weir et al. (2012) recommended that studies describe the functionalities of their CPOE such as level of decision support and order set integration to combat this issue.  Although CPOE studies primarily focused on physicians prescribing, and BCMA technology research 27  focused on nursing medication administration errors, the medication administration process is inherently interdisciplinary, and therefore other organizational processes need to be considered when studies describe their results (Fowler et al., 2009; Weir et al., 2012).  BCMA technologies may contribute to medication administration errors if not user-friendly or if they were implemented improperly, and therefore researchers should supplement their findings with a detailed description of the technology’s design and training strategies to strengthen the overall evidence (Weir et al., 2012).  When studies rely on measuring medication administration errors with incident reports it may create self-reporting bias; as nurses tend to underreport near misses or errors that are not serious (Young et al. 2010).  Although BCMA technology adoption is rapidly increasing worldwide, there is still limited evidence regarding the effectiveness of these BCMA technologies in reducing adverse events (Young et al., 2010).  Currently, many gaps in the literature exist pertaining to how CPOE and BCMA components impact quality of patient care.  2.5 Electronic Nursing Information Systems and Quality of Patient Care The recognition that accurate and accessible health information can improve patient care has led institutions to invest in developing computerized nursing information systems (NIS) (Urquhart et al., 2009).  A NIS is a record of care that is used by nurses to document various nursing processes; communicate with other healthcare professionals; and include patients in their health plan (Urquhart et al., 2009).  For the purpose of this review, a NIS encompassed computerized nursing documentation platforms, nursing record systems, computerized care plans, and clinical pathways.  NISs can be incorporated with other EHR components to prompt evidence-based clinical decision making (Daly, Buckwalker, & Maas, 2002).  Integrating 28  standardized nursing languages into a NIS can benefit nurses by streamlining multiple terms into specific words and concepts (Daly et al., 2002).  These specific concepts provide a reference for nurses to use for documenting nursing processes, as well as create opportunities for nursing data to be analyzed for nurse-sensitive quality indicators (Daly et al., 2002).  For example, the CNA advocates for the use of the systematized nomenclature of medicine-clinical terms (SNOMED CT) and the international classification for nursing practice (ICNP) as the standardized clinical terminologies in Canada for EHR documentation (CNA, 2017).  Despite the potential advantages, integrating standardized nursing languages into electronic nursing documentation remain challenging due to various interoperability issues (Infoway, 2018d).  The systematic review by Urquhart et al. (2009) indicated that the current body of NIS literature focused on measuring documentation time, quality and accuracy; nursing assessment, diagnosis, and interventions; as well as interprofessional and nurse-patient communication.  These studies indicated that NISs may influence nursing practices and performance (Urquhart et al., 2009; Walters, 1986).  Walters (1986) discussed how efficient patient health information translated into efficient and quality patient care.  However, there was a significant gap in the research regarding the impact of NISs and nurse-sensitive patient outcomes (Daly et al., 2002; Urquhart et al., 2009).  Daly, Buckwalter and Mass (2002) studied the comparison of computerized nursing care plan versus a paper-based nursing care plan on patient outcomes in a long term care facility. Patient outcomes were measured through patient pain scores, activities of daily living, and cognition status through validated tools, while the number of medications, bowel function, nutritional status, and altered skin integrity were extracted via chart reviews (Daly et al., 2002). The findings of this study indicated that the use of standardized language in the computerized 29  care plan increased documentation of nursing processes compared to the paper-based version, but had non-significant effects on patient outcomes (Daly et al., 2002).  Since patient outcomes did not improve in the intervention group despite the significant increase in documentation, further evaluation between documentation and nursing action was recommended (Daly et al., 2002).  Similarly, Walters (1986) reported that when nurses are enabled to quantify and qualify their work, there is an increase in communication and continuity of care, thus leading to improved quality of patient care.  However, there was no evidence that documented nursing activities were related to the actual nursing interventions implemented to patients in this study (Walters, 1986).  The review by Urquhart et al. (2009) suggested that healthcare professionals believed that there is a positive relationship between using a NIS and the quality of patient care. The quality of documentation and quality of care may be affected by NISs with standardized languages, but the available evidence regarding patient outcomes and nurse-sensitive quality indicators were limited (Urquhart et al., 2009).  Further research is needed in order to link nurses’ assessed quality of care and objective nurse-sensitive patient outcomes by measuring quality indicators and NIS use.   2.5.1 Critique of Nursing Information System Research The current literature discussed the influence of NISs on nursing processes and performance, but failed to produce adequate evidence pertaining to the impact of NISs on patient outcomes (Daly et al., 2002; Urquhart et al., 2009; Walters, 1986).  Although researchers argued that simply changing paper documentation to electronic versions will not change patient care quality, integrating standardized nursing languages into a computerized care plan still resulted in 30  non-significant results in one study (Daly et al., 2002).  The heterogeneity of what comprises a NIS, as well as vague purposes of the NIS further contributed to the sparse evidence of NIS and quality of patient care (Urquhart et al., 2009).  Urquhart et al. (2009) recommended that more qualitative work should be conducted in order to gain a deeper understanding of how specific NIS functions influence nursing processes and attitudes.  Furthermore, Urquhart et al. (2009) advocated for quantitative studies in this field to increase their methodological rigour.  For example, Daly et al. (2002) incorporated psychometrically validated tools for measuring patient outcomes, but had a sample size of ten nurses per group who worked on the same hospital unit, creating possible crossover effects.  All of these recommendations could potentially enhance the conceptual clarity of NIS and patient outcome results in this body of literature.  Finally, by involving bedside nursing staff into the design and implementation process of NIS researchers may increase the adoption of NIS, which could facilitate more research opportunities.  2.6 Nurses’ Individual Characteristics and Quality of Patient Care In order to contextualize health research findings, it is essential to acknowledge extraneous influences on the impact of EHR systems and components on the quality of patient care.  Nurses’ characteristics reported in studies examining the effects of EHR systems and components on quality of patient care included: age, race, years of nursing experience, gender, educational level, previous computer experience, previous EHR experience, EHR user satisfaction, computer literacy and comfort level (Andre et al., 2008; Daly et al., 2002; Eley et al., 2009; Fowler et al., 2009; Kirkendall, 2013; Kossman et al., 2008; Lin et al., 2016; Stalpers et al., 2016; Walker-Czyz, 2016; Zadvinskis, 2013).  However, few studies included these variables in their qualitative analyses (Andre et al., 2008; Kossman et al., 2008), or quantitative 31  analyses (Eley et al., 2009; Fowler et al., 2009; Lin et al., 2016; Stalpers et al., 2016; Walker-Czyz, 2016).  The systematic reviews by Young et al. (2010), and Weir et al. (2012) did not report any nurses’ characteristics in the studies that measured the impact of CPOE and BCMA on patient outcomes.  In the systematic review of NISs by Urquhart et al. (2009), only one out of nine studies mentioned nurses’ characteristics as potential confounders in their analysis.  The effects of gender are not commonly analyzed in current nursing informatics research possibly because the nursing sample consists of mostly females, and nurses have similar work behaviours across all genders (Andre et al., 2008; Stalpers et al., 2016).   In an Australian nursing survey concerning EHR adoption, there was a significant negative correlation between nurses’ age and their perception of computer knowledge (i.e., older nurses had lower computer knowledge) (Eley et al., 2009).  Lin et al. (2016) also reported similar findings, in which nurses’ age had an inverse relationship to information system use satisfaction.  However, the study by Kossman et al. (2008) reported that age had no impact on EHR use.  In some studies, nurses’ age and nurses’ work experience were highly co-related and therefore one of the two items were not measured or included in the analysis (Andre et al., 2008; Stalpers et al., 2016).  Kirkendall et al. (2013) suggested that their nursing sample had an average of 10 years of working experience, and concluded that the sample was not biased toward less experienced nurses who are likely to be more computer oriented.  The researchers implied that younger and less experienced nurses may have higher quality of care and satisfaction scores than older and more experienced nurses when using EHRs without a thorough explanation (Kirkendall et al., 2013).  Similarly, the number of years worked in nursing had a significant inverse relationship to perceived computer knowledge in another study (Eley et al., 2009).  Nurses who had more work experience were more likely to lack the EHR skills and confidence compared to recently trained 32  nurses because their training occurred before computers were habitual (Eley et al., 2009).  In one study, nursing experience was positively associated with higher quality and satisfaction score (Stalpers et al., 2016).  Thus, the current evidence lacks unison.  The concept of user satisfaction is intended to measure users’ overall awareness and psychological state resulting in using the technology and appreciating its potential influence (Lin et al., 2016).  Nurse satisfaction was measured via turn-over rates in the study by Walker-Czyz (2016) and were inconsistent across different nursing units post-EHR implementation; despite better patient outcomes on all units, the medical-surgical units experienced an increase in nursing turnover, while the turn-over rates in the intensive care units remained unchanged, indicating an inconsistent relationship between turnover and user satisfaction (Walker-Czyz, 2016).  In the study by Fowler et al. (2009), nurses’ satisfaction had a positive correlation with perceived patient safety.  Nurses were generally more satisfied in hospitals with high scores on nurse-sensitive indicators, and least satisfied in lower-scoring hospitals (Stalpers et al., 2016).  Nursing satisfaction of NISs was found to be positively associated with nursing care performance (Lin et al., 2016), and was related to quality of patient care and safety. Kirkendall et al. (2013) described that their nursing sample had an average of five years’ experience using EHRs, which could have contributed to higher user satisfaction results after EHR implementation.  In another study, nurses with more computer experience are believed to have more positive attitudes to the use of the EHRs compared to nurses with lower computer experience (Lin et al., 2016).  Previous EHR experiences influenced nurses’ attitudes because it may affect nurses’ sense of self-efficacy, confidence, and autonomy when using computers (Andre et al., 2008; Eley et al., 2009).  Nurses who have EHR experience may have higher computer skills, user confidence and positive attitudes towards EHR components compared to 33  nurses who have less EHR experience (Andre, 2008; Eley et al., 2009).  Oppositely, Fowler et al. (2009) found no significant difference in user satisfaction in EHRs between nurses who had previous computer training versus no training.  2.6.1 Critique of Nurses’ Characteristics in Research Individual characteristics of nurses can impact their use of EHR components and therefore may inherently affect quality of patient care.  In the literature, nurses’ characteristics are often acknowledged as confounding variables, but infrequently included in the analysis.  Nurses’ work experience may have negative impacts on nurses’ perceived quality of patient care in two studies (Eley et al., 2009; Kirkendall et al., 2013) but positively affects quality care in another (Stalphers et al., 2016).  Whereas previous EHR experience and user satisfaction suggest mostly positive effects (Andre et al., 2008; Fowler et al., 2009; Lin et al., 2016; Stalpers et al., 2016) with the exception for the mixed results of user satisfaction in one study (Walker-Czyz, 2016).  Therefore, nursing experience, EHR experience, and user satisfaction should be controlled when evaluating quality of patient care as they may impact quality of patient care.  Studies should include confounding factors into their analysis at the structural, organizational and individual levels when possible to increase the body of knowledge on EHR components and quality of patient care in the clinical environment.  2.7 Summary EHR adoption is being encouraged by organizations such as Infoway with mixed empirical evidence supporting the positive impact of EHR components and quality of patient care.  Although this is beyond the scope of this thesis, usability issues and weak implementation 34  strategies were mentioned in numerous studies as contributors to nurses’ dissatisfaction of EHRs which could contribute to a decrease in patient safety.  Main limitations in the literature included conceptual ambiguity and heterogeneity of EHR components and functions; various patient outcome measurement tools; and a lack of methodological rigour.  It is necessary for the effects of EHR components on nurses’ assessed quality of care to be addressed because of the possible relation to nurse-sensitive patient outcomes and patient safety.  When applicable, studies should include organizational and individual factors to model the clinical contexts in which nursing work takes place.  Therefore, the purpose of this thesis is to increase the understanding of specific EHR components and quality of patient care to address one aspect of this evidence gap in the literature. 35  Chapter 3: Methods  3.1 Introduction In this chapter the research methods used to assess the effect of EHR components on quality of patient care are discussed.  The research questions and secondary data analysis study design are described.  The survey questionnaire and data collection procedures from the primary study are explained.  Finally, the secondary data analysis, and ethical considerations are discussed.  3.2 Summary of Study Variables In Tables 3.1 and 3.2 the primary survey items that are included in the secondary data analysis are presented.  The research variables used in the secondary analysis were not previously defined in the questionnaire survey of the primary study.  In Tables 3.3 and 3.4 the conceptual and operational definitions of the variable items according to the literature for this secondary analysis are presented.  EHR component items with similar functions are conceptually grouped together to accurately represent how nurses use EHRs in healthcare settings.  For the purpose of this thesis, each study variable was given a code for use in statistical analysis and a short term for discussion. The eleven EHR component variables were recoded as binary or dichotomous items (0= do not use, 1= use).  Dichotomous variables are composed of two points of qualitatively different categories (Tabachnik & Fidell, 2013).  The individual nurses’ characteristics were measured as continuous variables with higher numbers indicating higher values.  Tabachnik & Fidell (2013) define continuous variables as “… measured on a scale that changes values smoothly.  These 36  values are within the range of the scale and the size of the numbers reflect the amount of the variable” (p. 6).  Likert scales can be treated in behavioural research as continuous even though they are ordinal as long as the data: 1) still meets the normality assumptions; and 2) there are numerous categories and these categories represent quantitative attributes (Tabachnik & Fidell, 2013).  A detailed description of any variable recoding is in section 3.7 of this chapter. 37  Table 3.1 Primary Study Variables: Nurses’ Characteristics and Quality of Care Operational Variable Code Original Question from Primary Surveya Response Choices Years of nursing experience NsgExp How many years have you been a nurse? 1= less than 5  2= 5-15 3= 16-25 4= more than 25 Experience viewing or retrieving information from electronic health record ExpView How long have you been using electronic record /clinical information systems to view or retrieve patient information in your MAIN care setting? (Electronic records may include an EMR, Hospital Information System, ADT system, lab viewer, drug profile viewers) 1= < 1 year 2= 1- 2 years 3= 3- 4 years 4= 5- 6 years 5 = 6+ years Experience documenting information into electronic health record ExpDoc How long have you been using electronic record /clinical information systems to enter clinical documentation on patient encounters in your MAIN care setting? 1= <1 year 2= 1- 2 years 3= 3- 4 years 4= 5- 6 years 5 = 6+ years User satisfaction of EHR UserSat How satisfied are you with the electronic record/ clinical information systems that you currently use in your MAIN care setting? 