Thesis

Electronic medical records and quality of care Exploration of the effects of EMR use on the quality of hospital care Rube van Poelgeest and quality of care

ELECTRONIC MEDICAL RECORDS AND QUALITY OF CARE EXPLORATION OF THE EFFECTS OF EMR USE ON THE QUALITY OF HOSPITAL CARE RUBE VAN POELGEEST

Printed by Ipskamp Printing | proefschriften.net Layout and design: Birgit Vredenburg, persoonlijkproefschrift.nl Copyright 2022 © Rube van Poelgeest The Netherlands. All rights reserved. No parts of this thesis may be reproduced, stored in a retrieval systemor transmitted in any formor by any means without permission of the author.

proefschrift compleet versie 20221011 10/26/2022 page 1 Electronic medical records and quality of care exploration of the effects of EMR use on the quality of hospital care De relatie tussen het Elektronisch Patiënten Dossier en de kwaliteit van de geleverde Nederlandse ziekenhuiszorg (met een samenvatting in het Nederlands) Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof.dr. H.R.B.M. Kummeling, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op 6 december 2022 te 2.15 uur door Rube van Poelgeest geboren op 30 maart 1947 te Amsterdam

proefschrift compleet versie 20221011 10/26/2022 page 2 Promotor: Prof. dr. C.B. Roes Copromotor: Dr. A.J.P. Schrijvers Beoordelingscommissie: Prof. dr. C. Veenhof (voorzitter) Prof. dr. C.J. Kalkman Prof. dr. W.W. van Solinge Prof. dr. C. Wagner Dr. W.J.G. Ros Dit proefschrift werd (mede) mogelijk gemaakt met financiële steun van het CZ Fonds

In memory of my father Christiaan Cornelis van Poelgeest, who, more than anyone I have known, demonstrated to me the importance of learning. Paranimfen: Dr. E.M. (Erik) van Poelgeest, M.D. PhD, internist and postdoctoral researcher, Charité Universitätsmedizin Berlin en Amsterdam UMC, Amsterdam. Drs. H.C. (Hans) Peucker, huisarts te Houten

CONTENTS Chapter 1 Introduction 9 Chapter 2 Profile of the digitization of patient medical records in Dutch hospitals. 19 Chapter 3 The Association between eHealth Capabilities and the Quality and Safety of Health Care in The Netherlands: Comparison of HIMSS Analytics EMRAM data with Elsevier’s ‘The Best Hospitals’ data. 41 Chapter 4 Patient safety outcomes and their association with the level of digitization in Dutch hospitals 53 Chapter 5 Level of digitization in Dutch hospitals and the lengths of stay of patients with colorectal cancer. 69 Chapter 6 Medical Specialists’ Perspectives on the Influence of Electronic Medical Record Use on the Quality of Hospital Care: Semi structured Interview Study. 85 Chapter 7 General discussion 109 Appendices Bibliography 117 Summary 119 Samenvatting 120 About the author 121 Acknowledgements 122

CHAPTER 1 Introduction

10 Chapter 1 Implementations of potentially transformative eHealth technologies are currently underway internationally, often with significant impact on national expenditure. 1 2 Such large-scale efforts and expenditures have been justified on the grounds that picture archiving and communication systems (PACS), electronic prescribing (ePrescribing) and associated computerized provider (or physician) order entry systems (CPOE), and computerized decision support systems (CDSS) are supposed to help to address the problems of variable quality and safety in modern health care. 3 However, the scientific basis of such claims, which are repeatedly made, remains to be firmly established. 2,4–7 This thesis has the objective to contribute to the scientific discourse on the relationship between the digitalization of hospital care and quality and safety of such care by exploring the experience in one European country with fairly advanced EMR capabilities: The Netherlands. The hypothesis to be tested is: advanced electronic medical record (EMR) capabilities are positively associated with quality and safety of hospital care. While electronic medical records (EMRs) and electronic health records (EHRs) are used interchangeably, there is a difference between the two terms. EMRs are the digital version of the paper charts in the clinician’s office. An EMR contains the medical and treatment history of the patients in one practice. However, EHRs focus on the total health of the patient-going beyond standard clinical data collected in the provider’s office and inclusive of a broader view on a patient’s care. This thesis focuses on the EMR i.e. the legal record created in medical centers and ambulatory environments. To measure the capabilities of EMRs in Dutch hospitals a specially developed maturity model, the socalled Electronic Medical Record Adoption Model, the EMRAM, is used. The Electronic Medical Record Systems The Electronic Medical Record (System) is an application environment composed of the clinical data repository, clinical decision support, controlled medical vocabulary, order entry, computerized provider order entry, pharmacy, and clinical documentation applications.8 This environment supports the patients electronic medical record across inpatient and outpatient environments, and is used by healthcare practitioners to document, monitor, andmanage health care delivery within a care delivery organization (CDO). The data in the EMR is the record of what happened to the patient during their encounter at the CDO and is owned by the CDO. The EMR environment is a complex and sophisticated environment. Its foundation is the clinical data repository (CDR), a realtime transaction processing database of patient clinical information for practitioners.

