This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Medical Informatics, is properly cited. The complete bibliographic information, a link to the original publication on http://medinform.jmir.org/, as well as this copyright and license information must be included.
The Health Information Technology for Economic and Clinical Health Act (HITECH) allocated $19.2 billion to incentivize adoption of the electronic health record (EHR). Since 2009, Meaningful Use Criteria have dominated information technology (IT) strategy. Health care organizations have struggled to meet expectations and avoid penalties to reimbursements from the Center for Medicare and Medicaid Services (CMS). Organizational theories attempt to explain factors that influence organizational change, and many theories address changes in organizational strategy. However, due to the complexities of the health care industry, existing organizational theories fall short of demonstrating association with significant health care IT implementations. There is no organizational theory for health care that identifies, groups, and analyzes both internal and external factors of influence for large health care IT implementations like adoption of the EHR.
The purpose of this systematic review is to identify a full-spectrum of both internal organizational and external environmental factors associated with the adoption of health information technology (HIT), specifically the EHR. The result is a conceptual model that is commensurate with the complexity of with the health care sector.
We performed a systematic literature search in PubMed (restricted to English), EBSCO Host, and Google Scholar for both empirical studies and theory-based writing from 1993-2013 that demonstrated association between influential factors and three modes of HIT: EHR, electronic medical record (EMR), and computerized provider order entry (CPOE). We also looked at published books on organizational theories. We made notes and noted trends on adoption factors. These factors were grouped as adoption factors associated with various versions of EHR adoption.
The resulting conceptual model summarizes the diversity of independent variables (IVs) and dependent variables (DVs) used in articles, editorials, books, as well as quantitative and qualitative studies (n=83). As of 2009, only 16.30% (815/4999) of nonfederal, acute-care hospitals had adopted a fully interoperable EHR. From the 83 articles reviewed in this study, 16/83 (19%) identified internal organizational factors and 9/83 (11%) identified external environmental factors associated with adoption of the EHR, EMR, or CPOE. The conceptual model for EHR adoption associates each variable with the work that identified it.
Commonalities exist in the literature for internal organizational and external environmental factors associated with the adoption of the EHR and/or CPOE. The conceptual model for EHR adoption associates internal and external factors, specific to the health care industry, associated with adoption of the EHR. It becomes apparent that these factors have some level of association, but the association is not consistently calculated individually or in combination. To better understand effective adoption strategies, empirical studies should be performed from this conceptual model to quantify the positive or negative effect of each factor.
The US Government passed the Health Information Technology for Economic and Clinical Health (HITECH) act [
Adoption of the EHR is a significant goal. International vernacular for the EHR varies; for example, electronic patient record, computerized patient records, electronic medical records (EMRs), and digital medical record. The defining difference, as defined by the Institute of Medicine, the health arm of the US National Academy of Sciences, focuses on the longitudinal and interoperable nature of the electronic patient record [
Studies estimate that adoption of the EHR could eventually save more than $813 billion annually, prevent 200,000 adverse drug events, and enhance the doctor-patient relationship through increased communication [
The environment of health care is unique in a competitive environment. The health care organization develops an organizational strategy based on the local environment. To increase an organization’s ability to compete, its strategy might also include cost reduction, and EHR adoption runs counter to this goal in the SR. The health care environment faces many sources of influence, including a reluctance to accept technology.