1= highly satisfied 2= moderately satisfied 3= neither satisfied or dissatisfied 4= moderately dissatisfied 5= very dissatisfied Quality of patient care QoPC Because of your use of electronic record /clinical information systems, the quality of the patient care you provide is: 1= much better 2= better 3= no change 4= worse  5= much worse Note: a Primary Survey “National survey of Canadian Nurses: Use of digital health technology in practice”. By Canada Health Infoway, 2017, retrieved from https://www.infoway-inforoute.ca/en/component/edocman/resources/reports/benefits-evaluation/3345-2017-national-survey-of-canadian-nurses-data-set?Itemid=101   38  Table 3.2 Primary Study Variables: EHR Components Question from Primary Studya   Please indicate which of the following you use in your MAIN care setting to support patient care. (Check all that apply) Original Variable from Primary Study Code Response Choices Electronic reminders for recommended patient care following clinical practice guidelines (e.g. due for mammogram, complete falls risk assessment) CDSS_Rem 1= I currently use this functionality in my practice at a computer station 2= I currently use this functionality in my practice on a mobile platform 3= this is available electronically but I do not have access 4= not available electronically in my practice    5= don’t know/ unsure Electronic clinical decision support tool (e.g. dose calculator by weight, BMI calculator) CDSS_Tool Electronic warning for adverse prescribing and/or drug interactions CDSS_DI Electronic ordering/order entry of laboratory tests CPOE_Lab Electronic ordering/order entry of diagnostic imaging tests (e.g. CT, mammogram, MRI) CPOE_Dx Electronic ordering/order entry of patient care (e.g. turn and position every 2 hours) CPOE_PC Electronic order entry/prescribing of patient medications CPOE_Med Electronic list of all medications taken by an individual patient  eMAR Electronic clinical documentation (e.g. assessments, progress notes) NIS_Doc Electronic flow sheet or checklist for management of patients with chronic disease. (e.g. clinical care pathway) NIS_Flow Electronic patient care plans NIS_Plan Note: a Primary Survey “National survey of Canadian Nurses: Use of digital health technology in practice”. By Canada Health Infoway, 2017, retrieved from https://www.infoway-inforoute.ca/en/component/edocman/resources/reports/benefits-evaluation/3345-2017-national-survey-of-canadian-nurses-data-set?Itemid=101     39  Table 3.3 Variable Definitions: Nurses’ Characteristics and Quality of PatientCare Concept Primary Survey Variable(s) Term Operational Definition Scale Quality of Patient Care Quality of patient care Quality of care Quality of patient care is a nurses’ self-assessment of quality of care that encompasses both subjective and objective measures of patient care and safety (Kutney-Lee & Kelly, 2011; Lin et al., 2016; Stalpers et al., 2016).   1= greatly decreased 2= decreased 3= no change 4= increased  5= greatly increased Nurses’ Characteristics Years of nursing experience Nursing experience Years of experience working as a regulated nurse (includes working as a RN, NP, LPN and RPN) (CIHI, 2016a). 1= 0-5 years  2= 6-15 years 3= 16-25 years 4= 26+ years Experience viewing or retrieving information from electronic health record  Experience viewing EHRs Length in time viewing patient data through summary views or task specific views in the EHR; this includes accessing, searching, retrieving, viewing, and interpreting relevant patient information from an EHR in order to provide appropriate and timely nursing care (Coiera, 2015). 1= < 1 year 2= 1- 2 years 3= 3- 4 years 4= 5- 6 years 5 = 6+ years Experience documenting information into electronic health record    Experience documenting EHRs Length of time documenting patient data into the EHR by using structured templates or unstructured narrative charting in order to communicate with other healthcare providers regarding the patient’s healthcare encounter (Coeira, 2015). 1= <1 year 2= 1- 2 years 3= 3- 4 years 4= 5- 6 years 5 = 6+ years User satisfaction of EHR    EHR Satisfaction Nurses' satisfaction with an EHR refers to how nurses' perceptions and awareness of an EHR meets or exceeds their expectations and how it affects their psychological state (Lin et al., 2016).  1= very dissatisfied 2= dissatisfied 3= neither 4= satisfied 5= very satisfied    40  Table 3.4 Variable Definitions: EHR Components Concept Primary Survey Variable(s) Term Operational Definition  Scale Computerized Clinical Decision Support System (CDSS) • Electronic reminders from clinical practice guidelines  • Electronic clinical decision support tool  • Electronic warning for adverse drug interactions  • CDSS reminders  • CDSS tool  • CDSS drug interactions Computerized clinical decision support system (CDSS) is a point-of-care, evidence-based information tool that provide clinicians with individualized clinical practice guidelines and alerts which help inform real-time decision making and streamline clinical workflow (Infoway, 2018b).  Do not use=0 Use= 1  Computerized Provider Order Entry (CPOE) • Electronic order entry of lab tests  • Electronic order entry of diagnostic imaging  • Electronic order entry of patient care  • Electronic order entry/ prescribing of patient medications  • CPOE lab • CPOE imaging  • CPOE patient care • CPOE medication  CPOE applications enable health care providers to enter orders for services electronically and include using standardized order sets based on current best practice guidelines to support patient care (Infoway, 2018a).  Do not use=0 Use= 1  Electronic Medical Administration Record (eMAR) • Electronic list of patient medications (eMAR)  • eMAR Electronic medication administration record (eMAR) is a technology that automatically documents the administration of medication into EHR systems and provides automated information on the five medication administration rights (HIMSS, 2018b). Do not use=0 Use= 1  Electronic Nursing Information System (NIS) • Electronic clinical documentation   • Electronic flowsheet/checklist for management of patients with chronic disease  • Electronic patient care plans  • NIS documentation  • NIS flowsheets   • NIS care plans A nursing information system (NIS) is used to fulfill and document various nursing processes and standardized care by nurses to share health information with other healthcare providers and patients (Urquhart et al., 2009).  Do not use=0 Use= 1    41  3.3 Research Question One research question and four hypotheses were addressed: Research Question 1: In direct patient care nursing, what are the relationships between EHR component use (CDSS, CPOE, eMAR, NIS) and quality of patient care when controlling for nurses’ characteristics (nursing experience, experience viewing EHRs, experience documenting EHRs, EHR satisfaction)? Hypothesis 1a: Nurses’ use of CDSSs are positively associated with higher quality of patient care. Hypothesis 1b: Nurses’ use of CPOEs are positively associated with higher quality of patient care. Hypothesis 1c: Nurses’ use of eMARs are positively associated with higher quality of patient care. Hypothesis 1d: Nurses’ use of NISs are positively associated with higher quality of patient care.  Upon examination of the results from the research question, a subsequent question and two additional hypotheses were added. Subsequent Question: What is the association between quality of patient care and nurses’ characteristics (nursing experience, experience viewing EHRs, experience documenting EHRs, EHR satisfaction)?  Hypothesis 2a: Nurses with greater experience are positively associated with quality of patient care. Hypothesis 2b: Nurses with greater EHR experience (viewing and documenting) and EHR satisfaction are positively associated with quality of patient care. 42  3.4 Research Conceptual Model  The research conceptual model used for this thesis is shown in Figure 3.1.  The conceptual model indicates the direction and strength of the EHR components and nurses’ individual characteristics on quality of patient care.  Figure 3.1  EHR Components and Quality of Care: Conceptual Model Computerized Clinical Decision Support System• CDSS Reminders• CDSS Tool• CDSS Drug InteractionsComputerized Provider Order Entry• CPOE Lab• CPOE Imaging• CPOE Patient care• CPOE MedicationsNursing Information Systems• NIS Documentation• NIS Flowsheets• NIS Care PlansElectronic Medication Administration Record• eMARQuality of Patient CareGreater Years of Experience Viewing EHRsGreater Years of Experience Documenting EHRsGreater Years of Nursing Experience Greater EHR SatisfactionNurses’ IndividualCharacteristicsElectronic Health Record Components+++++-++43  3.5 Study Design As stated earlier, this thesis involved a secondary data analysis of a questionnaire that measured nurses’ digital health technology use that was conducted by Infoway in 2017.  This survey is described in full in Section 3.6 below.  Secondary data analysis is a form of research in which the data collected by one researcher are reanalyzed by another investigator to confirm the primary study’s original findings or to answer new research questions by using a subset of a larger quantitative data set (Polit & Beck, 2014).  For this thesis, the subset of data that was used from the Infoway (2017a) questionnaire included direct care nurses’ quality of patient care score and various EHR component use.  Nurses’ demographic data were also extracted from the data set to examine characteristics that could affect the relationship between EHR components and quality of patient care.  Working with an existing data set requires researchers to work within the confines of whatever study design and measures were chosen for the primary study (Doolan & Froelicher, 2009).  This secondary data analysis can add new knowledge to nursing informatics research literature without having to recruit participants and collect data, thus focusing the research efforts on research hypothesis testing and data analysis.  3.6 Data Source  As described earlier, the primary data source is the ‘Nurses Digital Health Technology Use’ Survey conducted in 2017 by Infoway.  The self-administered survey that was created for the primary study is explained here.  The results of a survey are limited by the extent to which the sample of the population are capable and willing to respond to the questions (Polit & Beck, 2017), and therefore it is imperative to describe the contents and structure of the primary study’s questionnaire tool.  The final version of the survey tool was piloted by a working group of fifty 44  nurses from the CNA and CNIA. This working group consisted of nurses in direct patient care, administration and researcher roles.  The pilot testing focused on comprehension of the survey questions and the skip logic format of survey.  There was no psychometric validation of the survey questionnaire by the parent study before it was deployed.    3.6.1 Questionnaire Design The 2017 Infoway questionnaire was designed to capture the current use of clinical information systems by nurses in Canada and across clinical practice settings by using a web-based survey.  This survey was revised from the first Canadian national survey which was conducted in 2014 (Infoway, 2017b).  Bennett et al. (2011) defined a survey as a systematic method for gathering information from a sample of entities for the purpose of constructing data of the attributes of the larger population.  The Infoway (2017a) questionnaire was a self-administered, cross-sectional survey with Likert scale questions, tick boxes, and open-ended comment sections for a total of 65 questions (see Appendix A).  The Infoway (2017a) questionnaire utilized a skip logic method in the survey tool depending on the nurses’ healthcare setting employment, work duties and EHR system use.  The Infoway (2017a) survey questions covered nurses’ demographics and their perceptions of clinical information system use in their place of work.  The final version of the 2017 Infoway survey is found in Appendix A for reference.  The survey is useful for secondary analysis because it was answered by over two thousand nurses across Canada, and represented multiple nursing positions, digital health technology experiences and attitudes.  45  3.6.2 Sampling Methods in Primary Study Canadian nurses of all practice domains who were English and French speaking were included in the primary study.  Regulated nurses are composed of registered nurses (including Nurse Practitioners), licensed practical nurses and registered psychiatric nurses (CIHI, 2016b) and therefore all of these nursing professions were included in the secondary analysis.   In Canada, there are three regulated nursing professions, but each province and territory has its own legislation governing nursing practice, as well as its own body that regulates and licenses its members (CIHI, 2016b).  Registered nurses (RNs) are self-regulated health care professionals who work both autonomously and in collaboration with others to enable individuals, families, groups, communities and populations to achieve their optimal levels of health (CIHI, 2016b).  Nurse practitioners (NPs) and clinical nurse specialists (CNSs) are RNs with additional educational preparation and clinical experience who possess and demonstrate increased competencies within their legislated scope of practice (CIHI, 2016b).  RNs are currently regulated in all 13 provinces and territories (CIHI, 2016b).  Licensed practical nurses (LPNs) assess, plan, implement and evaluate care for clients in a variety of healthcare settings and are currently regulated in all 13 provinces and territories, but in the province of Ontario, licensed practical nurses are referred to as registered practical nurses (CIHI, 2016b).  Registered psychiatric nurses (RPNs) focus on mental health and addictions and are currently regulated in Manitoba, Saskatchewan, Alberta, British Columbia, and the Yukon (CIHI, 2016b).  No a-priori calculation of sample size to determine adequate power was completed prior to survey deployment of primary study.   46  3.6.3 Recruitment Methods in Primary Study To recruit participants, the CNA and the Canadian Nurses Informatics Association emailed the online survey link as well as added the survey link to their electronic newsletter to all members (Infoway, 2017b).  The survey links were posted on their websites and social media platforms.  The CNA also coordinated with all provincial nursing associations in Canada to promote the survey via e-mail addresses and electronic newsletters to their members.  Infoway (2017b) also promoted the survey invitation nationally to clinical engagement networks, national electronic newsletters and social media postings.  For nurses in Canada without access to email or online promotions, an advertisement was printed in the January 2017 issue of Canadian Nurse with the survey link in English and French.  It was estimated by CNA that over 20,000 nurses subscribe and receive the national magazine (Infoway, 2017b).  To ensure national representativeness, invitations were also sent in March, 2017 to nurses who subscribe to EnsembleIQ's 'Nurse Registrant's' research lists.  In total, 2,863 English invitations were disseminated and 2,256 French language invitations were disseminated via this method (Infoway, 2017b).  Canadian nurses registered to mdBriefCase's online education platform (n=78) were also invited to participate in March, 2017 to ensure representative samples were achieved from the provinces of Alberta, Nova Scotia, New Brunswick, Newfoundland and Quebec (Infoway, 2017b).  The parent study utilized multiple recruitment strategies to ensure nurses from all provinces had opportunities to access the survey.  3.6.4 Data Collection Methods in Primary Study Survey data collection and primary data analysis was conducted by Environics Research Company of Canada who were commissioned by Infoway (2017b).  The collection of online 47  surveys occurred between January 6, 2017 to March 19, 2017.  There was no pre-notification of survey deployment before the data collection period began.  There was no financial compensation for those who completed the survey from the parent study.  It was unknown how many returned survey responses were excluded due to being incomplete or ineligible.  The response rate was unavailable due to the snowball sampling method of the primary study which resulted in an unknown denominator.  A total of 2,058 nurses completed the survey, and 65-67% of the total number of respondents indicated that they provide direct care (N= 1342).  The total number of direct patient care nurses who used either electronic documentation systems or partially electronic documentation systems was sizable (N= 1031).  Data collection and recruitment strategies were piloted in the primary study by a third party vendor during a soft launch but details are unknown.   3.7 Data Extraction for Secondary Analysis Data for this study were extracted from the Abacus Dataverse Network website where the SPSS data file was publicly available (Abacus Dataverse Network, 2017).  3.7.1 Sampling Methods of Primary Data for Secondary Analysis The term direct care refers to only those registrants who provided services directly to clients and the definition of direct care can vary by profession and province (CNA, 2016). Nurses who were employed in non-direct patient care roles such as education, management, and government were excluded from the sample for this secondary analysis.   As described earlier, skip logic was used in the online survey which affected the available data.  Nurses’ individual characteristic variables were answered by all nurses.  All nurses who 48  self-reported using only paper documentation systems were excluded from answering the quality of patient care question in the primary survey and were therefore excluded from the sample for the secondary analysis.  In order to maintain a consistent study sample throughout the statistical analyses, all items were split into direct patient care nurses who used partial or complete EHR systems for all variables in the secondary data analysis (N=1031).  Canadian nurses who provide direct patient care and who use partial or complete EHR systems were included in the sample for this secondary data analysis (N=1031).   The sample size requirement is a technique typically computed before data analysis (Cohen, 1988).  According to a-priori power calculation, with approximately 15 independent variables, a sample size of about 950 individuals would provide 80% power to detect small effect sizes (i.e., f 2 = .02) for multiple regression analysis at alpha=.05 (Cohen, 1988; Soper, 2011a).  