11 Introduction In the following paragraphs EMR stands for the Electronic Medical Record System as defined in this paragraph. Maturity models of healthcare information systems and technologies. The concept of the ‘maturity’ of information systems has been developed to determine the adoption and use of information systems. Richard Nolan is considered the principal architect of the Information Systems Technology (IST) maturity approach. 9 IST maturity models fit in various types of organizations. There are several examples of maturity models focused on different areas of the organization and IST. 10 This study will use the so-called Electronic Medical Record Adoption Model (EMRAM) of HIMSS (Healthcare Information and Management Systems Society). This maturity model (Figure 1) is specifically developed to measure the maturity of information systems in hospitals.11 Figure 1 - The EMR adoption model of HIMSS for hospitals. It is important to mention that this EMRAM model explores which software the hospital has installed and whether the hospital actually uses the functionality offered by this software. HIMSS calls this ‘adoption and use’. The model has the advantage of being an accepted model for hospitals in the USA, Canada, and Europe. More than 15.000 1

12 Chapter 1 hospitals, with all their key organizational figures, are measured, whichmakes it possible to compare hospitals with each other. At the lowest level, the hospital is at the beginning and has not even automated the ancillary departments like the pharmacy, radiology, or laboratory. At the highest level, the hospital no longer uses physical paper or images, but has everything stored and encoded electronically. Clinical Decision Support Systems (CDSS) are used to analyze clinical information and thus improve the quality of care and patient safety. Clinical data are exchanged electronically between all parts of the hospital. Computerized Provider Order Entry (CPOE) systems are used and refer to the process of providers entering and sending treatment instructions, including medication, laboratory, and radiology orders, via a computer application rather than paper, fax, or telephone. Physician orders are standardized across the organization and maybe individualized for each doctor or specialty by using order sets. Orders are communicated to all departments and involved caregivers, improving response time and avoiding scheduling problems and conflicts with existing orders. And the hospital uses analytic tools and thus also has, from a management perspective, a view of all parts of the hospital. Only hospitals on the highest level (stage 7) are ready to communicate fully digitally with other healthcare providers, the patient, insurers, and other stakeholders. The scoring process is done by identifying the software used in the different functional areas of the hospital. Depending on the level of maturity, each hospital is presented with approximately 150 questions to focus on varied issues to include demographics, software functionalities, processes, integration standards, and usage in percentage by physicians and nurses. In order to monitor the quality of the scoring process, site visits or telephone interviews are conducted on selected hospitals. Validation is done by the quality assurance department of HIMSS Analytics Europe and the scoring by a proprietary scoring algorithm (HIMSS Analytics North America). If a hospital receives an EMRAM Stage 6 score, an additional 59 questions are asked by a validation team of international peer inspectors mostly from Stage 6 or 7 hospitals in the EU (see appendix 2 for example of results). Stage 6 hospitals can apply for a Stage 7 validation, consisting of a 2-day visit of peer inspectors (see appendix 3 for example of results). EMRAM scores lower than Stage 6 are not publicly shared by HIMSS Analytics and are confidential. See chapter 2 for more details of the model. Quality of healthcare The World Health Organization (WHO) definition of quality of care is “the extent to which health care services provided to individuals and patient populations improve desired health

13 Introduction outcomes. In order to achieve this, health care must be safe, effective, timely, efficient, equitable, and people-centered” 12. Central to the definition is the provision of health care by the health care provider to the individual patient or client. Quality monitoring 13 is becoming an accepted method for insurers, patients, and providers to evaluate the value of health care expenditures. Significant advances in the science of quality measurement have occurred over the past decade. Still, many challenges remain to be addressed so that quality monitoring may realize its potential as a counterforce to the demands of cost containment. In the next paragraph the quality and safety measures as used in this thesis are explained in more detail. Objective and outline of the thesis. This thesis has the objective to contribute to the scientific discourse on ‘the relationship between the digitization of hospital information and processes and the effect on the quality of care’. This aim is translated into the central research question: What is the relationship between the maturity of hospital information systems and the quality of care? This question is divided into three sub-questions: 1. How mature are the information systems of the Dutch hospitals and what are the influencing factors (determinants) for this degree of maturity? 2. What is the association between the degree of maturity of information systems of hospitals and the quality of care? 3. Which positive or negative aspects influence the relationship between EMR use and the quality of medical care according to medical specialists? The purpose of the first subquestion is to identify organizational and environmental factors that are associated with the adoption and use of hospital EMRs as determined using the EMRAM model as described (see chapter 2). The second subquestion (chapter 3, 4, 5) concerns the effects of the degree of maturity and its consequences for the quality of care. Relevant is to investigate the link between the EMRAM-score and the quality of care. The maturity concept assumes that this relationship exists in a positive sense: a higher score also translates to a higher level of 1

14 Chapter 1 the quality of care.14 The question is whether this association actually exists and to what extent especially in Dutch hospitals. To investigate this relationship, at first (chapter 3) we compared the quality and safety measures as used in Elsevier’s annual publication ‘The Best Hospitals’. Elsevier used select data from the publicly available basic quality and the safety set of the Health Care Inspectorate (IGJ) and the Dutch Health Care Transparency Program ‘Zichtbare Zorg’ (ZIZO) program. The next step (chapter 4) of this part of the study was to compare the adverse events and unplanned readmissions as measured by the Netherlands Institute for Health Services Research (NIVEL). NIVEL conducted three patient safety measurements with patient records from 2004, 2008 and 2011/2012, to keep track of changes in patient safety at a national level. As the EMRAM study took place in 2012-2014 we used the last study in 2011/2012 of this study of seventeen hospitals shared in the NIVEL study as well in the underlying EMRAM studies. For the last part (chapter 5) of the second sub question to examine the relationship between the EMRAM score and the quality of care we used the postoperative length of stay (LOS) as measured by the Dutch surgical colorectal audit (DSCA). DSCA is a nationwide audit used to monitor, evaluate and improve quality of care of primary colorectal cancer surgery. It provides feedback to all hospitals in the Netherlands on a set of quality measures and indicators. The purpose of the third sub question (chapter 6) is to examine the relationship between EMR use and the influencing factors according to the medical specialist. As there are indications that the role of the medical specialist is a particularly important factor in the ‘adoption and use’ of EMR systems in hospitals 15, we conducted a study of the role of the medical specialist toward the adoption of electronic medical records. Physicians are a main frontline user-group of EMR systems. Hence, it requires physicians to actively support and use EMR systems to benefit optimally from their use. To optimise EMR use, it is essential to understand what physicians perceive to be key factors that either support or hinder the use of EMR systems to positively impact the delivered medical treatment and care. The aim of this study was to examine how, and by which aspects, the relationship between EMR use and quality of care in hospitals is influenced according to medical specialists.

