Volume 2, Issue 1B
DECISION AND INFORMATION TECHNOLOGIES
The Missing Link: Electronic Health Record Adoption by Physician Practices Catherine Anderson, Ph.D. Candidate, Smith School of Business Abhay Mishra, Assistant Professor, Smith School of Business Ritu Agarwal, Professor and Robert H. Smith Dean’s Chair of Information Systems and Director of CHIDS Corey M. Angst, Research Assistant Professor and Associate Director of CHIDS Steve Labkoff, Director, Healthcare Informatics, Pfizer Inc
The vast majority of care delivery in the US takes place in physician practices, yet the investigation of the drivers of adoption of electronic health record systems (EHR) at the practice level has been limited. An electronic health record is simply the computerization of health record content and associated processes. EHRs (EHR systems) are the software platforms that physician offices and hospitals use to create, store, update, and maintain electronic health records for patients. They are a critical foundational technology through which interoperability of health information can take place. Barriers to adoption such as cost, privacy, and disruptions in workflow have been often referenced as deterrents to EHR adoption, but little empirical evidence is offered to support these assertions. In fact, the limited research that is available suggests that the benefits of EHR adoption outweigh the costs. Surprisingly, what one cannot dispute is the slow rate of diffusion of EHRs in physician practices, particularly the small practices. In this study, we investigate the drivers of EHR adoption in physician practices, hereafter “practices.” We argue that characteristics of the technology, health system as a whole, and unique aspects related to connections between clinicians and other stakeholders create a dynamic that necessitates a reframing of traditional adoption models. For example, nearly all practices – even sole-physician practices – have hospital-physician affiliations and/or admitting rights with one or more hospitals and clinics. In the past, physicians who were not a part of a hospital-physician affiliation would have felt little pressure to conform to specific practices of their admitting hospitals. Now, with the increased diffusion of EHRs into hospitals, independent practices may realize the benefits of interfacing with the hospitals. In addition, hospitals may impart pressure on the practices to provide electronic interfaces. Thus, while doctors may appear to
Spring 2007
be ‘free-agents,’ decisions about their practice related to technologies such as EHRs will have implications at multiple levels. The most practical way for health information to be made interoperable is through the linking of electronic health record systems. However, it will not suffice to have EHRs installed in all hospitals across the country. Nationwide interoperability of health information, as defined by the Office of the National Coordinator for Health Information Technology, can only be achieved if medical records are digitized in all health care facilities across the country including physician practices, which are the focus in this study. In a health system, interoperability essentially means that at a minimum, one computer system will exchange data with another computer system. In more detailed terms, interoperability should provide that the data embedded in a message convey the same meaning on the receiving system as it does on the sending system. Interoperability is important because it is the mechanism through which not only treating clinicians in various locations can have access to the same data, but it can also potentially provide links to many, if not all stakeholders (e.g., hospitals, insurance companies, laboratories, employers, pharmaceutical and medical device firms, etc.) in the health care value chain. EHRs are enabling technologies, but also disruptive innovations, not only because they introduce changes in workflow, but because of the potential impact they can have on the health system as discussed earlier. These seemingly opposing roles of EHRs as both enablers and disruptors, combined with the unique environment, make this an important area for adoption research. To inform the topic of EHR adoption at the practice level, we draw from motivation-ability literatures and suggest that the adoption decision is a function of the practice’s ability (i.e. infrastructure and human capital) and motivation (i.e. external pressure and internal beliefs). We conducted a national survey of small and medium-sized physician practices in the US in early 2007. Our survey yielded 273 usable responses including over 100 practices who had adopted EHRs. Research Framework As decision making in physician practices is influenced by both organizational and individual physician factors,
CHIDS • Van Munching Hall • College Park, MD • University of Maryland • www.smith.umd.edu/chids •
[email protected] • 301.405.0702 © 2007 Center for Health Information and Decision Systems. CHIDS Research Briefings are published on a quarterly basis to update CHIDS members.