Therefore, the secondary data analysis sample size (N=1031) had sufficient power to detect small effect sizes.    3.8 Data Analysis IBM SPPS version 25 was used for all data analysis procedures (IBM, 2017).  3.8.1 Data Screening and Recoding The raw data were screened for errors by checking frequencies for each variable and for scores that were out of range.  No extreme outliers were found in either univariate or multivariate situations.  Missing data points were confirmed by comparing the cleaned data set to the original survey skip logic and coded as missing = 999.  Missing data points were assessed and confirmed at less than 5% of the total data set, which is small enough to not impact the generalizability of 49  the large data set (Tabachnik & Fidell, 2013).  Therefore, no dummy coding, estimation or deletion of data points were required for this data set.   The variables were transformed and recoded when appropriate for statistical use with guidance from the thesis committee members.  For example, the eleven EHR component items had five categorical options originally: 1=used at computer station, and 2=used on mobile platform were then recoded to 1= used. While 3=available but not able to access, and 4= not available, were recoded into 0=not used.  The value 5= don’t know/ unsure was coded to 97=not applicable.  The following Likert scale items that were originally labelled as nominal measures were changed to scale measures for descriptive analysis: nursing experience, experience viewing EHRs, experience documenting EHRs, EHR satisfaction and quality of care.  The EHR satisfaction and quality of care items were originally inversely coded, these were recoded to measure the items as continuous variables with higher numbers indicating higher values (see Table 3.1).  3.8.2 Descriptive Statistics Preliminary analysis using descriptive statistics including means, standard deviations, histograms and boxplots were computed for each study variable to confirm normality (Polit, 2010).  Normality of variables can be determined via skewness and kurtosis (Tabachnik & Fidell, 2013).  Skewness is related to symmetry of a distribution, while kurtosis relates to the peakedness of a distribution (Tabachnik & Fidell, 2013; Polit, 2010).  When distribution is normal, both skewness and kurtosis values are at zero (Tabachnik & Fidell, 2013). However, in a large sample, the impact of departure from zero in both skewness and kurtosis will not make a substantive difference in the analysis (Tabachnik & Fidell, 2013).  In larger samples, researchers 50  are encouraged to assess variables for possessing the bell-shaped curve of a normal distribution via histograms (Tabachnik & Fidell, 2013).  Additionally, the central limit theorem states that with large sample sizes, sampling distributions of means are normally distributed regardless of the distributions of variables (Tabachnik & Fidell, 2013).    3.8.3 Bivariate Analysis  Correlation analyses were used to describe the strengths and direction of linear relationships and to assess for multicollinearity between variables (Polit, 2010).  Included variables were verified in meeting the assumptions of normal distribution because extreme groups and outliers can increase the magnitude of the coefficient (r), whereas a restricted range of variables reduces the magnitude of r (De Winter, Gosling, & Potter, 2016; Polit, 2010; Tabachnik & Fidell, 2013).  Sample correlations may be lower than population correlations when there is a restricted range in sampling of cases or very uneven splits in the dichotomous variables (Polit, 2010; Tabachnik & Fidell, 2013).  Pearson’s r was chosen to test for bivariate correlations because of the large sample size, continuous and dichotomous variable measures and lack of severely skewed values or outliers (Pallant, 2016; Polit, 2010; Tabachnik & Fidell, 2013).  3.8.4 Regression Analysis Hierarchical multiple regression was the main method of data analysis used to examine the relationship between EHR components, nurses’ characteristics and quality of patient care.  Sequential or hierarchical multiple regression is computed by entering the predictors into the model in a series of steps, with the order of entry based on theoretical or logical reasoning and controlled by the researcher (Tabachnik & Fidell, 2013).  Adding the predictors with a presumed 51  association to the dependent variable allows observations of what that predictor (or blocks of predictors) adds to the equation at the point in which it is entered (Polit, 2010).  The predictor variables in the final model are evaluated for what they add to the prediction over and above all the other variables (Tabachnik & Fidell, 2013).   In multiple regression analysis, beta coefficients, R2 and R2 change statistics were examined to determine if a relationship existed and the magnitude of the relationship.  The beta coefficients indicate the change in dependent variable for one standard deviation increase in each independent variable (Tabachnik & Fidell, 2013).  The R2 change statistic indicates the percentage of variance in the dependent variable that can be attributed to each variable, and the R2 statistic identifies the percentage of variance in the dependent variable attributed to the full model (Tabachnik & Fidell, 2013).   The suggested hierarchical regression sample size criteria is N  > 50 + 8(m) (if m= number of independent variables) (Tabachnik & Fidell, 2013).  The recommended sample size for this regression analysis would be 170 cases, therefore the sample size requirement was greatly exceeded (N=1,031).  This large sample size leaves adequate cushioning in case substantial measurement error is expected from less reliable variables or questionnaire tools (Tabachnik & Fidell, 2013).   Multicollinearity occurs when independent variables (IVs) are highly correlated (r = .9 and above), which can be assessed by conducting a bivariate correlations test (Pallant, 2016).  Given that multiple regression is sensitive to outliers (those with standardized residual values above 3.3 or below -3.3), it is imperative to transform outliers before proceeding with regression analysis (Tabachnik & Fidell, 2013).  Multicollinearity was not present, thus transformation was not required for this study. 52  Three additional assumptions can be examined through residual plots in SPSS, and should be computed pre-analysis to avoid outcome bias (Tabachnik & Fidell, 2013).  Normality is the assumption that each variable and all linear combinations of variables are normally distributed, and the residuals scatterplot should reveal a concentration of residual in the center of the plot at each value of predicted score and normal distribution of residuals trailing off symmetrically from the center (Polit, 2010; Tabachnik & Fidell, 2013).  Linearity is the assumption that there is a straight line relationship with all pairs of variables (Polit, 2010).  Failure of linearity of residuals in regression does not invalidate an analysis so much as weaken it; which means the power of the analysis is reduced to the extent that the analysis cannot map the linear relationships among the independent and dependent variables (Tabachnik & Fidell, 2013).  Homoscedasticity is the assumption that variance in scores for one should be the same for all predicted scores (Polit, 2010).  Oppositely, heteroscedasticity occurs when some variables are skewed and others are not, and serious heteroscedasticity occurs when the spread in SDs of residuals around predicted values is three times higher for the widest spread as for the narrowest spread (Tabachnik & Fidell, 2013).  All hierarchical regression assumptions were met for each study variable.  3.9 Ethical Considerations Completion and submission of the online survey from the primary study indicated implied consent.  All survey respondents remained anonymous via assigned identification numbers in the primary data file.  All electronic data were kept on a password protected computer.  There was no research ethics board submission by the parent study.  As per the University of British Columbia Behavioral Research Ethics Board, this research study did not 53  qualify as requiring a formal research ethics approval, and therefore no ethics application was completed for this study.  54  Chapter 4: Results  In this chapter, study findings are presented in three key sections.  In the first section, descriptive statistics and bivariate correlations between study predictors and outcome variables are reported.  In the second section, multiple regression findings are discussed with respect to the study outcome variable.  In the third section, descriptive and hierarchical regression findings are summarized and examined based on each study hypothesis and research question.  4.1 Study Sample Descriptives  The secondary data analysis sample was compared to the regulated nurses census for age, sex and Canadian province employed in Table 4.1 (CIHI, 2016c).  The findings suggest that the average age, sex and province employed in this study’s sample are similar to the Canadian national nursing population. Table 4.2 indicates the frequencies of nursing designation of all direct patient care nurses who use partial or complete EHRs.  The primary survey variable of highest nursing designation indicated that only 2.2% of survey respondents selected the “other” category (n=23).  When the data were further analyzed, it was found that those who selected the “other” category recorded job roles such as nurse educator and CNS titles.     55  Table 4.1 Study Sample and National Population  Characteristic f (%) M(SD) National Population a Age (years)  - 43.8 (11.3) 44.4  Gender      Male 70 (6.8%) - 7.8%  Female 955 (92.6%) - 92.2%  Transgender  0 (0.0%) - -  Prefer not to answer 6 (0.6%) - - Province or Territory Employed     Alberta 123 (11.9%) - 35,446 (11.9%)  British Columbia 87(8.4%) - 36,892 (12%)  Manitoba 85 (8.2%) - 12,870 (4.3%)  New Brunswick 38 (3.7%) - 8,150 (2.7%)  Newfoundland and Labrador 17 (1.6%) - 6,172.00 (2.1%)  Nova Scotia 45 (4.4%) - 9,635 (3.2%)  Nunavut and Northwest Territories  4 (0.4%) - 1,063 (0.4%)  Ontario 417 (40.4%) - 104,775 (35.1%)  Prince Edward Island 30 (2.9%) - 1,633 (0.6%)  Quebec 154 (14.9) - 70,980 (23.8%)  Saskatchewan 25 (2.4%) - 10,713 (3.6%)  Yukon 4 (0.4%) - 414 (0.1%)  Outside of Canada 2 (.2%) - - Note. N = 1031.  a Extracted from “Regulated Nurses, 2016: Registered Nurse/Nurse Practitioner Data Tables.” By Canadian Institute for Health Information, 2016. Retrieved from: https://www.cihi.ca/en/regulatednurses-2016-registered-nursenurse-practitioner-data-table   Table 4.2 Highest Nursing Designation Highest Nursing Designation  f (%)  Registered Nurse 827 (80.2%)  Nurse Practitioner 92 (8.9%)  Clinical Nurse Specialist 43 (4.2%)  Licensed Practical Nurse  41 (4%)  Registered Psychiatric Nurse 5 (.5%)  Other  23 (2.2%) Note. N= 1031.   56  Table 4.3 displays direct patient care nurses self-reported domain of practice.  Tables 4.3 and 4.4 show that clinical care was the most common domain of practice reported (n = 924) by the study sample of direct patient care nurses who use EHRs.  Administrative, policy, education, research, and clinical informatics domains of practice were also reported, and recoded into a non-clinical care group for comparison (n = 107).  A Mann-Whitney U test was used to compare quality of patient care between groups because if one group is more than 1.5 times greater than the second group, an independent t-test may produce erroneous results (Polit, 2010).  The Mann-Whitney U test revealed no significant differences in the quality of care scores of nurses who work in clinical care roles (Md = 4.00, n = 924) and non-clinical care roles (Md = 4.00, n = 107), U = 51977, z = .92,  p = .22.  Therefore, it suggests that including the direct patient care nurses who reported working in non-clinical domains of practice would not skew any statistical findings.  All direct patient care nurses were therefore included in the inferential statistical analysis.    Table 4.3 Direct Patient Care Nurses’ Domain of Practice Domain of Practice  f (%)  Clinical Care 924 (89.6%)  Nursing Administration 36 (3.5%)  Nursing Education 46 (4.5%)  Nursing Research 5 (.5%)  Nurse Policy 5 (.5%)  Clinical Informatics 15 (1.5%) Note. N= 1031.   Table 4.4 displays the total quality of patient care item, indicating that out of a possible score ranging from 1-5, the average was 3.64 (.91).   57  Table 4.4 Domains of Practice and Quality of Patient Care Scores   Quality of Patient Care Domains of Practice  n M(SD) MD Clinical Care 924 3.63 (.92) 4.00 Non-Clinical Care 107 3.68 (.91) 4.00 Both Groups 1031 3.64 (.91) 4.00 Note. N = 1031.   The quality of patient care frequencies are shown in Table 4.5 and suggest that the largest number of nurses reported their quality of patient care from using EHR components as increased (n= 395, 38.3%).  This suggests that on average, using EHRs increased nurses’ quality of patient care for the better.  Overall, over half of the nursing sample reported that their EHR use increased or greatly increased their quality of care (n = 583, 56.5%).   Table 4.5 Quality of Patient Care Frequencies Quality of Patient Care f (%) Greatly decreased 11 (1.1%) Decreased 91 (8.8%) Neither 346 (33.6%) Increased 395 (38.3%) Greatly increased 188 (18.2%) Note. N = 1031.   4.2 Descriptives of Quality of Care and EHR Components In relation to the primary research question, it is important to understand the relationships between quality of care and each EHR component (see Table 4.6).  The EHR component item that was used the most by direct patient care nurses was NIS clinical documentation (76.0%), followed by eMAR (67.3%) and CPOE lab (64.3%).  Oppositely, the EHR component used the 58  least by direct patient care nurses was NIS flowsheets (35.1%) followed by CDSS drug interactions (41.0%) and CDSS reminders (41.5%).  As shown in Table 4.6, quality of patient care is higher when nurses reported using CDSS, eMARs and NIS components compared to nurses who did not use those components.  However, the reported findings of quality of patient care with CPOE use is mixed.  Nurses reported lower quality of patient care when using the CPOE lab and the CPOE patient care items, but reported higher quality of patient care when using the CPOE medication and CPOE imaging items.  Further inferential analysis is conducted to help explain these descriptive findings and answer the research question and is presented in section 4.4.1.  59  Table 4.6 EHR Component Use Frequencies and Quality of Patient Care Electronic Health Record Component Quality of Patient Care Use Does Not Use f (%) M(SD) f (%) M(SD) Computerized Clinical Decision Support System CDSS Reminders 428 (41.5%) 3.72 (.96) 604 (58.5%) 3.58 (.87) CDSS Tool  524 (50.8%) 3.73 (.96) 508 (49.2%) 3.54 (.86) CDSS Drug Interactions 423 (41.0%) 3.70 (.95) 609 (59.0%) 3.59 (.89) Computerized Provider Order Entry CPOE Lab  664 (64.3%) 3.60 (.93) 368 (35.7%) 3.70 (.88) CPOE Imaging 555 (53.8%) 3.64 (.93) 477 (46.2%) 3.63 (.90) CPOE Patient Care 444 (43.0%) 3.62 (.99) 588 (57.0%) 3.65 (.85) CPOE Medications 451 (43.7%) 3.72 (.97) 581 (56.3%) 3.58 (.87) Electronic Medication Administration Record eMAR 695 (67.3%) 3.69 (.94) 337 (32.7%) 3.54 (.86) Nursing Information System NIS Documentation 784 (76.0%) 3.68 (.93) 248 (24%) 3.51 (.85) NIS Flowsheets 362 (35.1%) 3.67 (.96) 670 (64.9%) 3.62 (.89) NIS Care plans  487 (47.2%) 3.65 (.98) 545 (52.8%) 3.63 (.86) Note. N = 1031.   60   4.3 Study Variable Correlations  The linear relationships between study variables were examined using Pearson’s correlation analyses (see Table 4.7).  A correlation coefficient (r) of .10 represents a weak or small association; a correlation coefficient of .30 is considered a moderate association; and a correlation coefficient of .50 or larger represents a strong or large correlation (Polit, 2010).  A positive correlation coefficient indicates a positive relationship, and a negative correlation coefficient represents an inverse relationship between variables (Polit, 2010).   Pearson’s correlation computation indicated that quality of patient care was statistically significant between five EHR component items and all four nurses’ individual characteristic items.  With respect to the nurses’ characteristics; nursing experience had a weak, negative correlation to quality of care (r= -.11, p <.01), which suggests that higher nursing experience is associated with lower quality of care.  Both experience viewing EHRs (r= .10, p<.01) and experience documenting EHRs (r= .23, p<.01) had a small, positive relationship to quality of care.  This suggests that higher experience viewing EHRs and experience documenting EHRs are associated with higher quality of care.  EHR satisfaction had a moderate, positive relationship to quality of care (r= .42, p<.01), indicating that higher EHR satisfaction is associated with higher reported quality of care. In total, EHR satisfaction has the highest correlation to quality of care in this correlation analysis. The strongest correlation between quality of care and an EHR component is the CDSS tool (r= .11, p <.01) followed by the CDSS Reminders (r= .08, p<.05).  These statistically significant correlations indicate a small, positive association.  CPOE Medications (r= .08, p<.05), eMAR (r= .08, p<.05), and NIS documentation (r= .08, p<.05) all indicate a small, 61  positive relationship with quality of care.  These five EHR component items suggest that direct patient care nurses who use them have slightly higher quality of care scores compared to nurses who do not use them.  All other correlations involving quality of care were not statistically significant.   The correlations between nurses’ characteristics with one another had some statistically significant relationships worth noting.  For example, nursing experience (r=.32, p<.01) has a moderate, positive correlation to experience viewing EHRs, and a smaller, positive relationship with experience documenting EHRs (r=.22, p<.01).  This suggests that higher nursing experience is associated with higher experience viewing EHRs and experience documenting in EHRs.  EHR satisfaction has a small, positive correlation with experience documenting EHRs (r=.09, p<.01), suggesting that higher experience documenting EHRs is associated with higher EHR satisfaction.  