15 Introduction The last chapter (chapter 7) encompasses the conclusions and implications of the findings in the previous chapters. It discusses the main findings, it summarizes the answer to the central research question, the scientific and policy relevance of the thesis and the strengths and weaknesses of the study. The perspective of the results for the digitization of hospitals in the Netherlands and beyond are also discussed apart from recommendations for practice. Finally, angles for further research are explored. In Figure 2 the conceptual relation between the different chapters of this thesis is schematized. Centrally displayed in this diagram is the measured EMRAM score (chapter 2) in relation to, respectively the influencing factors of EMR maturity in hospitals (also chapter 2), the quality measures as used by Elsevier best hospitals publication (chapter 3), the adverse events and unplanned readmissions of the NIVEL study (chapter 4), the LOS of the DSCA audit (chapter 5) and finally the results of the study of the role of the medical specialist (chapter 6). Chapter 7 encompasses the conclusions and implications of the findings in the previous chapters. organizational and environmental characteristics Chapter 2 the medical specialist Chapter 6 adverse events Chapter 4 length of stay Chapter 5 digital maturity of hospitals Chapter 2 detailed quality & safety indicators Chapter 3 Chapter 7 Figure 2 - Outline of the thesis: chapters and associations. 1

16 Chapter 1 BIBLIOGRAPHY 1. Hillestad R, Bigelow J, Bower A, et al. Can electronic medical record systems transformhealth care? Potential health benefits, savings, and costs. Health Aff. 2005;24(5):1103-1117. doi:10.1377/ hlthaff.24.5.1103 2. BuntinMB, BurkeMF, HoaglinMC, Blumenthal D. The benefits of health information technology: a review of the recent literature shows predominantly positive results. Health Aff (Millwood). 2011;30(3):464-471. doi:10.1377/hlthaff.2011.0178 3. Adhikari NKJ, Beyene J, SamJ, et al. Effects of Computerized Clinical Decision Support Systems on Practitioner Performance. J Am Medican Assoc. 2005;293(10):1223-1238. doi:10.1001/ jama.293.10.1223 4. Jansen T, Koppes L. Elektronische gegevensuitwisseling in de zorg: ervaringen en opvattingen van zorgverleners en zorggebruikers. 2015. http://www.nivel.nl/sites/default/files/bestanden/ Rapport-Elektronische-gegevensuitwisseling-in-de-zorg.pdf. Accessed June 10, 2015. 5. Cresswell KM, Mozaffar H, Lee L, Williams R, Sheikh A. Safety risks associated with the lack of integration and interfacing of hospital health information technologies: a qualitative study of hospital electronic prescribing systems in England. BMJ Qual Saf. 2016;(April):1-12. doi:10.1136/ bmjqs-2015-004925 6. Iroju O, Soriyan A, Gambo I, Olaleke J. Interoperability in Healthcare: Benefits, Challenges and Resolutions. Int J Innov Appl Stud. 2013;3(1):262-270. http://www.ijias.issr-journals.org/ abstract.php?article=IJIAS-13-090-01. 7. Wade M, Hulland J. the Resource-Based View and Information Systems Research: Review, Extension, and Suggestions for Future Research. MIS Q. 2004;28(1):107-142. doi:Article 8. Garets D, Davis M. Electronic Medical Records vs. Electronic Health Records: Yes, There Is a Difference A HIMSS Analytics TM White Paper. 2006. www.himssanalytics.org. Accessed October 20, 2021. 9. Nolan RL. Managing the crises in data processing. Harv Bus Rev. 1979. 10. Carvalho JV, Rocha Á, Abreu A. Maturity Models of Healthcare Information Systems and Technologies: a Literature Review. J Med Syst. 2016. doi:10.1007/s10916-016-0486-5 11. Pettit L. Understanding EMRAM and how it can be used by policy-makers, hospital CIOs and their IT teams. … Off J Int Hosp Fed. 2012. http://europepmc.org/abstract/med/24377140. Accessed June 1, 2015. 12. World Health Organization. quality health services definition. https://www.who.int/newsroom/fact-sheets/detail/quality-health-services. 13. Donabedian A. Evaluating the quality of medical care. Milbank Q. 2005;83(4):691-729. doi:10.1111/j.1468-0009.2005.00397.x 14. Lin YK, Lin M, Chen H. Do electronic health records affect quality of care? Evidence from the HITECH act. Inf Syst Res. 2019;30(1):306-318. doi:10.1287/isre.2018.0813 15. Lakbala P, Dindarloo K. Physicians’ perception and attitude toward electronic medical record. Springerplus. 2014;3(1):63. http://www.pubmedcentral.nih.gov/articlerender. fcgi?artid=3918096&tool=pmcentrez&rendertype=abstract.

17 Introduction 1

CHAPTER 2 Profile of the digitization of patient medical records in Dutch hospitals. Rube van Poelgeest Lorren Pettit Rob J. de Leeuw and Guus Schrijvers Published in J Healthc Inf Manag. 2015;(Fall 2015):38-46.