p. 2
CHIDS Research Briefing
Volume 2, Issue 1B
the motivation and ability framework, which has been successfully applied at both levels, provides a strong foundation for our study. Literature in the organizational sciences suggests that motives, structures, and processes influence firm behaviors. Therefore, we expect salient motivating, infrastructure, and readiness factors to influence a practice’s decision making behavior with regard to EHR adoption. Likewise, at the individual level, motivation and ability are core components in the widely accepted and well established consumer information processing theory. We expect physicians, like consumers, will process information and make decisions if they are motivated and have the ability to do so. We expand on the motivation component by applying self-determination theory (SDT) which describes a continuum of autonomy that exists with extrinsic motivation being largely controlled externally at one end and intrinsic motivation largely autonomous at the other. SDT theory distinguishes between degrees of control and autonomy in various forms of extrinsic motivation that differentially influence the essentially positive influence of intrinsic motivation. The more controlling the extrinsic motivation, the more undermining its influence can be on intrinsic motivation. We anticipate an interactive effect of these influences on the EHR adoption decision in physician practices (see Figure 1). Controlled
Practice Motivation
Autonomous
Extrinsic Coercive Extrinsic Normative Intrinsic Perceived Pressure Pressure Value
Practice Ability IT Infrastructure IT-Related Intangibles
EHR Adoption
Physician/Staff IT Readiness Signifies a moderated relationship
infrastructure, IT related intangibles and physician/staff IT readiness. Collectively these resources reflect the practice’s ability to overcome the technical and organizational barriers associated with complex IT and successfully adopt the new IT. Thus we predict that superior IT infrastructure is associated with EHR adoption. IT-related intangibles refer to the extent to which an organization is using IT in value-adding ways. In the context of physician practices, if a practice is already demonstrating an ability to leverage IT to achieve intangible patient benefits such as customizing service or to create synergy by exchanging data electronically with authorized parties, it is demonstrating a high level of IT capability which makes the practice more likely to adopt new technology that will further provide such benefits. Such practices will see the value in adopting a technology that can enable them to continue along a path to achieving IT related intangible benefits. Thus, superior ITrelated intangibles will be associated with EHR adoption. A significant body of research in IT has argued that human capital is a key aspect of successfully leveraging IT within an organization. Our focus is on the IT-related skills of physicians and staff. The more the clinical and support staffs have experience with and use technology, the more receptive they are likely to be to new technology. Similar to infrastructure, the human resource component can serve as an enabler to innovation and adoption. Therefore we predict that superior physician/staff IT readiness is associated with EHR adoption. Motivation Merely having the capability to perform is insufficient; action is strongly contingent upon the drive of the focal entity. Practices face pressure from external sources such as regulatory agencies, insurance companies and pharmaceutical companies as well as other practices and its own patients. Because the decisions in a physician practice are strongly influenced by individual physicians and their perceptions, we use the literature on organization behavior grounded in psychology for insight into the factors that motivate individuals.