A strong, positive relationship between experience viewing EHRs and experience documenting EHRs (r=.68, p<.01) suggests that higher years of experience viewing EHRs is associated with higher years of experience documenting EHRs.  Experience viewing and experience documenting EHRs has the highest correlation between nurses’ individual characteristics in this correlation analysis.  CPOE lab and CPOE imaging are highly correlated (r=.65, p<.01), indicating that CPOE lab use is positively associated with using CPOE imaging.  All other EHR component variables have small to moderate, positive correlation with each one another; with CDSS drug interactions and CPOE medications showing the second highest correlation between EHR components (r=.49, p<.01), and NIS documentation and CPOE imaging showing the lowest correlation between EHR components (r= .10, p<.01).  This suggests that EHR component use are associated with one another which may be related to how the EHR systems are designed.62  Table 4.7 Pearson’s Correlations of Study Variables  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1. NsgExp -                2. ExpView .32** -               3. ExpDoc .22** .68** -              4. UserSat 1.01 -.02 .09** -             5. QoPC -.11** .10** .23** .42** -            6. CDSS_Rem .01 -.01 .07* .20** .08* -           7.CDSS_Tool .01 .02 .04 .18** .11** .31** -          8. CDSS_DI -.02 -.03 .01 .19** .06 .33** .43** -         9. CPOE_Lab -.05 -.01 .00 .07* -.05 .25** .26** .27** -        10. CPOE_Dx -.08* -.04 .01 .09** .00 .27** .22** .30** .65** -       11. CPOE_PC -.11** -.04 .01 .07* -.01 .39** .28** .31** .42** .39** -      12. CPOE_Med -.02 -.07* .02 .20** .08* .35** .39** .49** .40** .42** .41** -     13. eMAR -.01 -.05 .01 .22** .08* .30** .31** .31** .23** .24** .21** .41** -    14. NIS_Doc .03 -.04 .16** .21** .08* .33** .24** .19** .11** .10** .28** .26** .27** -   15. NIS_Flow -.03 .04 .09** .17** .02 .41** .35** .35** .28** .31** .36** .37** .27** .29** -  16. NIS_Plan -.18** -.06* .03 .14** .01 .37** .25** .24** .19** .15** .42** .29** .22** .32** .42** - Note. N = 1031.  NsgExp= Nursing experience.  ExpView= Experience viewing EHRs. ExpDoc= Experience documenting EHRs.  UserSat= EHR satisfaction.  QoPC= Quality of patient care.  CDSS_Rem= CDSS reminders (0= does not use, 1= use).  CDSS_tool= CDSS tool , (0= does not use, 1= use).  CDSS_DI= CDSS drug interactions, (0= does not use, 1= use).  CPOE_Lab= CDSS lab, (0= does not use, 1= use).  CPOE_Dx= CPOE imaging, (0= does not use, 1= use).  CPOE_PC= CPOE patient care, (0= does not use, 1= use).  CPOE_Med= CPOE medications, (0= does not use, 1= use).  eMAR (0= does not use, 1= use).  NIS_Doc= NIS documentation, (0= does not use, 1= use).  NIS_Flow=  NIS flowsheets, (0= does not use, 1= use). NIS_Plan= NIS care plans, (0= does not use, 1= use). *p < .05, ** p < .01 for two tailed statistical significance.63  4.4 Hierarchical Regression Findings of Study Variables In multiple regression analysis, beta coefficients, R2 and R2 change statistics were examined to determine if relationships existed and the magnitude of the relationships if present.  Multiple regression coefficients were interpreted to reflect the number of units that the quality of patient care was expected to change for each unit change in a predictor variable when the effects of other predictors were held constant or statistically controlled (Polit, 2010; Tabachnik & Fidell, 2013).    4.4.1 EHR Components and Quality of Care  Hierarchical multiple regression was used to determine the relationship between the EHR components (4 CPOEs, 3 NISs, 3 CDSSs, and 1 eMAR) on quality of care.  As the DoI model does not include theoretical reasoning for entering the variable sequences, logical reasoning was applied based on a review of the correlations’ strength and direction (see Table 4.8).  The EHR component items were then organized into their EHR component groupings.  In accordance with Tabachnick and Fidell (2013), the EHR component variables were entered into the regression model beginning with the smallest to largest correlations to quality of patient care (see Appendix B).  Next, the nurses’ characteristics (nursing experience, experience viewing EHRs, experience documenting EHRs, and EHR satisfaction) were added to the regression from smallest correlations to largest to assess their impact on EHR components and quality of care.  The total variance (R2) was statistically significant in all seven models of the regression analysis (see Appendix B).   64  Table 4.8 Final Model of Regression Findings Predicting Quality of Care Predictors B SE B b CI (95%) R2     LL UL       .26 CPOE Lab -.20 .07 -.10** -.34 -.06  CPOE Imaging .04 .07 .02 -.10 .17  CPOE Patient Care -.03 .06 -.02 -.15 .10  CPOE Medications .07 .07 .04 -.06 .20  NIS Documentation -.06 .07 -.03 -.19 .08  NIS Flowsheets -.11 .06 -.06 -.24 .01  NIS Care Plans -.13 .06 -.07* -.25 -.01  CDSS Reminders .07 .06 .04 -.05 .19  CDSS Tool .16 .06 .09** .04 .27  CDSS Drug Interactions -.05 .06 -.03 -.17 .08  eMAR -.00 .06 -.00 -.12 .12  Nursing Experience -.15 .03 -.18*** -.20 -.10  Experience Viewing EHRs -.00 .03 -.00 -.05 .05  Experience Documenting EHRs .15 .02 .24*** .10 .19  EHR Satisfaction .33 .02 .41*** .28 .37  Note. N= 1031. b= standardized beta coefficient. CI= confidence intervals, LL= lower limit, UL = upper limit. Model 7: F(15, 1015)= 24.03, p<.001.   *p<.05, **p<.01, ***p<.001  Research Question 1:  The final model for examining the relationship between EHR components (CPOE, NIS, CDSS, eMAR) and quality of patient care yielded three EHR components that were statistically significant: CPOE lab, NIS care plans and CDSS tool (see Table 4.8).  They are reported by the research hypotheses. Hypothesis 1a: Of the three CDSS components available, only the CDSS tool (b= .09, p<.01) was positively associated with quality of care and statistically significant.  On average, nurses who used the CDSS tool had quality of care scores that were .09 SD units higher than nurses who do not use the CDSS tool after controlling for other variables in the model.  The availability of evidence based information tools, which may include alerts and clinical practice 65  guidelines to inform real-time decision-making is a significant predictors of quality of care.  Since CDSS reminders and CDSS drug interactions were both insignificant, hypothesis 1a is partially accepted. Hypothesis 1b: Of the four variables in the CPOE components available, the CPOE lab was the only statistically significant variable; however, it was not in the direction hypothesized.  Specifically, CPOE lab (b=-.10, p<.01) is inversely associated with quality of care.  This suggests that nurses who used CPOE lab have quality of care scores that were .1 SD units lower than nurses who do not use CPOE lab after controlling for all other predictors.  Hypothesis 1b is rejected because zero out of four CPOE components increased quality of care. Hypothesis 1c: This hypothesis was rejected as the eMAR component was not a statistically significant predictor of quality of patient care. Hypothesis 1d:  Of the three NIS components available, only the NIS care plans was statistically significant (b= -.07, p< .05) and computed an inverse relationship to quality of care. This suggested that nurses who used NIS care plans have quality of care scores that are .07 SD units lower than nurses who do not use care plans after controlling for all other predictors.  As a result, those who used standardized care plans report lower quality care.  NIS flowsheets and NIS documentation were found to be insignificant.  Hypothesis 1d is rejected because zero out of three NIS components positively predicted quality of care.  4.4.2 Nurses’ Characteristics and Quality of Care Upon examination of the results from the research question, a subsequent question and two additional hypotheses were added. 66  Subsequent Question:  What is the association between quality of patient care and nurses’ characteristics (nursing experience, experience viewing EHRs, experience documenting EHRs, EHR satisfaction)? Hypothesis 2a: Nursing experience was statistically significant and inversely associated with quality of care (b= -.18, p<.001), indicating that for every 1 unit increase in nursing experience, nurses reported a decrease in quality of care by .18 SD units, after controlling for all other variables.  Therefore, hypothesis 2a was accepted.  Hypothesis 2b: Experience documenting EHRs remained a statistically significant predictor of quality of care (b= .24, p<.001).  This suggested that for every 1 unit increase in experience documenting EHRs, nurses reported an increase in quality of care by .24 SD units after controlling for all other variables in this model.  Experience viewing EHRs was not statistically significant in the regression model.  In the final model, EHR satisfaction explained an additional R2Δ of 15% FΔ(1, 1015)= 204.98, p<.001, for a total R2 of 26% F(15, 1015)= 24.03, p<.001.  EHR satisfaction recorded the highest beta value over and above all predictors (b= .41, p< .001) of quality of care.   This suggests that on average, for every 1 unit increase in EHR satisfaction, quality of care increased by .41 SD units after controlling for all other predictors.  In this regression analysis, the EHR satisfaction variable displayed had the highest statistical significance in predicting quality of care (p<.001).  Therefore, hypothesis 2b was partially accepted.  4.5 Post-Hoc Descriptives of Nurses’ Characteristics To answer the subsequent question to this study, that is, to understand the association between quality of patient care and nurses’ characteristics (nursing experience, experience 67  viewing EHRs, experience documenting EHRs, EHR satisfaction), several additional sub-analyses were examined. The frequencies of nurses’ characteristics and quality of care (see Table 4.9) indicates that 34.5% of the nursing sample reported their nursing experience to be greater than 26 years, which is the largest category of the nursing experience item.  In the last column of Table 4.9, the nursing experience item indicates lower quality of patient care as nursing experience increases.  For example, the lowest quality of patient care reported was from nurses had more than 25 years of experience (M=3.50).  The highest quality of patient care score (M= 3.72) is when nurses reported possessing 5-15 years of nursing experience.  This suggests that nursing experience may be inversely related to quality of patient care.  The greatest category of experience viewing EHRs (41.8%) and documenting EHRs (29.5%) were over six years of experience.  Out of a possible range of 1-5, experience viewing EHRs has a slightly higher average (M= 3.62) than experience documenting EHRs (M=3.12), indicating that nurses may have greater years of experience viewing patient information than documenting patient information into EHR systems.  In Table 4.9, an increase in experience viewing EHRs shows an increase in quality of patient care, except for the 5-6 year category.  Overall, the quality of patient care increases as experience documenting EHRs increases.  This suggests that nurses who have more experience documenting into EHR systems, reported higher quality of patient care.   To further investigate how nursing experience may inversely affect quality of patient care, the quality of patient care scores are reported by experience viewing EHRs and the experience documenting EHRs by each nursing experience category (see Table 4.10).  Although some categories possess small samples sizes, two overarching themes emerged.  First, the quality of patient care decreases as the nursing experience categories increase.  Second, within each 68  nursing experience category, the quality of patient care increases as experience viewing and documenting EHRs increases, with few exceptions.  The increase in quality of patient care with higher years of experience documenting EHRs is more consistent than the experience viewing EHRs item.  The highest quality of patient care was reported by nurses who have less than five years of experience and have 5-6 years of experience viewing EHRs (M= 4.20) as well as over 6 years of experience documenting EHRs (M= 4.20).  Oppositely, the lowest quality of patient care was reported by nurses who had more than 25 years of nursing experience and less than one year of experience viewing EHRs (M= 2.72).  Greater nursing experience is negatively associated with quality of patient care, while greater experience viewing and documenting EHRs is positively associated quality of patient care. Overall, 43.7% of nurses reported that they were satisfied using their EHRs, which represents the largest sample of nurses for possible EHR satisfaction categories (see Table 4.9).  Nurses reported an average of 3.24 (1.15) out of a possible EHR satisfaction score ranging from 1-5, reporting that on average, nurses were neither satisfied nor dissatisfied with using their EHR system.  The highest quality of patient care (M= 4.32) is when nurses reported being very satisfied with using their EHRs, and the lowest quality of patient care (M= 3.07) was when nurses reported being very dissatisfied.  Greater EHR satisfaction may increase nurses’ quality of patient care. Table 4.11 displays quality of patient care and EHR satisfaction by nursing experience categories.  The highest quality of patient care was reported by nurses who have less than five years of experience and are very satisfied (M= 4.89).  The lowest quality of patient care was reported by nurses who have more than 25 years of experience are very dissatisfied (M= 2.59).  69  Higher EHR satisfaction is positively associated with quality of patient care within each nursing experience category, and nursing experience is negatively associated with quality of patient care.  Table 4.9 Nurses’ Characteristics and Quality of Patient Care Nurses’ Characteristics Categories f (%) Quality of Patient Care  M(SD) Nursing Experience All 1031(100%) - 0-4 years 137(13.3%) 3.72 (.99) 5-15 years 308(29.9%) 3.75 (.83) 16-25 years 230(22.3%) 3.65 (.90) More than 25 years 356(34.5%) 3.50 (.95) Experience Viewing EHRs All 1031(100%) - Less than 1 year 96 (9.3%) 3.18 (.96) 1- 2 years 162 (15.7%) 3.69 (.93) 3- 4 years 211 (20.5%) 3.70 (.86) 5- 6 years 131 (12.7%) 3.63 (.90) More than 6 years 431 (41.8%) 3.70 (.91) Experience Documenting EHRs All 1031(100%) - Less than 1 year 216 (21.0%) 3.21 (.92) 1- 2 years 176 (17.1%) 3.61 (.92) 3- 4 years 211 (20.5%) 3.73 (.84) 5- 6 years 124 (12.0%) 3.77 (.87) More than 6 years 304 (29.5%) 3.84 (.87) EHR Satisfaction All 1031(100%) - Very dissatisfied 86 (8.3%) 3.07 (1.02) Dissatisfied 226 (21.9%) 3.23 (.86) Neither 170 (16.5%) 3.29 (.78) Satisfied 451 (43.7%) 3.94 (.79) Very satisfied 98(9.5%) 4.32 (.71) Note. N = 1031.     70  Table 4.10 EHR Experience and Quality of Patient Care by Years of Nursing Experience Nursing Experience EHR Experience Quality of Patient Care f M(SD) Less than 5 Years Experience viewing EHRs Less than a year 36 3.22(.99) 1-2 years 45 3.80(.99) 3-4 years 46 4.00(.89) 5-6 years 5 4.20(.84) More than 6 years 5 3.60(.89) Experience documenting EHRs Less than a year 52 3.71 (.92) 1-2 years 37 4.05 (.97) 3-4 years 42 4.05 (.85) 5-6 years 1 4.00 (.00) More than 6 years 5 4.20 (.45) 5 – 15 Years Experience viewing EHRs Less than a year 23 3.35 (.83) 1-2 years 49 3.85 (.82) 3-4 years 60 3.73 (.80) 5-6 years 51 3.73 (.85) More than 6 years 126 3.80 (.83) Experience documenting EHRs Less than a year 62 3.32 (.83) 1-2 years 53 3.68 (.08) 3-4 years 57 3.81 (.74) 5-6 years 47 3.94 (.82) More than 6 years 89 3.96 (.81) 16 – 25 Years Experience viewing EHRs Less than a year 19 3.32 (.95) 1-2 years 33 3.61 (1.00) 3-4 years 42 3.62 (.79) 5-6 years 26 3.62 (.94) More than 6 years 110 3.74 (.88) Experience documenting EHRs Less than a year 37 3.30 (.94) 1-2 years 39 3.51 (.97) 3-4 years 43 3.53 (.85) 5-6 years 31 3.77 (.88) More than 6 years 80 3.89 (.81) More than 25 Years Experience viewing EHRs Less than a year 18 2.72 (1.02) 1-2 years 36 3.39 (.87) 3-4 years 63 3.49 (.88) 5-6 years 49 3.47 (.92) More than 6 years 190 3.61 (.96) Experience documenting EHRs Less than a year 65 3.08 (1.00) 1-2 years 47 3.26 (.87) 3-4 years 69 3.61 (.86) 5-6 years 45 3.58 (.89) More than 6 years 130 3.79 (.94) Note. N = 1031. 71  Table 4.11 EHR Satisfaction and Quality of Patient Care by Years of Nursing Experience  Quality of Patient Care Nursing Experience EHR Satisfaction f M(SD) Less than 5 Years Very dissatisfied 9 3.56 (1.33) Dissatisfied 28 3.11 (.96) Neither satisfied or dissatisfied 19 3.21 (.81) Satisfied 72 3.97 (.82) Very satisfied 9 4.89 (.33) 5-15 Years Very dissatisfied 32 3.47 (.88) Dissatisfied 67 3.39 (.82) Neither satisfied or dissatisfied 48 3.42 (.71) Satisfied 131 4.02 (.77) Very satisfied 30 4.20 (.61) 16-25 Years Very dissatisfied 18 2.83 (.71) Dissatisfied 50 3.22 (.86) Neither satisfied or dissatisfied 35 3.31 (.63) Satisfied 110 3.96 (.81) Very satisfied 17 4.41( .71) More than 25 Years Very dissatisfied 27 2.59 (1.01) Dissatisfied 81 3.14 (.86) Neither satisfied or dissatisfied 68 3.22 (.88) Satisfied 138 3.81 (.78) Very satisfied 42 4.24 (.79) Note. N = 1031.  4.5.1 Post-Hoc Power Analysis According to Cohen (1988), “power of a statistical test is the probability that it will yield statistically significant results” (p.1).  In multiple regression, the estimated effect size is a function of the value of R2 (Polit & Beck, 2017).  Given the total variance in the final regression model (R2 = .26), with a sample of 1031 participants, 15 predictors, a 5% chance of a type 1 error and a 20% chance of a type 2 error, the study’s effect size is large (f 2 =.33)(Cohen, 1988; Polit 72  & Beck, 2017; Stoper, 2011b).  A large effect size suggests the degree to which the null hypothesis is false is extensive (Cohen, 1988).   4.6 Summary of Findings  In this section, the key findings described above are summarized with respect to the research questions.   In direct patient care nursing, what are the relationships between EHR components and quality of care when controlling for nurses’ characteristics? The significant predictors of quality of care CDSS after controlling for nurses’ characteristics was the CDSS tool; as the CDSS tool was a positive predictor of quality of care.  Therefore, out of eleven EHR components tested, one research hypothesis related to the CDSS tool was accepted.  This means that in predicting patient quality of care, the strongest predictor is whether nurses used electronic clinical decision support to assist with their work.  