20 Chapter 2 ABSTRACT Background Much research has been conducted on the organizational and environmental factors associated with the adoption and use of electronic medical records (EMRs) in hospitals. With much of these studies focused on U.S. hospitals, there are limited studies at this time surrounding the adoption of EMRs in Dutch hospitals. The purpose of this study is to profile the organizational and environmental factors associated with the adoption and use of EMR technologies in Dutch hospitals. Methods Using the HIMSS Analytics Electronic Medical Record Adoption Model to define a hospital’s EMR capabilities, acute care hospitals in the Netherlands (NL) were surveyed regarding their EMR capabilities. From this data, we determined the proportion of hospitals that had a comprehensive EMR system in use in various clinical areas of the hospital and then examined the relationship between the hospital’s EMR capabilities and various intervening variables to include environmental factors, hospital characteristics and information and communications technology (ICT) characteristics. Results The results of this study indicate that Dutch hospitals reflect a varied array of EMR capabilities. Of the 72 hospitals surveyed between 2012 and 2014 (77.4% of all NL hospitals), 15.3% had a comprehensive EMR system present in at least one clinical unit. The findings also revealed notable EMR capability differences by organizational and environmental characteristics. Larger hospitals and academic affiliated hospitals were more likely to have advanced EMR systems. There also appears to be a positive association between EMR capabilities and the size of a hospital’s IT budget. Conclusions The findings of this research project support studies from the U.S. that hospital organizational and environmental factors are associated with the adoption and use of EMR technologies. The findings generally supported half of the hypotheses forwarded in the study design. There was no support for example for the hypothesis that EMR capabilities in the Netherlands are positively associated with hospital competition and population density. This latter finding suggests the need for subsequent research studies surrounding a ‘leadership and culture’ hypothesis.

21 Profile of the digitization of patient medical records in Dutch hospitals Keywords: Hospital, EMR 2

22 Chapter 2 INTRODUCTION The EMR has emerged as a very significant component of the health information technology landscape. 1 EMRs are “systems that integrate electronically originated and maintained patient-level clinical information, derived from multiple sources, into one point of access and replaces the paper medical record as the primary source of patient information.” 2 EMRs are expected to drastically change healthcare by making care more efficient while also improving quality through the automation of care and the more complete documentation and dissemination of individual medical records. 3 However, the implementation and use of EMRs in acute care hospitals has been slow. 4 The barriers to EMR adoption are varied and include cost, concerns regarding information security, and physician resistance. 5 Although there has been discussion of widespread EMR use for several years, a national interoperable EMR system in The Netherlands (NL) has yet to emerge. A national interoperable EMR system would allow patients, payers, and providers to document and widely share health information for individuals quickly using computers, but would require a standardized format, confidentiality regulations, nearly unanimous support, and a large financial investment. 6 While this type of a system does not yet exist, several NL healthcare providers have already implemented EMRs and reported their experiences. While the overall adoption of EMRs has been slow, it has not been completely stagnant. Purpose and research question Given that IT is assumed to be fundamental to an organization’s survival and growth, they face the critical challenge of integrating, building, and reconfiguring IT resources so as to obtain competitive advantage and superior performance. Recently, a number of researchers 7–10 have applied resource based view and resource dependence theory to investigating IT business value, with mixed results. One major research stream explores the relationships between IT and environmental issues. Another major stream explores the relationships between IT and other organizational factors (i.e., organizational strategy, organizational process, organizational culture, organizational structure). A third major stream explores the relationships between IT and methods of organizing IT resources to align themwith enterprises. However, at present, we know very little about these relationships in hospitals. The purpose of this study is to identify organizational and environmental factors that are associated with the adoption and use of hospital EMRs. The results of this study may guide policy and practice by identifying specific barriers to hospital EMR use.

23 Profile of the digitization of patient medical records in Dutch hospitals Theoretical model For this study we used the Resource dependence theory model. This theory begins with the premise that organizations are not in control of all of the resources they need to survive. As such, many of the organization’s strategies for survival include attempts to reduce their dependence on external resources in times of uncertainty by securing necessary inputs. Moreover, Iroju O., et al 6 claim that the omnipresence of information and communications technology (ICT) makes information about quality and prices more readily available, generally lowering dependence among buyers and suppliers able to develop alternatives more readily. This may disturb the power balance. We elected to apply the Resource Dependence Theory to the adoption of hospital EMRs because this theory allows us to develop a fairly comprehensive model 11 (see Figure 1) to identify significant predictors and barriers to EMR use. Hypotheses to test According to the Resource Dependence theory, environmental uncertainty may motivate organizational action or strategy. 12 Organizations in areas of greater uncertainty aremore likely to take action to secure resources than organizations in areas of less uncertainty. After all, organizations with certain access to necessary resources do not need to secure inputs from the environment, while organizations in uncertain environments must adapt to their surroundings in order to survive. Since EMRs may lead to better hospital performance and outcomes as well as increasing efficiency, some hospitals may use EMRs as a strategy to combat this environmental uncertainty. From this model we deducted the following hypotheses. H1: Hospitals in a lower population density area are less likely to have advanced EMR capabilities. As hospitals are scarcer in areas of lower population areas, (potential) patients have less choice and hospitals have less urgency to adopt advanced technologies like EMRs As hospitals are scarcer in areas of lower population areas, (potential) patients have less choice and hospitals have less urgency to adopt advanced technologies like EMRs. H2: As environmental competition increases, the likelihood of having advanced EMR capabilities increases. The level of competition in an external environment, according to Resource Dependency theory, is a large predictor of organizational strategy and action. In an area with a great deal of competition, hospitals must compete for the same resources, thus making inputs 2