Figure 1. Research Model Ability Physician practices posses different levels of ability based on their unique capabilities and resources that can be brought to bear in the event they choose to implement an EHR. This baseline level of “stock” is considered during the adoption decision making process. Indeed, empirical support exists for the mediating influence of system usage on the relationship between technology competence and firm value. We focus on three core-IT resources: IT
Intrinsic motivation involves an individual acting out of an internal belief that the activity is interesting, good, satisfying or right. In contrast, extrinsic motivation involves action propelled by a desire to achieve or gain some external reward or benefit or to avoid a consequence. Originally, scholars believed that that intrinsic and extrinsic rewards were additive and that the more intrinsic and extrinsic rewards an individual received; the more motivated he became. However,
Spring 2007 p. 3
several studies showed that some forms of extrinsic motivation actually attenuate and undermine intrinsic motivation, by diminishing an individual’s feeling of autonomy. In the context of our study, if a physician believes that EHRs will enable him to diagnose more efficiently, increase the quality of care provided to his patients, and complement his own knowledge, he is likely to understand the value of EHRs and be intrinsically motivated to adopt the technology as a result. As these beliefs are not imposed upon him by an outside entity, he has a high level of autonomy over them, which is consistent with the professional culture to which he is accustomed. However, intrinsic motivation is not the only form of motivation that must be considered in the EHR context. Pressure exerted by external entities such as regulatory bodies, vendors, and insurance companies is coercive and extrinsically imposed. Although this extrinsic coercive pressure may be a motivating factor, it is a form of controlled motivation which has been shown in prior studies to undermine the positive influence of intrinsic motivation. It has been suggested that the dampening effect of extrinsic motivation is especially true for complex tasks such as the adoption and implementation of EHR. Thus, we hypothesize that extrinsic coercive pressure will dampen the relationship between intrinsic perceived EHR value and EHR adoption. A second form of extrinsic motivation for practices is normative in nature and addresses the question of how many other practices with which the focal practice routinely interacts have already adopted EHRs. For example, in the provision of care for a particular patient, a pulmonary specialty practice may share information with oncology, radiology and cardiology practices. If these three other practices have already adopted EHRs and the focal practice has not, it may feel pressure to adopt in order to conform and keep pace with its peer practices. Furthermore, through its connections to other practices, the focal practice can learn about EHR and its associated costs and benefits. In a study of interorganizational linkages adoption, this form of normative pressure exhibited the strongest influence on organization-level technology adoption. We therefore predict that extrinsic normative pressure amplifies the relationship between intrinsic perceived EHRs value and EHR adoption. In much the same way that extrinsic coercive pressure moderates the relationship between intrinsic perceived value and adoption, we hypothesize that extrinsic coercive
pressure negatively moderates the relationship between extrinsic normative pressure and EHR adoption. Research Methods To test our research model, we collected survey data from physician practices spread across the U.S. Based on input from healthcare IT experts, healthcare informatics professionals, and physicians actively involved in the implementation of healthcare IT, we refined the preliminary survey instrument and administered it electronically through announcements provided by the American College of Physicians (ACP), American Medical Informatics Association (AMIA), and ACPnet, a practice-based research network that volunteers to examine health care processes. Through three rounds of collection, we gathered data from 273 respondents representing different physician practices. Analysis and Results Table 1 presents descriptive statistics for all research variables. We estimated the research model using binary logistic regression with interaction terms. In this approach the model is estimated initially using only main effects and interaction effects are added to the base model subsequently. Table 1. Descriptive Statistics and Correlations Constructs and Variables (V1) IT Infrastructure (V2) IT-Related Intangibles (V3) Physician/ staff IT readiness (V4) Extrinsic Coercive Pressure (V5) Extrinsic Normative Press. (V6) Intrinsic Perceived Value (V7)EHR Adoption
Mean
SD
V1
V2
V3
V4
V5
V6
V7
16.29 4.23 1.0 8.64
4.03 .61 1.0
4.31
2.62 .41 .66 1.0
16.48
.14
1.0
.07 .24 .19
.14
1.0
20.52 4.56 .29 .43 .29
.12
.03 1.0
-.12
.27
.26
.41
5.12 .05 .13 .24
.50
.40 .37 .08
.13
1.0
We first introduced control variables into the specification, followed by main effects, followed by interactions. Two of the three ability-related variables are significant. As predicted, IT infrastructure and IT-related intangibles are both positively associated with EHR adoption (see Table 2). Our other hypothesized ability component, physician/staff readiness is not significant. In the presence of our interaction terms, there remains only one significant motivational main effect for extrinsic normative pressure which is positively associated with EHR adoption. All three of the interaction terms are significant and negatively associated with EHR adoption.