What is the association between quality of patient care and nurses’ characteristics (nursing experience, experience viewing EHRs, experience documenting EHRs, EHR satisfaction)?  Nurse characteristics were also predictors of quality of care, specifically, nurses’ satisfaction with EHRs and the extent of experience documenting in EHRs. 73  Chapter 5: Discussion  The purpose of this study was to examine the unique effects of EHR components and nurses’ characteristics on quality of patient care.  In this chapter, key study findings are explained and interpreted using Rogers’ DoI model.  The first section of this chapter discusses expected and unexpected study findings.  Strengths and limitations associated with this study are addressed in the second section.  Lastly, the research, practice, education and policy implications of these findings are discussed in the third section.  5.1 EHR Components and Quality of Care One out of eleven EHR components were statistically significant as hypothesized in the final regression model: the CDSS tool.  In final the regression model, it was unexpected for the other ten EHR components to be statistically insignificant or inversely associated with quality of patient care (Table 4.8).  One possible reason for this was the extent to which nurses in the study used all of the EHR components.  For example, the frequencies of EHR component (see Table 4.6) indicates that each of the components were not used to the same extent as each other within their EHR component groupings.   Overall, the quality of care averages are higher when nurses reported using EHR components, except for CPOE lab and CPOE patient care (see Table 4.6).  The systematic review by Rouleau et al. (2017) suggests that access to specific EHR components and their functionalities may differ from one healthcare setting to another.  Similarly, several studies noted that the heterogeneity of EHRs used by nurses was a main barrier to synthesizing EHR use and nurse-sensitive outcomes, as different components and their functions affects nursing care in 74  multiple ways (Rouleau et al., 2017; Urquhart et al., 2009; Zadvinskis et al., 2013).  Additionally, due to the complexity of EHR functionalities, nurses’ perceptions of use may differ from actual use (Rouleau et al., 2017).  This may further contribute to the amount of EHR component use remaining unclear (Rouleau et al., 2017).  This may be related to the way the EHR component items were chosen and described in the primary survey, which will be discussed in the limitations section.  The next section discusses the specific components of the EHR that may be the best predictors of quality patient care.  5.1.1 Computerized Provider Order Entry and Quality of Care The four CPOE components available for this study were lab, imaging, patient care, and medications.  In the final regression model, CPOE lab inversely predicted quality of care, which was an unexpected finding.  CPOE components are more commonly used by physicians because ordering bloodwork, diagnostics, and medications are not usually within the scope of regulated nurses (with the exceptions of NPs), which may have contributed to the inverse results (CNA, 2016; Kirkendall et al., 2013).  Furthermore, despite the literature implying that CPOE can influence nursing processes, CPOE research is mostly concerned with investigating physician adoption due to their prescribing responsibilities (Nelson & Staggers, 2017; Weir et al., 2012).  Since CPOE research has focused on physicians, there may be indirect and unanticipated consequences of using electronic order entry of lab tests in nursing (Rogers, 2003).  Indirect and unanticipated consequences of an innovation are difficult to measure, but just as imperative to understand (Rogers, 2003).  Since there is limited knowledge as to how CPOE may influence nursing care, future inquiry of how CPOE can impact quality of care is warranted.  Based on the 75  current findings, quality of patient care resulting from nurses’ use of EHR does not seem to be predicted by CPOE components.  5.1.2 Nursing Information Systems and Quality of Care The NIS components of EHRs used by nurses for this study were the NIS documentation, flowsheets, and care plans.  These systems are used to document various nursing processes and provide standardization of care, and to communicate with other healthcare providers.  A key finding was that all NIS components were not positively associated with quality of patient care, and in fact the NIS care plan component was negatively associated.  The review on NISs by Urquhart et al. (2009) posits that there is a positive relationship between using a NIS and quality of care.  However, the available evidence regarding patient outcomes and nurse-sensitive quality indicators from nurses using NISs remain limited.  This finding is unsettling because if nurses perceive that a NIS decreases their quality of care, they may reject using that NIS by performing workarounds (Fowler et al., 2009; Young et al., 2010).  Workarounds can unintentionally increase nursing errors if the technologies are not used as they are designed to be used (Fowler et al., 2009; Young et al., 2010).  It is important that the consequences of nurses utilizing NIS are explored, as nurses’ perception of care quality is related to patient safety and nurse-sensitive outcomes (Urquhart et al., 2009).  The unexpected findings of NIS care plans may have also been due to the wording of the primary survey tool, which may have influenced how nurses responded to this question.  This should be explored in future research.  Moreover, examining direct relationships between NISs and specific patient outcomes and nurse-sensitive quality indicators may be beneficial to understand their impact on patients.     76  5.1.3 Computerized Decision Support and Quality of Care   The CDSS tool was statistically significant and positively associated with quality of care, which was an expected finding for this study.  In the literature, using CDSSs to improve guideline adherence such as vaccinations and diagnostic test ordering have shown positive results (Fossum et al., 2011).  However, the direct impact CDSSs have on patient care through clinician guideline adherence and alert responsiveness largely remain unclear (Fossum et al., 2011; Rood et al., 2005).  Similarly, the studies by White and Mungall (1991) and Bouyer-Ferrullo et al. (2015) indicate that there was no significant difference in patient outcomes with CDSS use.  The inconclusive research findings may have been because many CDSS studies did not describe their EHR components and contexts of use in detail, and the majority of CDSS studies focused on nursing care processes such as documentation alone (Rouleau et al., 2017).  One study suggests that quality of patient care corresponds to nurse-sensitive outcomes (Stalphers et al., 2016).  The finding that the CDSS tool increased quality of patient care in this study demonstrates that EHR component use may impact patient care.  It is imperative to further explore how CDSSs affects quality of care in the future.    5.2 Nurses’ Individual Characteristics and Quality of Care There are several characteristics of nurses that have been used to describe the relationship between EHR components and quality of patient care.  The nurses’ characteristics that were significant predictors in the regression model will be discussed, followed by other demographics that further inform research about EHR use in nursing.   77  5.2.1 Nursing Experience and Quality of Care In the final model of the regression analysis shown in Table 4.8, nursing experience was found to be statistically significant and inversely associated to quality of patient care.  This was consistent with the study by Eley et al., (2009), in which they found that nursing experience was negatively associated with EHR use and quality of care.  However, Stalphers et al. (2016) described nurses’ work experience to be positively associated with quality of care and EHR use, but implied that this was an unexpected finding in their study.  Several studies acknowledged nursing experience as an influencing factor when measuring quality of care and EHR use, but did not control for it (Kirkendall et al., 2013; Kossman, 2008; Walker-Czuz, 2016; Zadvinskis et al., 2013).  Other studies chose to measure the impact of nurses’ age on quality of patient care and EHR use instead of nursing experience (Kossman et al., 2008; Lin et al., 2016).   Previous researcher examining the relationship between EHR components and quality of patient care have explored the nuances as to whether age or amount of nursing experience are significant predictors. In this sample, the average age of nurses in years (M= 43.8) was similar to the Canadian national nursing population (M=44.4), which falls within the age range of 35 to 48 years as reported by others Kossman et al. 2008; Lin et al., 2016; Eley et al., 2008; Daly, 2002).  Two studies indicated that age had a negative correlation to EHR use and quality of care (Eley et al. 2008; Lin et al., 2016), while one study indicated that nurses’ age had no difference in EHR use and quality of patient care (Kossman et al., 2008).  Similarly, the DoI theory states that age has shown both positive and negative relationships to the adoption of innovations (Rogers, 2003).  There remains varying evidence pertaining to how nurses’ age effects quality of care and EHR use.  In two studies, the age of nurses was commonly associated with the amount of nursing experience in the literature (Andre et al, 2008; Stalpers et al., 2016), which was also found to be 78  highly correlated to nursing experience when analyzing correlations for this study.  As multicollinearity between variables may overinflate standard errors of regression coefficients, the age item was not included for inferential analysis (Tabachnik & Fidell, 2013).  Based on this, several noteworthy points are important to discuss with regards to nursing experience.   The greatest category of nursing experience for this study was over 25 years, which is greater than the nursing literature: Daly (2002), reported an average of 8 years; Kossman et al. (2008) reported an average of 9 years; Kirkendall et al. (2013) reported 10 or more years; Stalphers et al. (2016) reported 16.76 years; and the nurses in the study by Zadvinskins et al. (2013) had an average of 3-20 years of work experience.  This is a noteworthy finding, as the average quality of patient care decreased as the nursing experience categories increased (Table 4.9).  Furthermore, when analyzing nurses’ characteristics by nursing experience categories in Table 4.10, it was clear that greater nursing experience decreased quality of patient care averages in almost every category.  This study highlights the importance of capturing nurses’ work experience as it may influence nurses’ perceptions regarding EHR use. Nursing experience may have an opposite effect on quality of care when nurses use EHRs because the transition from paper-based systems to computer-based systems may impact the complex nursing processes that experienced nurses have been practicing for longer than less experienced nurses (Eley et al., 2009).  Nurses with less experience and more recent entry-to-practice education may be more open to change, thus making it easier to accept newer technologies into their practice (Rogers, 2003).  Moreover, nurses with more working experience were more likely to have received their basic nursing education without computers and may have less computer skills (Eley et al., 2009).  Possessing adequate computer skills likely affects nurses’ perception of care quality when using EHR components (Andre et al., 2008).  The 79  inverse relationship between nursing experience and quality of care when all other variables are controlled for was expected.  However, given that personal technology use has increased dramatically over the last decade, there should be careful consideration of the findings related to nursing experience and EHR use, as some of the cited publications are over ten years old.  The inconsistencies of how nursing experience affects quality of care and EHR use in the literature suggests that additional research should be performed (Andre et al., 2008; Eley et al., 2009; Stalphers et al., 2016).  However, nursing experience cannot be used as a predictor without understanding the context of the amount of EHR experience that nurses possess as well, which is discussed in the next section.  5.2.2 Experience Using EHRs and Quality of Care Experience viewing EHRs was not statistically significant, but experience documenting EHRs was significantly positively associated with quality of care as shown in Table 4.8.  In the literature, nurses’ experience using EHRs were conceptualized as one item, and several articles acknowledged the impact of previous EHR experience on nurses’ attitudes and perceptions of EHR use (Andre et al., 2008; Eley et al., 2009; Fowler et al., 2009; Kirkendall et al., 2013; Kossman et al., 2008; Lin et al., 2016; Zadvinskis et al., 2013).  It is imperative to not underestimate the effects of nurses having previous experiences with computers and EHRs because “earlier adopters have greater knowledge of innovations than do later adopters” (Rogers, 2003, p. 291).  Previous EHR experiences influenced nurses’ attitudes towards adopting EHRs because their sense of self-efficacy, confidence, and autonomy is higher than nurses without previous EHR experience (Andre et al., 2008; Eley et al., 2008).  According to Rogers (2003), 80  individual attitudes towards change can affect overall organizational innovativeness, such as EHR implementation. Although the strong association between previous EHR experience and quality of care is recognized, few studies have analyzed its effects (Andre et al., 2008; Eley et al., 2008; Lin et al., 2016).  Unlike this study, most studies considered EHR experience as one concept; as nurses’ EHR experience was found to be positively associated with quality of care in two studies (Andre et al., 2008; Eley et al., 2009), but had no significance on quality of care in another (Lin et al., 2016).  Experience documenting EHRs may have displayed significance over experience viewing EHRs because inputting information represents a higher EHR functionality compared to only viewing information (Nelson & Staggers, 2017).  It is possible that being able to communicate with other healthcare members through electronic documentation impacts nurses’ quality of care more than nurses passively viewing patient information (Nelson & Staggers, 2017).  This study explores how specific EHR functions may impact patient care quality because the primary survey split EHR experience into two separate concepts.  Given that many EHR systems and components have unique functions, further inquiry should be conducted in order to distinguish how experience viewing EHRs and experience documenting EHRs may affect nurses’ quality of care (Coiera, 2015; Rouleau et al., 2017). Overall, the nurses of this secondary analysis had more EHR experience compared to the nursing literature, as the majority of longitudinal studies occurred six months after implementation (Bouyer-Ferrullo et al., 2015; Fossum et al., 2011; Fowler, 2009; Kaye, 2017; Kossman et al., 2008; Zadvinskis, 2013).  The amount of EHR experience could change how nurses perceive the EHR components because the more experience nurses have with using an EHR, the more confident they become at using that technology (Andre et al., 2008).  Therefore, 81  it may be important to acknowledge the amount of EHR experience a nursing sample possesses when studying perceptions of EHR use. As shown in Table 4.9, nurses reported more experience viewing EHRs compared to documenting in EHRs.  This may be related to how EHR systems are implemented in direct patient care settings, which coincides with the electronic medical record adoption management (EMRAM) measurement tool (HIMSS, 2018a).  For example, EMRAM stage one includes the use of laboratory and diagnostic imaging, which is similar to the experience viewing EHRs item.  EMRAM stage three identifies the use of nursing electronic documentation which corresponds to the experience documenting EHRs item (HIMSS, 2018a).  In clinical settings, EHR systems with less sophisticated functionalities such as viewing patient information, may be implemented in the first stages of implementation, whereas higher levels of functionalities such as inputting information into an EHR systems may occur at a later date (HIMSS, 2018a).   To understand the significance of documenting EHRs over viewing EHRs a subanalysis of nursing experience and EHR experience was reported (see Table 4.10).  In general, the highest quality of care averages were from nurses with lower nursing experience but higher EHR experience.  More specifically, nurses with less work experience but greater experience documenting EHRs had higher quality of care averages.  This finding may help to elucidate why experience documenting EHRs was significant in the regression model.  To this researchers’ knowledge, this is the first study to measure these specific nurse characteristics through descriptive and inferential analyses, and further research should be conducted to investigate these noteworthy findings.  Overall, experience documenting is more meaningful to the relationship between EHR and quality of care than is experience viewing EHRs.  82  5.2.3 EHR Satisfaction and Quality of Care As shown in Table 4.8, EHR satisfaction was statistically significant and positively associated with quality of care in the final regression model and explained the largest variance of quality of care.  The finding of EHR satisfaction having a strong association to the dependent variable quality of care was an expected result.  In the literature, EHR satisfaction is associated with quality of care in several studies (Fowler et al., 2009; Lin et al., 2016; Stalphers et al., 2016; Walker-Czyz et al., 2016).  In one study, nurses’ satisfaction with using nursing information systems was largely contingent on their perceived nursing performance which embodied quality of care and patient safety (Lin et al., 2016).  In another study, nurses were generally more satisfied in hospitals with higher nurse-sensitive indicators scores, and least satisfied in lower-scoring hospitals (Stalpers et al., 2016).  