24 Chapter 2 potentially scarcer and placing hospitals under more pressure to distinguish themselves from competitors, thus securing their market share of patients. If patients have more choices, they may elect where to go for healthcare and will likely choose a hospital that offers new or better services such as EMRs. Hospitals may reason that EMRs will make them more appealing to the patient population in an area of high competition where patients have choices of where to receive care. In the Netherlands the bargaining of hospital services is done exclusively by the healthcare insurance companies. In some areas certain healthcare insurance companies have a prevalent position. H3: Larger hospitals are more likely to have advanced EMR capabilities. Perhaps the greatest barrier to hospital EMR adoption is the cost of implementation and maintenance. With adequate financial resources, hospitals are likely more able to purchase the often-expensive EMR systems and equipment. However, not all hospitals have the financial means to implement and use complete EMR systems. Those with smaller operating margins are less likely to have the funds to buy and implement EMRs. Organizational power is often associated with organizational size since larger organizations tend to have greater impact on a community than smaller organizations. More powerful organizations may also be those that control vital resources in an environment, and for this reason, these organizations may be in a better position to name the terms of exchange. The power associated with size allows hospitals to more easily achieve economies of scale for services, and larger purchases will likely lead to more negotiation power with suppliers. H4: Academic affiliated hospitals are more likely to have advanced EMR capabilities. University and top teaching hospitals provide a great deal of specialized care andmedical research, as well as provide the training and education of many of the nation’s healthcare workforce. According to Retchin and Wenzel, 13 university health centers, as top teaching hospitals, can easily adapt to the use of EMRs because they “have the expertise to resolve remaining software issues, the components necessary for the integrated delivery, a culture for innovation in clinical practice, and a generation of future providers that can be acclimated to the requisites for computerized records” (p.493). Another reason for this increased likelihood is that medical training occurs in these hospitals, and younger medical trainees tend to be more comfortable with computers as they have recently used them in school. Because of this, the staff resistance to EMR use may not be as great as in other hospitals.

25 Profile of the digitization of patient medical records in Dutch hospitals FEATURE: PROFILE OF THE DIGITIZATION H2: As environmental competition increases, the likelihood of having advanced EMR capabilities increases. The level of competition in an external environment, according to Resource Dependency theory, is a large predictor of organizational strategy and action. In an area with a great deal of competition, hospitals must compete for the same resources, thus making inputs potentially scarcer and placing hospitals under more pressure to distinguish themselves from competitors, thus securing their market share of patients. If patients have more choices, they may elect where to go for healthcare andwill likely choose a hospital that offers new or better services such as EMRs. Hospitals may reason that EMRs will make themmore appealing to the patient population in an area of high competitionwhere patients have choices of where to receive care. In the Netherlands the bargaining of hospital services is done exclusively by the healthcare insurance companies. In some areas certain healthcare insurance companies have a prevalent position. H3: Larger hospitals are more likely to have advanced EMR capabilities. Perhaps the greatest barrier to hospital EMR adoption is the cost of implementation and maintenance. With adequate financial resources, hospitals are likely more able to purchase the often-expensive EMR systems and equipment. However, not all hospitals have the financial means to implement anduse completeEMRsystems. Those with smaller operating margins are less likely to have the funds to buy and implement EMRs. Organizational power is often associated with organizational FIGURE 1: Comparable model (also used by Kazley, A.S.; Ozcan, Y.A., Organizational and environmental determinants of hospital EMR adoption: a national study)11 ENVIRONMENTAL FACTORS n■ Population density n■ Competition ICT FACTORS n■ ICT budget n■ Number of ICT employees ORGANIZATIONAL FACTORS n■ Financial resources n■ Hospital size n■ Teaching status n■ Leadership and culture? EMR ADOPTION AND USE? Figure 1: Comparabl m del (also used by Kazley, A.S.; Ozcan, Y.A., Organizational and environ ental determinants of hospital EMR adoption: a nat onal study 11) H5: Hospi als with rel tively higher ICT budgets t nd to hav more advanced EMR capabilities. EMR implementation requires several tens of millions of euro’s budget over the years to bring result to success. ICT budgets typically fall in the range from 2% to 10% of a hospital’s total expenditure. It is to be expected that the higher the structural budget the better the EMR adoption. H6: Hospitals with relatively more ICT employees tend to have more advanced EMR capabilities. The successful implementation of EMR systems needs a broad range of expertise to include information architects, project managers, infrastructure specialists, and maintenance specialists. We expect that having more and better resources lead to better EMR use. 2

26 Chapter 2 METHODS For themeasurement of the level of implementation of information systems, the concept of maturity of information systems has been developed. There is a large number of methods or models available to measure the level of implementation of IT. 14 This study will use the EMRAM scoring approach developed by HIMSS Analytics. 15 EMRAM is an eight-stage maturation model reflecting the EMR capabilities in hospitals, ranging from a completely paper-based environment (Stage 0) to a highly advanced digital patient record environment (Stage 7). The EMRAMmodel 16 is perhaps one of themost commonly cited EMR maturation models in the world, as its scoring approach has been applied to over 10,000 hospitals in the U.S., Canada, Europe, the Middle and Far-East, and Australia. To adjudicate a hospital’s EMRmaturation, the CEOs of every hospital in the Netherlands (93) were invited to participate in this study. Seventy-two hospitals (77.4%) joined the study. The scoring process was done by identifying the software used in the different functional areas of the hospital. Depending on the level of maturity, each hospital CEO was presented with approximately 150 questions (available from the authors) to focus on varied issues to include demographics, software functionalities, processes, integration standards, and usage in percentage by physicians and nurses. In order to monitor the quality of the scoring process, the authors conducted site visits on selected hospitals. Validation was done by the quality assurance department of HIMSS Analytics Europe and the scoring by a proprietary scoring algorithm (HIMSS Analytics North America). If a hospital received an EMRAM Stage 6 score, an additional 59 questions were asked by a validation team of international peer inspectors mostly from Stage 6 or 7 hospitals in the EU. Stage 6 hospitals can apply for a Stage 7 validation, consisting of a 2-day visit of peer inspectors. EMRAM scores lower than Stage 6 are not publicly shared by HIMSS Analytics. The Netherlands was awarded its first EMRAM Stage 7 hospital in February 2015. This outcome is outside the scope of this project (2012-2014). RESULTS Non-Response analysis A profile of the 21 Dutch hospitals that did not participate in the study is reflected in Table 1.