CHIDS Research Briefing
Spring 2007 p. 4
Volume 2, Issue 1B
Table 2. Logit Analysis Results: Results: Full model Variables
Interaction
Motivation
Ability
Control Variables
Constant
Parameter estimates
pvalue
-.107
.139
Practice Size
.000
.449
Number of Insurance Plans
-.002
.718
Number of Professional Societies
.049
.742
No. Hospitals with Admitting Privileges
.048
.774
Champion
-.825
.384
IT Infrastructure
1.664
.000
IT-Related Intangibles
1.522
.000
Physician/staff IT Readiness
.263
.250
Extrinsic Coercive Pressure
-.227
.546
Extrinsic Normative Pressure
4.560
.003
Intrinsic Perceived Value
.776
.073
Ext. Coer. Press* Intrinsic Perceived Value
-.79
.017
Ext. Norm. Press*Intrinsic Perceived Value
-2.438
.049
Ext. Coer. Press*Extrinsic Norm. Pressure
-3.089
.036
Nagelkerke’s R-square
.611
Conclusion While the ability components of IT infrastructure and ITenabled intangibles have a straight-forward positive relationship to EHR adoption, the motivational components and their relationships to adoption are more complex. We hypothesized that extrinsic coercive pressure would moderate the positive relationship between intrinsic perceived EHR value and EHR adoption. As shown in Table 2, the interaction term is significant and the beta value is negative. This serves to attenuate the marginally significant and positive relationship of value on adoption, thus this relationship is supported. We argued that extrinsic normative pressure would positively moderate the effect of intrinsic motivation on EHR adoption. This expectation was based on the logic that to the degree physicians have a strong professional culture and identity, external pressure from similar others would be positively reinforcing. Contrary to predictions, we find the same negative interaction as with coercive pressure, suggesting that physicians perceive the pressure from other physicians as “undesirable” and value their autonomy more. Finally, we hypothesized that extrinsic coercive pressure would moderate the positive relationship between extrinsic normative pressure and EHR adoption. As shown in Table 2, the interaction term is significant and the beta value is negative. This attenuates the significant and positive relationship of extrinsic normative pressure on adoption.
can be undermined by external pressure exerted by outside sources such as regulatory agencies and pharmaceutical companies. In addition, a physician practice can be positively influenced by other practices it interacts with that have already adopted EHR. Although these other practices are external, they are peer influences which culturally play a strong role in the physician profession. However, if this positive influence is present in combination with the external pressure exerted by outside forces, the positive influence is again undermined. Implications Our findings suggest that care should be taken to implement policies and design incentives targeted toward increasing HER adoption among practices. External pressure can have an adverse impact on adoption, and must be used carefully. Policy makers need to judiciously reconsider the nature and extent or pressure they want to impose on practices to adopt EHR. Suggested Citation: Anderson, C., Mishra, A.N., Agarwal, R., Angst, C.M., and Labkoff, S., "The Missing Link: Electronic Health Record Adoption by Physician Practices " CHIDS Research Briefings (2:1B), Center for Health Information and Decision Systems, Robert H. Smith School of Business, University of Maryland, Spring 2007, pp 1-4. CHIDS Contact Information Director – Ritu Agarwal, Professor and Robert H. Smith Dean's Chair of Information Systems Associate Director – Corey Angst, Research Assistant Professor Center for Health Information and Decision Systems Robert H. Smith School of Business University of Maryland Van Munching Hall College Park, Maryland 20742 Ph: 301.405.0702
[email protected] www.smith.umd.edu/chids CHIDS Membership Information Become a corporate member of CHIDS and realize the benefits of having access to a world-class team of researchers. All levels of membership provide your organization with the following benefits: - Early insights from research - Access to top scholars - Quarterly updates on key HIT activities - Participation in bi-annual conferences - Opportunity to influence CHIDS’ research agenda - Priority access to research projects CHIDS Corporate Members Johnson & Johnson Health Care Systems, Inc. Pfizer Inc CapMed, a Division of Bio-Imaging Technologies, Inc. CTIS Inc.
We find that EHR adoption in physician practices is based on a combination of ability and motivation. Although a physician practice may see the value in adopting an EHR, the positive influence of this intrinsically motivating factor CHIDS • Van Munching Hall • College Park, MD • University of Maryland • www.smith.umd.edu/chids •
[email protected] • 301.405.0702 © 2007 Center for Health Information and Decision Systems. CHIDS Research Briefings are published on a quarterly basis to update CHIDS members.