The literature suggests that nurses are more likely to be satisfied with using EHR components if they perceive they will benefit their quality of care because of their responsibility for patients’ wellbeing (Fowler et al., 2009; Lin et al., 2016; Stalphers et al., 2016).  These literature findings as well as this study’s findings highlight the importance of nurses’ EHR satisfaction as an important indicator to patient care quality. On average, nurses reported that they were satisfied with their EHR use (see Table 4.9).  Furthermore, when EHR satisfaction is broken down by nursing experience, nurses with less nursing experience but higher EHR satisfaction had the highest quality of patient care (Table 4.11).  Even nurses who had more than 25 years of experience had high quality of patient care averages if they were very satisfied.  The Pearson’s correlation findings in Table 4.7 also indicates that EHR satisfaction and quality of patient care are moderately correlated (r = .42).  These descriptive and bivariate correlations suggest that higher satisfaction is strongly associated with higher quality of care.   83  Similarly, in the study by Lin et al. (2016), the nursing care performance and satisfaction of NIS usage were strongly correlated (r= .51), which they reported as an expected finding.  Using structural equation modelling, their study concluded that satisfaction with NIS explained 22.6% variance in nursing care performance (Lin et al., 2016).  Lin et al. (2016) conceptualized nursing care performance as quality of clinical care and patient safety through NIS usage, and concluded that NIS satisfaction is highly related to nurses’ perception of quality care.  Stalpers et al. (2016) conceptualized EHR satisfaction as a sub-concept of nurse assessed quality of care, and supported the notion of how interdependent the terms are with one another.  Stalpers et al. (2016) reported that a 10% increase in nurses’ satisfaction with the quality of care was associated with a 0.6 to 2.0-point increase in nurse sensitive outcome performance scores.  Fowler et al. (2009) reported higher nurse satisfaction of barcode medication administration technologies with nurses’ perceptions of increased patient safety.  Thus, user satisfaction of EHR systems may impact nurses’ perceived quality of care.   5.2.4 Nursing Designation and Domain of Practice Nursing designations and nurses’ domain of practice were reported for this study but were not used for inferential analyses.  The frequencies of the nursing designation suggest that 2.2% of direct patient care nurses reported the ‘other’ category (n=23) (see Table 4.2).  When further analyzed, it was found that those who selected the ‘other’ category reported job roles such as nurse educator and various clinical nurse specialist roles.  A nursing designation is related to the provincial nursing college that is mandated to oversee the regulation of nurses’ scope of practice and licensure, along with the shared responsibility of governments, the public, educational institutions, employers, and nurses (CNA, 2016; CNA, 2015).  For example, the 84  nursing designations in 2017 as reported by CIHI were 70.7% RNs (including NPs and CNSs), 27.9% LPNs and 1.4% RPNs (CIHI, 2017).  The conceptual ambiguity of the term nursing designation versus nursing job roles reported by the study sample in the primary survey created difficulties in recoding the ‘other’ category for secondary analyses.  In the literature, few studies measured the effects of nurse designations or job roles on EHR use and quality of patient care.  For example, Eley et al. (2008) analyzed nursing job levels and computer access; reporting that RNs with greater job status have more computer access and computer experience at work compared to RNs of lower job roles.  However, no other studies included nurse designations in their analyses. Although information on the effects of nursing designations and job roles on EHR use and quality of care are limited, four studies included nursing education levels as an individual nurses’ characteristic (Kossman et al., 2008; Stalphers, 2016; Walker_Czuz, 2016; Zadvinskis, 2013).  The findings of one study indicated that higher education increased quality of patient care (Stalphers, 2016), and the other three studies did not include education in statistical analyses.  The DoI theory suggests that those with higher education tend to adopt innovations more readily (Rogers, 2003).  Although the primary study did not include education level in the questionnaire, in the future it would be worth while investigating how nurses’ education level may affect EHR component use and quality of care.  CIHI indicates that Registered Nurses (including NPs and CNSs) practice in five domains: Clinical care, education, administration, research and policy (CIHI, 2016d).  Although the domain of practice item in the primary survey questionnaire is similar to CIHI’s nursing practice domains (with the exception of the additional clinical informatics domain), it was answered by all regulated nursing professions in the primary survey (see Table 4.3).  85  Additionally, the CNA describes four nursing areas of responsibility as direct care, administration, education and research (CNA, 2016).  Self-reporting as working within clinical care as a domain of practice, then in a separate question reporting if one worked in direct patient care setting may have been conceptually confusing to respondents of the primary study survey (see Appendix A).  Therefore, frequencies of domains of practice was not selected for inferential statistical analysis in this study.  Further explanation of the nursing designation item and the domain of practice item from the primary survey are discussed later in the limitation section of this chapter.  5.3 Strengths and Limitations 5.3.1 Strengths  This study has two main strengths.  Firstly, the sample characteristics.  The sample size is large, thereby increasing the power and reducing the chances of a type 2 error (Polit & Beck, 2017).  A sample size that successfully represents the larger population also increases the generalizability of the study (Polit & Beck, 2017).  Generalizability is further enhanced because the nurses’ demographics are similar to the CIHI national nursing census (2016c).  The primary survey includes English and French speaking participants across all provinces and territories through using national nursing bodies, thus this survey is generalizable to speakers of both languages in Canada.   Secondly, this secondary analysis was able to explore the relationships between EHR component use and quality of care while controlling for nurses’ characteristics by using hierarchical regression.  This study also measured quality of patient care, which is not often used as the dependent variable in health informatics research (Bani-issa, Al Yateem, Al Makhzoomy, 86  & Ibrahim, 2016;  Chow, Chin, Lee, Leung & Tang, 2011; Hurley et al., 2007; McBride, Tietze, Hanley & Thomas, 2017).  It is important to examine individual factors because they are critical in influencing technology adoption in healthcare, which several previous studies did not measure using inferential analyses (Andre et al., 2008; Daly et al., 2002; Eley et al., 2009; Fowler et al., 2009; Kirkendall, 2013; Kossman et al., 2008; Lin et al., 2016; Stalpers et al., 2016; Walker-Czyz, 2016; Ward, 2013; Zadvinskis, 2013).  The large sample size and inferential statistical analysis allowed this study to contribute exciting and new data to the nursing body of knowledge.    5.3.2 Limitations  Despite these strengths, the results of this study should be interpreted with caution due to several limitations.  A secondary data analysis study must work within the confines of the sample and data from the primary study (Doolan & Froelicher, 2009; Polit & Beck, 2017).  Additionally, the findings of the primary study are related to the measurement tool that was used to collect the data (Doolan & Froelicher, 2009).  The primary study’s questionnaire tool was not psychometrically validated and therefore may have significant flaws in the wording of items.  This may influence the results of this study.  Most importantly, the phrasing of the question of quality of patient care (“Because of your use of electronic record/clinical information systems, the quality of the patient care you provide is: …..”) may have influenced the participants’ responses (see Appendix A).  Ideally, the question regarding quality of patient care should have been a direct question and not a conditional question.  In addition, the primary survey tool did not provide accurate examples and descriptions for each EHR component item.  A body mass index calculator was provided as an example for the CDSS tool’s function when CDSS 87  functionalities include alerts and reminders, medication dosing, risk screening and prevention, clinical guidelines, and resource links (Nelson & Staggers, 2017).  Another important example is how the three NIS components did not reflect the nursing literature.  For instance, the flowsheet item may have been reported for chronic diseases only, when many acute diagnoses have flowsheets and standardized order sets in direct patient care settings as well (Nelson & Staggers, 2017; Urquhart et al., 2009).  Furthermore, the primary survey tool did not define quality of patient care and EHR satisfaction, so nurses might have had differing conceptualizations of those terms. Lastly, the questionnaire did not define the term use for the EHR component items.  For example, one could argue that in direct patient care settings, the EHR components such as CPOE lab is only used by physicians and NPs because they are responsible for prescribing and ordering.  However, nurses may also perceive that they use CPOE lab because they review and verify these orders before acting on them as part of their scope of practice.  Overall, there may have been some confusion when answering the EHR component items, which may explain why there were many unexpected findings associated with quality of patient care and specific EHR components. Another noted limitation to the primary survey tool was that exact description of the study sample (particularly job designation and domain of practice) was unknown due to primary survey question reporting.  In the future, it would be useful to include definitions of what a nursing designation versus what a nursing job role is to help facilitate more clarity to the respondents.  Understanding how EHR component use can affect quality of patient care for different nursing designations would benefit the nursing body of literature as many institutions consider staffing decisions based on skill mix (CNA, 2012).  Furthermore, it would have been helpful to include nurses who used paper-based documentation in reporting their quality of 88  patient care and in order to compare their quality of care scores to nurses who use EHRs.  In the future, the primary survey tool being used to measure EHR component use should be pretested by experienced researchers who understand EHRs and clinical contexts in which they are used to increase validity and reliability of the questionnaire.  5.4 Implications This section discusses the nursing implications of this study’s research findings.  Specifically, this section discusses the consequences and future recommendations by the four domains of research, education, practice and policy.   5.4.1 Nursing Research The primary survey by Infoway (2017a) demonstrated that gathering nursing information at the national level is possible.  By utilizing the CNA and CNIA for recruitment, the Infoway (2017a) survey was able to attain a very large sample size to accurately reflect the Canadian nursing population.  This secondary study took advantage of this large sample size to produce regression analyses to explore the relationships between nurses’ characteristics, EHR component use, and quality of care.  However, sampling could have been randomized to increase the validity of the primary study (Polit & Beck, 2017). The EHR components in the primary study may not reflect the most common EHR components in the clinical setting.  For example, barcoding medication administration technologies was not included as an EHR component in the 2017 Infoway survey (See Appendix A), even though they affect patient safety and quality of care (Fowler et al., 2009; Weir et al., 2012; Young et al., 2010; Zadvinskis et al., 2014).  This situation was also noted in the literature, 89  as the systematic review by Rouleau et al. (2017) described the need for EHRs to be defined and unified to better capture how their functionalities and the impact on nurses’ perceptions.  Several researchers describe the health technologies they are studying as a whole system, EHRs have different capabilities and user interfaces which affects how the end users perceive their use (Rouleau et al., 2017; Urquhart et al., 2009; Zadvinskis et al., 2013).  The heterogeneity of EHRs create difficulties in comparing the outcomes for systematic reviews and meta analyses (Rouleau et al., 2017).  This issue can be solved in part by researchers describing in detail the health technologies’ components and functionalities to increase the generalizability of their research (Rouleau et al., 2017; Urquhart et al., 2009; Zadvinskis et al., 2013). This study analyzed the relationships between nurses’ individual characteristics, EHR component use and quality of patient care.  Few studies included nurses’ characteristics such as nurses’ age, sex, nursing experience, EHR experience, and EHR satisfaction into their analyses when measuring quality of care and EHR use (Andre et al., 2008; Eley et al., 2009; Fowler et al., 2009; Kossman et al., 2008; Lin et al., 2016; Stalpers et al., 2016; Walker-Czyz, 2016).  The conflicting findings of EHR use and quality of care in the literature may be associated with interdependent organizational and individual level factors (Irurita, 1999; Rogers, 2003; Urquhart et al., 2009).  Although the findings of this study highlight how nurses’ individual factors may impact quality of care and EHR use, more research needs to be conducted to corroborate these findings.  Poor research methodologies such as small sample sizes and short research timelines may further contribute to these mixed findings (Bouyer-Ferrullo, et al., 2015; Daly, 2002; Dowding et al., 2011; Fossum et al., 2011; Hessels et al., 2015; Rouleau et al., 2017; Urquhart et al., 2009; Rantz et al., 2009; White & Mungall, 1990).  For example, the majority of longitudinal studies occurred six months to one year after EHR implementation (Bouyer-Ferrullo et al., 2015; 90  Fossum et al., 2011; Fowler, 2009; Kaye, 2017; Kossman et al., 2008; Zadvinskis, 2013).  Longitudinal research methods are recommended for innovation research, because nurse’s perceptions of using technologies will likely change over time (Andre et al., 2008; Rogers, 2003).  Including EHR experience as a variable in future studies may uncover more accurate findings related to EHR use and quality of care.  Additional research needs to be conducted to assess the impacts of individual and organizational factors on EHR use and quality of care.   EHR satisfaction is commonly used as a dependent variable in the literature, and is often utilized as an indication of nursing acceptance or adoption of technologies (Bani-issa et al., 2016; Chow et al., 2011; Hurley et al., 2007; McBride et al., 2017).  For example, in one study, nurses were 2.76 times more likely to be satisfied with the EHR when decision support functionality was present and used compared with nurses who indicated that decision support functionality was not present (OR, 2.76; 95% CI, 1.67–4.57; p< .001) (McBride et al., 2017).  Complex in nature, quality of care and EHR satisfaction should be researched together as the dependent variables because of the direct association to patient care and patient safety (Lin et al., 2016).  For example, using structured equation modeling, one study reported that nurses’ satisfaction with NIS usage had a significant influence on nursing care performance (R2 = 22.6) (Lin et al., 2016).  Quality of care is likely affected by EHR satisfaction because positive attitudes towards an innovation may affect the use of that innovation (Lin et al., 2016; Rogers, 2003).  In the future, nurse researchers could conduct structural equation modeling as a statistical analysis technique to compute how much quality of care and EHR satisfaction are related, because these variables are currently not well understood (Lin et al., 2016).  If satisfaction contributes to helping improve nursing care quality, more research should be conducted to explore this relationship.   91  Currently in the EHR literature, there remains ambiguity towards the definition of quality of patient care, as some refer to quality of care encompassing nurse-sensitive outcomes, nursing and healthcare processes, patient safety, and patient or nurses’ perception of quality of care (Campbell, Roland, & Buetow, 2000; Kutney-Lee & Kelly, 2011; Lin, Chiou, Chen, & Yang, 2016; Purdy et al., 2010; Stalpers, Linden, Kaljouw, & Schuurmans, 2016; Wong et al., 2010).  Furthermore, researchers interested in EHR use have not taken advantage of reliable tools to accurately measure quality of care (Rouleau et al., 2017).  For example, quality of care items that are internationally validated by Aiken et al. (2012), are utilized in nursing research regarding work environments.  For the purpose of this thesis, the term quality of patient care was conceptualized as the combination of nurse-sensitive patient outcomes and nurses’ subjective view of their quality of care because nurses’ assessed quality of patient care often incorporates both subjective and objective measures (Kutney-Lee & Kelly, 2011; Lin et al., 2016; Stalpers et al., 2016).  In order to further explore the linkage between EHR use and quality of patient care, the term quality of patient care needs to be further evaluated, and researchers should consider established measurements of quality of care that exhibit strong psychometric properties.  The combination of rich description of EHR components, adequate sample sizes, longitudinal research designs, including organizational and individual factors in the analyses, and using reliable tools to measure quality of care will increase the body of knowledge of how EHR use impacts nurses’ quality of patient care.  