27 Profile of the digitization of patient medical records in Dutch hospitals Table 1: Non-Response Analysis N total non response N nresp response N resp Number of beds large (>=562) 31 16.1% 5 83.9% 26 medium (>340; <562) 31 9.7% 4 87.1% 27 small (<=340) 31 38.7% 12 61.3% 19 Type of hospital university 8 25.0% 2 75.0% 6 general 55 29.1% 16 70.9% 39 teaching hospital 30 10.0% 3 90.0% 27 Region east 16 31.3% 5 68.8% 11 north 19 36.8% 7 63.2% 12 south 21 4.8% 1 95.2% 20 west 37 21.6% 8 78.4% 29 Small hospitals were less likely to participate. Remarkably is the high participation in the south. Representing approximately 22% of all Dutch acute-care hospitals, perhaps the most remarkable characteristic differentiating survey participates was hospital size. Over 57% of non-participants were small hospital providers as defined by the number of beds associated with the hospital (see Table 1). Hospital size as an influencer on survey participation is not unique to this survey effort and may be the result of a multiplicity of factors (e.g., limited staff availability to complete the survey). Regional variances also seemed to play a factor in survey participation as hospital non-participation rates ranged from 4.8% in the southern part of the Netherlands, to 36.8% in northern NL. The remarkably high non-participation rate in the northern part of the country maybe due to the relatively high concentration of smaller hospitals in this region compared to other regions. EMRAM scores of Dutch hospital Table 2 describes the different EMRAM stages and profiles of the EMRAM distribution of NL hospitals. 2

28 Chapter 2 TABLE 2: Frequency Distribution of EMRAM Scores (Q2 2015) HIMSS level Frequency Percent Characteristics (EU model) Stage 7 0 0 Complete EMR; CCD transactions to share data; Data warehousing feeding outcomes reports, quality assurance, and business intelligence. 6 11 15.3 Physician documentation interaction with full CDSS (structured templates related to clinical protocols trigger variance & compliance alerts) and Closed loop medication administration. 5 31 43.1 Full complement of PACS displaces all filmbased images. 4 2 2.8 CPOE in at least one clinical service area and/or for medication (i.e., e-Prescribing); may have Clinical Decision Support based on clinical protocols. 3 0 0 Nursing/clinical documentation (flow sheets); may have Clinical Decision Support for error checking during order entry and/or PACS available outside Radiology 2 27 37.5 Clinical Data Repository (CDR) / Electronic Patient Record; may have Controlled Medical Vocabulary, Clinical Decision Support (CDSS) for rudimentary conflict checking, 1 1 1.4 Major ancillary clinical systems are installed (pharmacy, laboratory, radiology) or laboratory, pharmacy, radiology information system data output is delivered to the hospital for online access and processing if the ancillary service is not provided inhouse, but by external service providers 0 0 0 All Three Ancillaries (Laboratory, Radiology and Pharmacy) not Installed OR not processing Laboratory, Radiology and Pharmacy data output online from external service providers. Total 72 100 Stage 3, Stage 6, and Stage 7 requirements present as challenges for NL hospitals. The distribution profile is significant in that it suggests some EMRAM requirements are more challenging for NL hospitals than others. EMRAM Stage 3 presents as the first notable challenge to NL hospitals, as 37.5% of the hospitals in this study have yet to satisfy the requirements of this stage. Once EMRAM Stage 3 requirements are met, the EMRAM profile then suggests NL hospitals are likely to be challenged in meeting the requirements of EMRAM Stage 6 (Closed loop medication administration [CLMA] and

29 Profile of the digitization of patient medical records in Dutch hospitals decision support [CDSS]). Progressing past EMRAM Stage 5, CLMA must be live in all inpatient units (EMRAM Stage 7 requirement) or at least in one clinical unit (EMRAM Stage 6 requirement). Only 11 hospitals in the NL had surpassed EMRAM Stage 5 requirements during the time of this study. Note that in March 2015, one NL hospital successfully met all of the requirements to satisfy EMRAM Stage 7 (UMC Radboud in Nijmegen), becoming the third hospital in all of Europe to achieve this distinction (Hospital de Dénia Marina Salud in Denia, Spain, University Medical Center Hamburg-Eppendorf, Germany, and RadboudUMC in Nijmegen, The Netherlands). RESULTS OF TESTS REGARDING THE HYPOTHESES University hospitals in the Netherlands are a special group. They have different financial models, have special teaching and research missions, and are in general a lot larger than other hospitals in the NL. They combine ICT efforts for healthcare, education, and research, and have more advanced ICT organizations; although, some are struggling with the more advanced position. Teaching hospitals and general hospitals are more focused on patient throughput than university hospitals; this has to do with the different missions and financial model. For this reason, the university hospitals were excluded, with the exception of hypothesis H4, where those hospitals are compared with the other hospitals. H1: Hospitals in a lower population density area are less likely to have advanced EMR capabilities. The ‘Randstad’ area of NL (defined by the area around Amsterdam, The Hague, Rotterdam, and Utrecht, with 40% of the total population of the Netherlands) has a much higher concentration of individuals (population density =3267/km2) than the other regions of the country (average population density of the non-Randstad region = 1186/ km2) (see Table 3). In comparing the EMR profiles of the densely populated Randstad region hospitals to hospitals in the other lower densely populated regions, we find no support for our first hypothesis. On the contrary, the lowest scoring hospitals (52.2% versus 30.2%) are in areas of higher population density. 2

30 Chapter 2 TABLE 3: Population Density by EMRAM Stage Stage 0-2 Stage 3-5 Stage 6 and 7 Total Populationdensitykm2 populationdensitykm2 populationdensitykm2 populationdensitykm2 Mean Count Row N % Mean Count Row N % Mean Count Row N % Mean Count Row N % non-Randstad 951 13 30.2% 1215 25 58.1% 1649 5 11.6% 1186 43 100% Randstad 3534 12 52.2% 3021 8 34.8% 2859 3 13.0% 3287 23 100% The Randstad region has the highest population density, but the lowest EMRAM score.

31 Profile of the digitization of patient medical records in Dutch hospitals H2: As environmental competition increases, the likelihood of having advanced EMR capabilities increases. The ‘Randstad’ area of NL (defined by the area around Amsterdam, The Hague, Rotterdam and Utrecht with 40% of total population of the Netherlands) has the highest concentration of hospitals in all of NL. With the density of hospitals per 10km in the Randstad area equaling 4.5, compared to a hospital density of 1.3 in the non-Randstad area, 17 the findings of this study provide no support for the hypothesis that EMR adoption increases based on environmental competition (Table 4). On the contrary, the nonRandstad area generally presents as having a more advanced EMR profile as the largest grouping of hospitals in the non-Randstad area (58.1%) are in Stage 3-5, while in the Randstad area 52.2% are in Stages 1 and 2. TABLE 4: Competition Density by EMRAM Stage competion density Average numbers of hospitals within a 10km area Stage 0-2 Stage 3-5 Stage 6 and 7 Count Row N% Count Row N% Count Row N% non-Randstad 1.3 13 30.2% 25 58.1% 5 11.6% Randstad 4.5 12 52.2% 8 34.8% 3 8.3% Non-Randstad region has more digital potential H3: Larger hospitals are more likely to have advanced EMR capabilities. Although not statistically significant, the general pattern emerging from this study (Table 5) supports the hypothesis that larger hospitals tended to have more advanced EMR capabilities than smaller hospitals. This pattern is reflected in the number of staffed beds. 2

32 Chapter 2 TABLE 5: Hospital Size by EMRAM Stage Stage 0-2 Stage 3-5 Stage 6 and 7 Mean Row N% Count Mean Row N% Count Mean Row N% Count number of beds 422 37.9% 501 50.0% 582 12.1% number of beds 0<340<562 large 28.6% 6 57.1% 12 14.3% 3 medium 30.8% 8 57.7% 15 11.5% 2 small 57.9% 11 31.6% 6 10.5% 2 Larger hospitals tended to have more advanced EMR capabilities than smaller hospitals H4: Academic affiliated hospitals are more likely to have advanced EMR capabilities. Academic affiliated hospitals not only have a higher percentage of Stage 6 and Stage 7 hospitals (21.2%) than general hospitals (10.3%) but have a much lower percentage of hospitals in the entry EMRAM stages (27,3% versus 48,7% respectively) (see Table 6). These findings support the hypothesis that hospital type tends to influence a hospital’s EMR capabilities. TABLE 6: Hospital type by EMRAM Stage Stage 0-2 Stage 3-5 Stage 6 and 7 Count Row N% Count Row N% Count Row N% academic affiliated Yes 9 27.3% 17 51.5% 7 21.2% No 19 48.7% 16 41.0% 4 10.3% Academic affiliated hospitals have more advanced EMR capabilities. H5: Hospitals with a relatively higher ICT budget tend to have more advanced EMR capabilities. Findings of this study (reflected in Table 7) support the hypothesis that hospitals with advanced EMR capabilities tend to have higher ICT budgets than those hospitals with lower EMR capabilities. The budgetary demands appear to grow in a fairly linear manner, raising notable concerns worthy of future exploration.