Lastly, future research can also explore the linkages between EHR use and actual patient outcomes such as patient falls and medication errors.  92  5.4.2   Nursing Education In this study, nurses with less nursing experience but higher EHR documentation experience had the highest quality of patient care (Table 4.11).  In a case study by Borycki, Frisch, Moreau, and Kushniruk (2015), nursing students reported that using an educational EHR could significantly improve their quality of care by providing structure to their nursing practice.  EHR systems may encourage novice nurses to increase their patient-care quality by providing them with best practice guidelines reminders and more efficient ways of communicating across disciplines (De Gagne, Bisanar, Makowski, & Neumann, 2012).   Integrating EHRs into the nursing curricula has the potential to provide invaluable experiential opportunities for students to become not only familiar, but proficient with documenting, retrieving, synthesizing and applying electronic patient information (Zhang, Ura & Kaplan, 2014).  Embracing EHRs into nursing student’s workflows and decision making processes in simulated learning environments affords them an opportunity to experience EHRs as integral to safe, as well as quality, patient care (Zhang et al., 2014).   Furthermore, using EHRs in nursing school may increase nursing students’ readiness to adopt these systems (Gagnon et al. 2012; Rogers, 2003).  Acceptance of, and attitudes towards EHRs may be key factors in how nurses use EHRs in practice (Gagnon et al. 2012).  Providing explicit learning opportunities to use tools such as CDSS and NIS as intended may prevent beginner nurses resorting to work-arounds (Gephart, Bristol, Dye, Finley, & Carrington, 2016).  As such, having previous experience using EHR components may improve quality of care.  Upon entering the workforce, new-graduate nurses are expected to perform tasks that require technological skills to care for patients.  To support this growing need, the Canadian Association of Schools of Nursing (CASN) received funding from Infoway to promote 93  informatics education and practice development in Canada (Canadian Association Schools of Nursing[CASN], 2015).  Specifically, CASN has been promoting awareness among nurse educators, informatics experts, and nursing students on integrating nursing informatics into entry-to-practice competencies; increasing the capacity of Canadian nurse educators to teach nursing informatics; and engaging nursing’s key stakeholders in developing nursing curricula (CASN, 2015). Nursing informatics curriculum implementation may play an important role in improving patient outcomes because it assists in bridging the knowledge gap between nursing students and employed staff nurses who are expected to be proficient in using health technologies (De Gagne et al., 2012).  Despite the potential benefits, many undergraduate nursing programs lack EHR education (Borycki, Frisch, Moreau, & Kushniruk, 2015).  Potential barriers to integrating EHR education into nursing curricula programs includes; lack of health informatics competencies of the nursing educators, insufficient funding for purchasing and sustaining EHR systems, and inadequate administrative support (Borycki et al., 2015; De Gagne et al., 2012).  Facilitators for educational EHRs are to create stronger partnerships between the undergraduate programs and the health centers that are using EHRs, and for academic agencies to provide opportunities for continuing education for faculty regarding health technologies (De Gagne et al., 2012).  As greater experience documenting in EHR may improve quality of care, nursing students should have the opportunity to document in EHRs before they enter practice.   5.4.3 Nursing Practice  Direct patient care nurses are the largest portion of healthcare workers in Canada (CIHI, 2016a).  Health organizations should invest in creating spaces for direct patient care nurses to be 94  satisfied with their EHR system, as EHR satisfaction can impact quality of care (Fowler et al., 2009; Lin et al., 2016; Stalphers et al., 2016; Walker-Czyz et al., 2016).  In clinical settings, greater EHR satisfaction can be sustained in several ways; which includes increasing nurse staffing levels during EHR implementation (McGinn et al., 2011; Nguyen et al., 2014); offering twenty-four hour technical support (McGinn et al., 2011; Nguyen, Bellucci & Nguyen, 2014); adequate staff training (Gagnon et al. 2012; McGinn et al., 2011; Nguyen et al., 2014; Poe, Abbott & Pronovost, 2011); creating peer mentor positions (Gagnon et al. 2012; Nguyen et al., 2014; Poe et al., 2011); including direct patient care nurses in the design and implementation of the EHR (Bani-issa et al., 2016; Gagnon et al. 2012; McGinn et al., 2011; Nguyen et al., 2014; Smith, Morris & Janke, 2011); and conducting pretesting with end-users before go-live (Gagnon et al. 2012; Lin et al., 2016; Nguyen et al., 2014).  A detailed analysis of what influences EHR satisfaction is challenging due to the complexity of these interrelated factors (Ammenwerth et al., 2003; Gagnon et al. 2012).  Although this is beyond the scope of this thesis, a deeper exploration of such factors may impact the consequences of EHR use in clinical settings. Healthcare administrators responsible for choosing and purchasing health technologies should select EHRs that utilize standardized nursing languages.  Integrating standardized nursing terminologies into EHRs can benefit direct patient care nurses by streamlining electronic nursing documentation into specific concepts (CNA, 2017; Daly et al., 2002).  For example, the CNA advocates for the use of the systematized nomenclature of medicine-clinical terms (SNOMED CT) and the international classification for nursing practice (ICNP) as the standardized clinical terminologies in Canada for EHR documentation (CNA, 2017).  Utilizing standardized nursing languages in EHR components provides a frame of reference for nurses when documenting their work processes, and also provides a unified nursing voice (Daly et al., 2002; Lin et al., 2016).  95  Furthermore, standardized nursing terminologies create opportunities for nursing data to be analyzed for nurse-sensitive quality indicators and patient outcome research (Lin et al., 2016).   Health organizations and nurse administrators should prioritize measuring the consequences of EHR use post-implementation (Rogers, 2003; Stalpers et al., 2016).  This can be done through purchasing EHRs that have the capabilities to capture nurse-sensitive quality indicators such as medication administration errors; falls; hospital acquired infections; and pressure ulcers (Dowding et al., 2012; Furukawa et al., 2010; Pillemer et al. 2012; Rantz et al., 2009; Stalpers et al., 2016; Walker-Czyz, 2016).  Nurse-sensitive indicators provide a quantitative basis to monitor and evaluate nursing care and are defined as patient outcomes that are directly associated with nurses’ practices (Stalpers et al., 2016; Walker-Czyz, 2016).  In the future, organizations should choose EHRs that use standardized clinical terminologies and data analytics to capture nurse-sensitive quality indicators.  If these recommendations are done, then nurses’ quality of care can be evaluated and potentially improved (Lin et al., 2016; Stalpers et al., 2016).      5.4.4 Nursing and Healthcare Policy National and provincial health policies should support the planning, implementation, and sustainability of EHRs in order to address the needs of healthcare workers and patients.  For nurses in particular, policies should support the design and use of EHR components such as NIS and CDSS in a point-of-care work to enhance quality of care.   In this study, experience documenting in EHRs increased quality of care, while viewing EHRs did not.  Health information and management system society (HIMSS) standards indicate that healthcare centers become electronic medical record adoption model (EMRAM) level 3 once 96  they obtain 50% nursing and allied health electronic documentation (HIMSS, 2018a).  However, I propose that the EMRAM level three should be change to when clinical settings reach 100% nursing and allied health electronic documentation.  With increased funding from Infoway and other national organizations, nurse informaticians can help encourage increased electronic documentation and decision support through research and knowledge dissemination opportunities. Health authorities should advocate for EHR systems to include CDSS components as it may increase quality of care.  Hospitals should be supported by provincial and national funding and policies in providing frontline staff adequate time and resources to learn and use EHRs. This will increase their experience using EHR components, decrease workarounds, all leading to better care and safer patient environments.  National organizations like the CNA and CNIA can together to help improve nurses’ perceptions of EHRs by continuing to advocate for health organizations to include nurses in the planning and implementation of EHR systems so that EHR components such as NIS documentation can better reflect what nurses need to provide better care.  Lastly, the 2017 Infoway survey on which this secondary study was based, should continue to be deployed in order to capture invaluable data and trends related to nurses’ health technology use.  The next survey should include psychometrically validated measures, expert nurse opinions and continue to obtain a large pan-Canadian sample size to achieve maximum potential for the findings.  This can only be accomplished if Infoway designates sufficient resources towards this project in the future.  Improving this Canadian nursing survey tool should be Infoway’s priorities as it sheds light on how EHR use impacts patient care.  97  5.5 Conclusion In complex healthcare systems, quality of patient care is likely dependent upon multiple, interconnected concepts.  In this study, individual characteristics and specific EHR component use influenced nurses’ perceived quality of care.  Clinical decision support use, EHR satisfaction and EHR documentation experience increased quality of care.  Greater nursing experience, electronic nursing documentation and CPOE lab use decreased quality of patient care.  According to the literature, CDSS and NIS have the potential to increase quality of patient care, but the evidence is limited.  Furthermore, there is inadequate evidence to the effects of nursing experience and EHR experience on quality of care.  Nurse researchers should include rich descriptions of the EHRs’ functionalities, use longitudinal research designs, and validated measurement tools to increase the understanding of how EHR use impacts quality of care.  In practice, direct patient care nurses should have adequate EHR training and be involved in the planning and implementation of EHRs to increase adoption rates; which may decrease work-arounds and improve patient safety.  This secondary analysis would not have been possible without the primary study survey tool conducted by Infoway.  Although the survey had a large sample size, the questionnaire could be improved with psychometric validation in the future.  It is imperative for Infoway to continue deploying this national nursing survey on health technology use, as nurses are the largest end-users of EHRs and play a direct role in ensuring patient safety.  98  References Abacus Dataverse Network, UVic Dataverse, Canada Health Infoway Data Collection Dataverse (2017). 2017 National survey of Canadian nurses: Use of digital health technology in practice. Retrieved Dec 1, 2017 from: http://dvn.library.ubc.ca/dvn/dv/CHI/faces/study/ StudyPage.xhtml?globalId=hdl:11272/0499.  Adler-Milstein, J., Everson, J., & Lee, S. D. (2015). EHR adoption and hospital performance: Time-related effects. Health Services Research, 50(6), 1751-1771. doi:10.1111/1475 6773.12406 Aiken, L. H., Sermeus, W., Van den Heede, K., Sloane, D. M., Busse, R., McKee, M., . . . Kutney-Lee, A. (2012). Patient safety, satisfaction, and quality of hospital care: Cross- sectional surveys of nurses and patients in 12 countries in Europe and the United States. The BMJ, 344(7851), 1-14. doi:10.1136/bmj.e1717 Ancker, J. S., Kern, L. M., Edwards, A., Nosal, S., Stein, D. M., Hauser, D., . . . with the HITEC Investigators. (2015). Associations between healthcare quality and use of electronic health record functions in ambulatory care. Journal of the American Medical Informatics Association, 22(4), 864-871. doi:10.1093/jamia/ocv030 André, B., Inger Ringdal, G., Loge, J. H., Rannestad, T., Laerum, H., & Kaasa, S. (2008). Experiences with the implementation of computerized tools in health care units: A review article. International Journal of Human-Computer Interaction, 24(8), 753-775. doi:10.1080/10447310802205768 Auditor General of Canada. (2010). Electronic health records in Canada: An overview of federal and provincial audit reports. Retrieved April 3, 2018 from: http://www.oag bvg.gc.ca/internet/English/ parl_oag_201004_07_e_33720.html 99  Bani-issa, W., Al Yateem, N., Al Makhzoomy, I. K., & Ibrahim, A. (2016). Satisfaction of health-care providers with electronic health records and perceived barriers to its implementation in the United Arab Emirates: Electronic health records. International Journal of Nursing Practice, 22(4), 408-416. doi:10.1111/ijn.12450 Bennett. C., Khangura. S., Brehaut, J. C., Graham, I. D., Moher, D., Potter B. K., & Grimshaw, G. M. (2011) Reporting guidelines for survey research: An analysis of published guidance and reporting practices. PLoS Medicine, 8(8), 1-11. doi:10.1371/journal.pmed.1001069 Borycki, E. M., Frisch, N., Moreau, J., & Kushniruk, A. W. (2015). Integration of electronic health records into nursing education: Issues, challenges and limitations. Studies in Health Technology and Informatics, 208, 88-92. doi:10.3233/978-1-61499-488-6-88 Bouyer-Ferullo, S., Androwich, I. M., & Dykes P. C. (2015). Clinical decision support and perioperative peripheral nerve injury: A quality improvement project. CIN: Computers, Informatics, Nursing, 33 (6), 238-248. doi: 10.1097/CIN.000000000000014 Bright, T. J., Wong, A., Dhurjati, R., Bristow, E., Bastian, L., Coeytaux, R. R., . . . Lobach, D. (2012). Effect of clinical decision-support systems: A systematic review. Annals of Internal Medicine, 157(1), 29-43.  Campbell, S. M., Roland, M. O., & Buetow, S. A. (2000). Defining quality of care. Social Science & Medicine, 51(11), 1611-1625. doi:10.1016/S0277-9536(00)00057-5  Canada Health Infoway. (2017). National survey of Canadian nurses: Use of digital health technology in practice. Retrieved December 10, 2017 from: https://www.infoway inforoute.ca/en/component/edocman/resources/reports/benefits-evaluation/3345-2017 -national-survey-of-canadian-nurses-data-set?Itemid=101 100  Canada Health Infoway. (2017). National survey of Canadian nurses: Use of digital health technology in practice final executive report. Retrieved December 10, 2017 from: https://www.infoway inforoute.ca/en/component/edocman/3320-2017-national-survey of-canadian-nurses-use of-digital-health-technology-in-practice/view document?Itemid=0  Canada Health Infoway. (2018). Computerized Provider Order Entry. Retrieved April 3, 2018 from: https://www.infoway-inforoute.ca/en/solutions/clinicians-e-services/computerized provider-order-entry  Canada Health Infoway. (2018). Decision Support and Workflow. Retrieved April 3, 2018 from: https://www.infoway-inforoute.ca/en/solutions/clinicians-e-services/decisionsupport-and workflow Canada Health Infoway. (2018). Electronic Health Records. Retrieved March 5, 2018 from: https://www.infoway-inforoute.ca/en/solutions/electronic-health-records Canada Health Infoway. (2018). Summary Corporate Plan 2018-2019. Retrieved April 3, 2018 from: https://www.infoway-inforoute.ca/en/plan  Canada’s Health Informatics Association (COACH). (2017). Strategic Plan 2017-2020. Retrieved April 3, 2018 from: https://ams.coachorg.com/inventory/ PurchaseDetails.aspx?Id=d7e20f49 57af-4612-9a3587f868988650 Canadian Association of Schools of Nursing. (2015). Nursing Informatics: Entry-to-Practice Competencies for Registered Nurses. Retrieved November 22, 2017 from: https://www.casn.ca/2014/12/casn-entry-practice-nursing-informatics-competencies/ Canadian Institute for Health Information. (2016). Regulated Nurses 2015. Retrieved April 3, 2018 from: https://secure.cihi.ca/free_products/Nursing_Report_2015_en.pdf  101  Canadian Institute for Health Information. (2016). Regulated Nurses, 2016: Methodology Guide. Retrieved June 20, 2018 from: https://www.cihi.ca/en/regulated-nurses-2016  Canadian Institute for Health Information. (2016). Regulated Nurses, 2016: Registered Nurse/Nurse Practitioner Data Tables. Retrieved June 20, 2018 from: https://www.cihi.ca/en/regulated nurses-2016-registered-nursenurse-practitionerdata tables  Canadian Institute for Health Information. (2016). Registered Nurses/Nurse Practitioners. Retrieved June 20, 2018 from: https://www.cihi.ca/en/registered-nursesnurse practitioners  Canadian Institute for Health Information. (2017). Regulated Nurses: Data tables. Retrieved June 20, 2018 from: https://www.cihi.ca/en/regulated-nurses-2017-data-tables  Canadian Nurses Association. (2006). E-Nursing strategy for Canada. Retrieved April 3, 2018 from https://www.cna-aiic.ca/~/media/cna/page-content/pdf-en/e-nursing-strategy-for canada.pdf?la=en  Canadian Nurses Association (2012). Staff Mix Decision-making Framework for Quality Nursing Care. Retrieved August 10, 2018 from: https://www.cna-aiic.ca/en/nursing practice/evidence based-practice/staffing-patient-outcomes/staff-mix-framework Canadian Nurses Association. (2015). Framework for the practice of Registered Nurses in Canada. Retrieved June 20, 2018 from: https://www.cna-aiic.ca/~/media/cna/page/ content/pdf/en/framework-for-the-pracice-of-registered-nurses-in-canada.pdf  Canadian Nurses Association. (2016). Registered Nurses Profile (including Nurse Practitioners), Canada, 2016. Retrieved June 20, 2018 from: https://www.cna-aiic.ca/en/nursing- practice/thepracticeof-nursing/health-human-resources/nursing-statistics/Canada  102  Canadian Nurses Association. (2017). Joint position statement Nursing informatics. Retrieved March 5, 2018 from: https://www.cna-aiic.ca/en/~/media/cna/page-content/pdf- fr/nursing-informatics joint-position-statement Chow, S. K., Chin, W., Lee, H., Leung, H., & Tang, F. (2012). Nurses’ perceptions and attitudes towards computerisation in a private hospital. Journal of Clinical Nursing, 21(11-12), 1685-1696. doi:10.1111/j.1365-2702.2011.03905.x Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, N.J: L. Erlbaum Associates. College of Registered Nurses of British Columbia. (2017). Nursing Documentation. Retrieved April 3, 2018 from: https://www.crnbc.ca/Standards/ PracticeStandards/Pages/documentation.aspx  Coiera, E. (2015). Guide to Health Informatics (3rd Ed.). Boca Raton, Florida: CRC Press. Daly, J. M., Buckwater, K., & Maas, M. (2002). Written and computerized care plans: Organizational Processes and effects on patient outcomes. Journal of Gerontological Nursing, 28(9), 14-23.   