33 Profile of the digitization of patient medical records in Dutch hospitals TABLE 7: Average Hospital ICT Budgets by EMRAM Stage Stage 0-2 Stage 3-5 Stage 6 and 7 Mean Mean Mean ICT budget in € millions 5.1914 6.7400 8.3000 ICT budget as percentage of hospital budget 3.06% 3.63% 4.17% Hospitals with higher IT budgets score better. H6: Hospitals with more ICT employees have more advanced EMR capabilities. The findings of this study support the hypothesis that hospitals withmore ICT employees tend to have more advanced EMR capabilities. As reflected in Table 8, the number of ICT workers per hospital bed increases with advancing EMR capabilities. These findings unfortunately do not detail the type of ICT workers. Future research efforts might want to consider if there is a shifting demand in the type of ICT worker as hospitals progress in their EMR capabilities. TABLE 8: Average Number of ICT Employees by EMRAM Stage Stage 0-2 Stage 3-5 Stage 6 and 7 Mean Mean Mean numbers of ICT employees per hospital bed 0.0782 0.0831 0.1015 Hospitals with more ICT employees tend to have more advanced EMR capabilities DISCUSSION Leveraging the HIMSS Analytics EMRAM to define the basic EMR as one that exceeds the infrastructure of an EMRAM Stage 5 hospital, we found that only 15.3% of NL hospitals during our study period had a basic EMR system in at least one clinical area (EMRAM Stage 6). With so few NL hospitals operating a basic EMR system, we are left to question those factors that might be at work influencing a hospital’s EMR progression. This study considered two major classes of variables: the EMRAM requirements themselves and intervening organizational and environmental forces. With respect to EMRAM requirements, the findings of this analysis suggest NL hospitals may be particularly challenged in addressing one distinct requirement of the HIMSS Analytics EMRAM. With 37.5% of hospitals having successfully met the requirements of Stage 2 but not Stage 3, the requirements of Stage 3 appear to be a challenge for a sizeable percentage of NL hospitals. And indeed, when looking further into the data, we find that electronic nursing/clinical documentation presents as a particular challenge. We are uncertain as 2

34 Chapter 2 to why so many hospitals in the Netherlands have been apparently slow to implement a nursing documentation system but do question if it has something to do with the value placed on nurses in Dutch hospitals. While nurses play a significant role in patient care in a hospital, their role (and by extension perceived value) are often times secondary to the role of the attending physician. If true, then it is possible IT leaders are placing a greater emphasis on EMRAM applications targeted towards physicians than nurses. A notable observation: physician related applications (Computerized Physician Order Entry; Physician Documentation) are higher order applications in EMRAM. This hypothesis certainly warrants further exploration as there is a strong argument to be made in prioritizing the automation of nursing documentation, especially as a means of reducing the transmission of erroneous patient information. Once EMRAM Stage 3 requirements are met, the EMRAM profile then suggests NL hospitals are likely to be challenged in meeting the requirements of EMRAM Stage 6 (Closed loopmedication administration [CLMA] and advanced decision support [CDSS]), with 43.1% of the hospitals in EMRAM Stage 5. The CLMA process includes ePrescribing, medication dispensing and tracking, and administration and documentation in the electronic Medication Administration Record (eMAR). A CDSS function (i.e., alerts) must be available at the point of care immediately prior to administration to ensure the five rights of administration check (right patient, right medication, right dose, right route, and right time). Especially the guarantee of the right medication and the right dose is a challenge in the NL, as bar coded unit doses are not always readily available from the pharmaceutical industries by lack of European bar code standards for drugs. The other class of variables influencing EMR adoption in NL hospitals involves organizational and environmental forces. By considering a wide array of relevant variables, the results of this study support the general assertion that EMR adoption is influenced by organizational and environmental forces. More specifically, variances in EMR adoption rates varied notably by hospital size and hospital type. The same holds for smaller hospitals. Smaller hospitals are unlikely to have the financial or human resource means to implement and use an EMR system. This is consistent with previous research that has identified cost as the greatest barrier to EMR adoption and use. 2,18 When hospitals make an investment in an EMR system and when the implementation is successful, the payers and purchasers also benefit. This misalignment of incentives represents perhaps the single most important barrier to moving ahead. Additionally, it is possible that a smaller hospital may not have the human resources available to run such a system. If this is the case, these smaller hospitals may need to form a coalition to investigate the feasibility of a

35 Profile of the digitization of patient medical records in Dutch hospitals group purchase and implementation of EMRs. Because EMRs are expensive, and larger hospitals have begun using EMRs more than smaller hospitals, it is possible that without greater economies of scale for implementation, EMRs are too costly for the smaller hospitals. If EMRs do in fact improve hospital quality or efficiency, policymakers should take steps to encourage hospital EMR use. These steps could include programs that aid hospitals in implementing and using EMRs with EMR hardware and software, as well as training and personnel to help with implementation. These programs will be especially important to smaller hospitals. Policymakers may also, at some point, offer greater financial reimbursement for hospitals that use EMRs as a way to encourage hospital use. Additionally, more regulations frompayer groups and policymakers can ensure that hospital EMR use is practiced. These regulations may be in the form of requirements for certification, endorsements, or accreditation. Previous research has concluded similar incentives are necessary for more widespread EMR adoption. 19 Three of the six university hospitals represented in this study (total of eight university hospitals in the Netherlands) have full EMR systems in use in at least one clinical unit. University hospitals have far more resources than non-university hospitals and have a different financial model with more government funding. They have more IT budget and more ICT employees. Teaching hospitals are well represented in the group that is implementing additional functionalities, which also reflects other findings. They have higher and better resources. Hospitals with higher ICT budgets were expected to do better, which is the case. CONCLUSION Although small hospitals and hospitals located in the northern part of the Netherlands were underrepresented in the study, the 72 hospitals that did participate provided a fairly good representation of the total population of the Netherland’s 93 hospitals. Although the group is relatively small, and no robust statistically significant conclusions could be drawn, three of the six hypotheses were supported. Adoption and use of the EMR tends to be positively associated with larger hospitals and teaching hospitals. Not found was a relationship between hospital density and population density, as a measure for the competition level, and the EMRAM score. According to the Resource Dependence theory, resources in areas with low density populations are scarcer in areas of lower population, (potential) patients have less choice, and hospitals feel less urgency to adopt advanced technologies like EMRs. No proof was found for this statement in the Netherlands in this study. Competition in the Netherlands is a new and developing phenomenon. The healthcare insurance companies and the Government play a dominant role. So 2

RkJQdWJsaXNoZXIy MjY0ODMw