De Gagne, J. C., Bisanar, W. A., Makowski, J. T., & Neumann, J. L. (2012). Integrating informatics into the BSN curriculum: A review of the literature. Nurse Education Today, 32(6), 675-682. doi:10.1016/j.nedt.2011.09.003 De Sousa, P. A. F., Dal Sasso, G. T. M., & Barra, D. C. C. (2012). Contributions of the electronic health records to the safety of intensive care unit patients: An integrative review. Text & Context Nursing Journal, 21(4), 971-979. 103  De Winter, J. C. F., Gosling, S. D., & Potter, J. (2016). Comparing the pearson and spearman correlation coefficients across distributions and sample sizes: A tutorial using simulations and empirical data. Psychological Methods, 21(3), 273–290. doi: 10.1037/met0000079 Doolan, D. M., & Froelicher, E. S. (2009). Using an existing data set to answer new research questions: A methodological review. Research and Theory for Nursing Practice, 23(3), 203-215. Dowding, D. W., Turley, M., & Garrido, T. (2011). The impact of an electronic health record on nurse sensitive patient outcomes: An interrupted time series analysis. Journal of American Medical Informatics Association, 19, 615-620. doi:10.1136/amiajnl-2011 000504 Doyle, G. J., Garrett, B., & Currie, L. M. (2014). Integrating mobile devices into nursing curricula: Opportunities for implementation using Rogers' diffusion of innovation model. Nurse Education Today, 34(5), 775. doi:10.1016/j.nedt.2013.10.021 Eley, R., Fallon, T., Soar, J., Buikstra, E., & Hegney, D. (2009). Barriers to use of information and computer technology by Australia's nurses: A national survey. Journal of Clinical Nursing, 18(8), 1151. doi:10.1111/j.1365-2702.2008.02336.x Fossum, M., Alexander G. L., Ehnforsa, M., & Ehrenbergd A. (2011). Effects of a computerized decision support system on pressure ulcers and malnutrition in nursing homes for the elderly. International Journal of Medical Informatics 80, 607–617. doi:10.1016/j.ijmedinf.2011.06.009 Fowler, S. B., Sohler, P., & Zarillo, D. F. (2009). Barcode technology for medication administration: Medication errors and nurse satisfaction. Medsurg Nursing, 18(2), 103- 109.  104  Furukawa, M. F., Raghu, T. S., & Shao, B. B. M. (2010). Electronic medical records, nurse staffing, and nurse-sensitive patient outcomes: Evidence from California hospitals, 1998 2007. Health Services Research 45(4), 241-262. doi: 10.1111/j.1475-6773.2010.01110.x Gagnon, M., Desmartis, M., Labrecque, M., Car, J., Pagliari, C., Pluye, P., . . . Légaré, F. (2012). Systematic review of factors influencing the adoption of information and communication technologies by healthcare professionals. Journal of Medical Systems, 36(1), 241-277. doi:10.1007/s10916-010-9473-4 Gephart, S. M., Bristol, A. A., Dye, J. L., Finley, B. A., & Carrington, J. M. (2016). Validity and reliability of a new measure of nursing experience with unintended consequences of electronic health records. Computers, Informatics, Nursing, 34(10), 436-447. doi:10.1097/CIN.0000000000000285 Health Information Management Systems Society (2018). Electronic medical record adoption model (EMRAM). Retrieved April 3, 2018 from: https://www.himssanalytics.org/emram  Health Information Management Systems Society. (2018). Electronic Health Records. Retrieved April 5, 2018 from: http://www.himss.org/library/ehr.  Health Information Management Systems Society. (2018). Electronic Medication Administration Record (eMAR). Retrieved April 3, 2018 from: http://www.himss.org/electronic medicationadministration-record-emar.  Health Informatics Society of Australia. (2017). Position Statement: Nursing Informatics. Retrieved April 3, 2018 from: https://www.hisa.org.au/nia/  Hurley, A. C., Bane, A., Fotakis, S., Duffy, M. E., Sevigny, A., Poon, E. G., & Gandhi, T. K. (2007). Nurses' satisfaction with medication administration point-of-care technology.105   The Journal of Nursing Administration, 37(7/8), 343-349. doi:10.1097/01.NNA. 0000285114.60689.02  IBM SPSS Statistics (Version 25.0) [IBM Corp]. Armonk, NY: IBM Corp International Medical Informatics Association Towards (2007). IMIA 2015 - the IMIA Strategic Plan. Retrieved April 3, 2018 from: http://imia-medinfo.org/wp/imia-strategic-plan/  Institute of Medicine (2012). Health IT and Patient Safety: Building Safer Systems for Better Care. Washington, D.C: The National Academy Press.  Kaye, S. P. (2016). Nurses’ attitudes toward meaningful use technologies: An integrative review. Computers, Informatics, Nursing, (35)5, 237-247. doi:10.1097/CIN.0000000000000310 Kazley, A. S., & Ozcan, Y. A. (2008). Do hospitals with electronic medical records (EMRs) provide higher quality care? An examination of three clinical conditions. Medical Care Research and Review, 65(4), 496-513. doi:10.1177/1077558707313437 Kirkendall, E. S., Goldenhar, L. M., Simon, J. L., Wheeler, D. S., & Spooner, S. A. (2013). Transitioning from a computerized provider order entry and paper documentation system to an electronic health record: Expectations and experiences of hospital staff. International Journal of Medical Informatics, 82, 1037–1045. doi: 10.1016/j.ijmedinf.2013.08.005 Kossman, S. P. & Scheidenhelm, S. L. (2008). Nurses’ perceptions of the impact of electronic health records on work and patient outcomes. CIN: Computers, Informatics, Nursing, 26(2), 69–77.  Kruse, C. S. & Beane, A. (2018). Health Information Technology Continues to Show Positive Effect on Medical Outcomes: Systematic Review. Journal of Medical Internet Research, 20(2), 1-8. doi:10.2196/jmir.8793 106  Kurihara, Y., Kusunose, T., Okabayashi, Y., Nyu, K., Fujikawa, K., Miyai, C., & Okuhara, Y. (2001). Full implementation of a computerized nursing records system at kochi medical school hospital in japan. Computers in Nursing, 19(3), 122. Lin, H., Chiou, J., Chen, C., & Yang, C. (2016). Understanding the impact of nurses' perception and technological capability on nurses' satisfaction with nursing information system usage: A holistic perspective of alignment. Computers in Human Behavior, 57, 143-152. doi:10.1016/j.chb.2015.12.001 Linnen, D. (2016). The promise of big data: Improving patient safety and nursing practice. Nursing, 46(5), 28-34. doi: 10.1097/01.NURSE.0000482256.71143.09 Lee, T., Lee, T., Lin, K., & Chang, P. (2005). Factors affecting the use of nursing information systems in Taiwan. Journal of Advanced Nursing, 50(2), 170-178. doi:10.1111/j.1365 2648.2005.03376.x Lee, E. J., Lee, M., & Moorhead, S. (2009). Developing an electronic nursing record system for clinical care and nursing effectiveness research in a Korean home healthcare setting. Computers, Informatics, Nursing, 27(4), 234-244. doi:10.1097/NCN.0b013e3181a91b75 McBride, S., Tietze, M., Hanley, M. A., & Thomas, L. (2017). Statewide study to assess nursesʼ experiences with meaningful use based electronic health records. Computers, Informatics, Nursing, 35(1), 18-28. doi:10.1097/CIN.0000000000000290 McGinn, C. A., Grenier, S., Duplantie, J., Shaw, N., Sicotte, C., Mathieu, L., . . . Gagnon, M. (2011). Comparison of user groups' perspectives of barriers and facilitators to implementing electronic health records: A systematic review. BMC Medicine, 9(1), 46 46. doi:10.1186/1741-7015-9-46 107  Moreland, P. J., Gallagher, S., Bena, J. F., Morrison, S., & Albert, N. M. (2012). Nursing satisfaction with implementation of electronic medication administration record. Computers, Informatics, Nursing, 30(2), 97-103. doi:10.1097/NCN.0b013e318224b54e Nelson, R. & Staggers, N. (2017). Health Informatics: An interprofessional approach (2nd Ed.). St. Louis, Missouri: Elsevier.  Nguyen, L., Bellucci, E., & Nguyen, L. T. (2014). Electronic health records implementation: An evaluation of information system impact and contingency factors. International Journal of Medical Informatics, 83(11), 779-796. doi:10.1016/j.ijmedinf.2014.06.011 Pillemer, K., Meador, R. H., Teresi, J. A., Chen, E. K., Henderson Jr., C. R., Lachs, M. S., . . . Eimicke, J. P. (2012). Effects of electronic health information technology implementation on nursing home resident outcomes. Journal of Aging and Health, 24(1), 92–112. doi: 10.1177/0898264311408899 Poe, S. S., Abbott, P., & Pronovost, P. (2011). Building nursing intellectual capital for safe use of information technology: A before-after study to test an evidence-based peer coach intervention. Journal of Nursing Care Quality, 26(2), 110-119. doi:10.1097/NCQ.0b013e31820b221d Polit, D. F. (2010). Statistics and data analysis for nursing research (2nd ed.). New Jersey: Pearson Education Inc. Polit, D. F., & Beck, C. T. (2017). Nursing research: Generating and assessing evidence for nursing practice (10th ed.). Philadelphia: Wolters Kluwer Health/Lippincott Williams & Wilkins. 108  Polit, D. F., & Beck, C. T. (2014). Essentials of nursing research: Appraising evidence for nursing practice (8th ed.). Philadelphia: Wolters Kluwer Health/Lippincott Williams & Wilkins. Powers, P. (2009). A growing e-health IT partnership: Canadian healthcare and HIMSS analytics. Healthcare Quarterly, (12)2, 122-125.   Purdy, N., Spence Laschinger, H. K., Finegan, J., Kerr, M., & Olivera, F. (2010). Effects of work environments on nurse and patient outcomes. Journal of Nursing Management, 18(8), 901-913. doi:10.1111/j.1365-2834.2010.01172.x Rantz, M. J.,  Hicks, L., Petroski, G. F., Madsen, R. W., Alexander, G., Galambos, C., . . . Greenwald, L. (2010). Cost, staffing and quality impact of bedside electronic medical record (EMR) in nursing homes. Journal of the American Medical Directors Association, 11(7), 485-493. doi:10.1016/j.jamda.2009.11.010 Rogers, E. M. (2003). Diffusion of innovations (5th Ed.). NY, New York: Free Press. Rood, E., Bosman, R. J., Spoel, J. I. V., Taylor, P., & Zandstra, D. F. (2005). Use of a computerized guideline for glucose regulation in the intensive care unit improved both guideline adherence and glucose regulation. Journal of American Medical Informatics Association 12, 172–180. doi: 10.1197/jamia.M1598. Rouleau, G., Gagnon, M., Côté, J., Payne-Gagnon, J., Hudson, E., & Dubois, C. (2017). Impact of information and communication technologies on nursing care: Results of an overview of systematic reviews. Journal of Medical Internet Research, 19(4), 1-33. doi: 10.2196/jmir.6686 109  Smith, D., Morris, A., & Janke, J. (2011). Nursing satisfaction and attitudes with computerized software implementation: A quasi-experimental study. Computers, Informatics, Nursing, 29(4), 245-250. doi:10.1097/NCN.0b013e3181f9dcfa Soper, D. (2011). A-priori online statistical calculator for multiple regression. Retrieved July 25, 2018 from www.danielsoper.com  Soper, D. (2011). Effect Size Calculator for Multiple Regression. Retrieved July 25, 2018 from www.danielsoper.com  Stalpers, D., Kieft, Renate A. M. M, Linden, D., Kaljouw, M. J., & Schuurmans, M. (2016). Concordance between nurse-reported quality of care and quality of care as publicly reported by nurse-sensitive indicators. BMC Health Services Research, 16(1), 120. doi:10.1186/s12913-016-1372-z Stokowski, L. A. (2001). Using technology to improve medication safety in the newborn intensive care unit. Advances in Neonatal Care, 1(2), 70-83. doi: 10.1053/adnc.2001.29591 Tabachnik, B. G., & Fidell, L. S. (2013). Using multivariate statistics (6th Ed.). California, USA: Pearson Education.  Urquhart, C., Currell, R., Grant, M. J., & Hardiker, N. R. (2009). Nursing record systems:  Effects on nursing practice and healthcare outcomes. Cochrane Database of Systematic Reviews, 1, 1-66. doi: 10.1002/14651858.CD002099.pub2. Walker-Czyz, A. (2016). The impact of an integrated electronic health record adoption on  nursing care quality. Journal of Nursing Administration, 46(7-8), 366-372. doi: 10.1097/NNA.0000000000000360 110  Walters, S. (1986). Computerized care plan help nurses achieve quality patient care. Journal of Nursing Administration, 16(11), 33-39.  Ward, R. (2013). The application of technology acceptance and diffusion of innovation models in healthcare informatics: A literature review. Health Policy and Technology (2), 222 228. doi: 10.1016/j.hlpt.2013.07.002 Weir, C. R., Staggers, N., & Laukert, T. (2012). Reviewing the impact of computerized provider order entry on clinical outcomes: The quality of systematic reviews. International Journal of Medical Informatics, 81(4), 219-231. doi:10.1016/j.ijmedinf.2012.01.009 White, R. H. & Mungall, D. (1991). Outpatient Management of warfarin therapy: Comparison of computer predicted dosage adjustment to skilled professional care. Therapeutic Drug Monitoring, 13(1), 46-50.  Wong, C. A., Spence Laschinger, H. K., & Cummings, G. G. (2010). Authentic leadership and nurses' voice behaviour and perceptions of care quality. Journal of Nursing Management, 18(8), 889-900. doi:10.1111/j.1365-2834.2010.01113.x Yanamadala, S., Morrison, D., Curtin, C., McDonald, K., & Hernandez- Boussard, T. (2016). Electronic health records and quality of care: An observational study modeling impact on mortality, readmissions, and complications. Medicine, 95(19), 1-6. DOI: 10.1097/MD.0000000000003332 Young, J., Slebodnik, M., & Sands, L. (2010). Bar code technology and medication administration error. Journal of Patient Safety, 6(2), 115-120.   Zadvinskis, I. M., Chipps, E., & Yen, P. (2013;2014;). Exploring nurses’ confirmed expectations regarding health IT: A phenomenological study. International Journal of Medical Informatics, 83(2), 89-98. doi:10.1016/j.ijmedinf.2013.11.001 111  Zhang, W., Ura, D., & Kaplan, B. (2014). A comparison study on integrating electronic health records into priority simulation in undergraduate nursing education. Journal of Nursing Education and Practice, 4(7) doi:10.5430/jnep.v4n7p123 112  Appendices Appendix A   Primary Study Questionnaire Tool  113   114   115   116   117   118   119   120   121   122   123   124   125   126   127   128   129   130    131  Appendix B  Full Model of Multiple Regression Findings for Quality of Patient Care     Predictors B SE B b CI (95%) R2 R2 Δ      Lower Upper    Model 1       .02**  .02**     CPOE_Lab -.21 .08 -.11** -.37 -.05       CPOE_Dx .07 .08 .04 -.08 .22       CPOE_PC -.06 .07 -.03 -.19 .07       CPOE_Med .21 .07 .12*** .08 .34   Model 2      .02** .01     CPOE_Lab -.21 .08 -.11* -.36 -.05       CPOE_Dx .08 .08 .04 -.08 .23       CPOE_PC -.08 .07 -.04 -.22 .06       CPOE_Med .19 .07 .10** .05 .32       NIS_Doc .16 .07 .08* .02 .30       NIS_Flow .01 .07 .00 -.13 .14      NIS_Plan -.02 .07 -.01 -.15 .11   Model 3      .03*** .01**     CPOE_Lab -.22 .08 -.12** -.38 -.06       CPOE_Dx .07 .08 .04 -.08 .23       CPOE_PC -.11 .07 -.06 -.25 .03       CPOE_Med .13 .07 .07 -.02 .27       NIS_Doc .12 .07 .06 -.02 .27       NIS_Flow -.06 .07 -.03 -.20 .09       NIS_Plan -.04 .07 -.02 -.18 .09       CDSS_Rem .12 .07 .07 -.01 .26       CDSS_DS .18 .07 .10** .05 .31       CDSS_DI .01 .07 .01 -.13 .15   Model 4       .03*** .00     CPOE_Lab -.23 .08 -.12** -.38 -.07       CPOE_Dx .07 .08 .04 -.08 .22       CPOE_PC -.10 .07 -.06 -.24 .04       CPOE_Med .11 .08 .06 -.04 .26       NIS_Doc .11 .07 .05 -.03 .26       NIS_Flow -.06 .07 -.03 -.20 .09       NIS_Plan -.05 .07 -.03 -.18 .09       CDSS_Rem .12 .07 .06 -.02 .25       CDSS_DS .17 .07 .09* .04 .30               132      Predictors B SE B b CI (95%) R2 R2 Δ      Lower Upper       CDSS_DI .01 .07 .00 -.13 .15     eMAR .07 .07 .03 -.07 .20   Model 5      .05*** .02***     CPOE_Lab -.22 .08 -.12** -.38 -.06       CPOE_Dx .05 .08 .03 -.10 .20       CPOE_PC -.12 .07 -.07 -.26 .02       CPOE_Med .12 .07 .07 -.03 .27       NIS_Doc .13 .07 .06 -.01 .28       NIS_Flow -.05 .07 -.03 -.19 .09       NIS_Plan -.10 .07 -.05 -.23 .04       CDSS_Rem .14 .07 .07* .00 .27       CDSS_DS .18 .07 .1** .05 .31       CDSS_DI .01 .07 .00 -.13 .14       eMAR .06 .07 .03 -.07 .20       NsgExp -.11 .03 -.13*** -.16 -.06   Model 6      .11*** .07***     CPOE_Lab -.21 .08 -.11** -.36 -.05       CPOE_Dx .03 .08 .02 -.12 .18       CPOE_PC -.10 .07 -.06 -.24 .03       CPOE_Med .12 .07 .07 -.02 .26       NIS_Doc .04 .07 .02 -.11 .18       NIS_Flow -.09 .07 -.04 -.22 .05       NIS_Plan -.11 .07 -.06 -.24 .02       CDSS_Rem .13 .07 .07 -.00 .26       CDSS_DS .18 .06 .10** .06 .31       CDSS_DI .02 .07 .01 -.12 .15       eMAR .08 .07 .04 -.05 .21       NsgExp -.15 .03 -.18*** -.21 -.10       ExpView -.03 .03 -.04 -.08 .03       ExpDoc .18 .03 .29*** .13 .23   Model 7      .26*** .15***     CPOE_Lab -.20 .07 -.10** -.34 -.06       CPOE_Dx .04 .07 .02 -.10 .17       CPOE_PC -.03 .06 -.02 -.15 .10       CPOE_Med .07 .07 .04 -.06 .20       NIS_Doc -.06 .07 -.03 -.19 .08       NIS_Flow -.11 .06 -.06 -.24 .01               133      Predictors B SE B b CI (95%) R2 R2 Δ      Lower Upper            NIS_Plan  -.13  .06  -.07*  -.25  -.01     CDSS_Rem .07 .06 .04 -.05 .19       CDSS_DS .16 .06 .09** .04 .27       CDSS_DI -.05 .06 -.03 -.17 .08       eMAR -.00 .06 -.00 -.12 .12       NsgExp -.15 .03 -.18*** -.20 -.10       ExpView -.00 .03 -.08 -.05 .05       ExpDoc .15 .02 .24*** .10 .19       UserSat .33 .02 .41*** .28 .37   Note. N= 1031. b= standardized beta coefficient. CI(95%)= 95% confidence intervals. CPOE_Lab (0= does not use, 1= use), CPOE_Dx (0= does not use, 1= use), CPOE_PC (0= does not use, 1= use), CPOE_Med (0= does not use, 1= use), NIS_Doc (0= does not use, 1= use), NIS_Flow (0= does not use, 1= use), NIS_Plan (0= does not use, 1= use), CDSS_Rem (0= does not use, 1= use),  CDSS_DS (0= does not use, 1= use), CDSS_DI (0= does not use, 1= use), eMAR (0= does not use, 1= use). Model 7: F(15, 1015)= 24.03, p<.001.   *p<.05, **p<.01, ***p<.001  

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}]}"
                            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.24.1-0372891/manifest

Comment

Related Items