There has been a tremendous amount of research dedicated to the study of acceptance of technology, specifically the Technology Acceptance Model (TAM) [
Organizational theories address influence, but none adequately addresses the complexity of the health care organization. Payers, providers, and patients all control resources that exert influence. The nature of the competitive environment will also exert influence on decisions. External influence from those who control resources can be explained through resource dependence theory [
According to resource dependence theory, health care organizations with the greatest level of dependence on other organizations that control the resources will feel the greatest level of environmental influence on its decisions [
Similar to Resource Dependence, the Diffusion of Innovation Theory describes a social system that influences through communication channels [
The concept of compatibility [
The concept of complexity [
The concept of trialability [
The concept of observability [
Relative advantage is a multifaceted concept for this study. In health care, the most important factor is provision of health, as well as the treatment and prevention of disease. If adoption of the EHR speaks directly to the health care organization’s primary purpose, then it might provide relative advantage over competitors that have not adopted it. Another concept is that of social prestige [
Strategy can be a multifaceted concept, and organizations around the world hire strategy experts to help identify and focus on a market forces. An operational definition of strategy is borrowed from education [
Studies have shown that decision making in the health care industry is often based on how the organization competes, whether in a single-market or multimarket environment [
Ash and Bates [
Factors that influence health information system (HIS) adoption in US hospitals have been studied by others as well (n=1441) [
A knowledge-based taxonomy of critical factors for adopting an EHR was developed from a systematic literature review [
Alternative measures of EHR adoption among hospitals have been studied [
Others have studied the relationship between hospital financial position and the adoption of the EHR [
Because commonly used elements of organizational strategy are difficult to change, several of the variables were categorized as internal organizational factors. Research has assessed variables of hospital influence in five categories: (1) capacity as measured by number of beds in groupings by intervals of 100, (2) management, or ownership, (3) organizational focus, or teaching status, (4) competitive location and alternatives, and (5) state regulatory pressures [
Although resources have been consumed to study factors associated with adoption of HIT, there is a gap in the literature that provides a conceptual model to guide the design of empirical models. It may seem backward to design a conceptual model after so many studies have already been conducted, but the gap remains. The aim of this study was to develop a conceptual model from a systematic literature review that associates both internal and external factors associated with adoption of the EHR. The intent of the conceptual model is to enable future empirical models.
Search terms were selected based on the experience of the authors in the field of health care administration. The time frame of 1993-2013 was selected as convenience. It was assumed that 2 decades would be sufficient to capture trends.
The 83 records were reviewed for content and evidence. After discarding 58 articles for lack of evidence, three additional references were added because they were key concepts upon which other studies were based. Of the remaining 25 articles, a list of factors was identified as IVs. Some factors were grouped under a similar category for the purposes of simplification of the conceptual model. The dependent variable (DV) started as adoption of the EHR, but the studies from those chosen were not as specific. From personal experience, many studies seem to discuss the EHR, but call it something else: most commonly the EMR. That is why EMR was included in the search. Because so few ERHs exist without some form of CPOE, the latter term was included in the search criteria.
Our study combines the influences highlighted by previous work and examines determinants of EHR adoption. Examining EHR adoption at the health care organization level will demonstrate validity between this study and others that have used the hospital as the unit of analysis.
Literature review process.
Several influences in the environment exert pressure on the health care organization to adopt EHR. Influences range from incentives from the federal government to the nature of local competitive community. US federal incentives provide a heavy influence for EHR implementation, under specific conditions, and imposes penalties for a lack of EHR implementation.
The internal politics of one organization serve as one source of influence. A hospital is part of a community, which serves as an external influence. Further, if a hospital is also part of a larger multihospital system (MHS), then the politics of the broad MHS will also exert influence on local decisions.
The patient is external to the organization, and for our study, the patient primarily serves as an external influence. Although some employees of the health care organization might also be patients, and this relationship could create a small internal influence, this study considers those few stake holders in the internal organizational factor of users. The providers are internal to the organization, and for our study, providers serve as an internal organizational influence. The payer is a significant influence [
The premise for an EHR adoption conceptual model is that that environmental influences affect organizational strategy of the health care organizations that adopt the EHR [
Elements of organizational strategy are not variables that can be easily changed [
Resource Dependence Theory explains environmental influences and the external interdependence of organizations [
Disparate stakeholders have different interests with reference to different components of the EHR. These interests may be different in the SR interests versus the LR interests. SR interests are those that are immediate, such as current year expenditures. LR interests are further out when all inputs are variable. The SR interests of cost can often compete with the LR potential of cost savings and greater safety. Both the SR and LR interests are affected by the external environment [
In a highly competitive environment, SR cost implications could often win over any long-term savings. The number of patients in a market is fixed in the SR, and a highly competitive market will affect each competitor’s share of that market. The SR costs of EHR implementation might be insurmountable by an organization in this market because it could not afford to lose ground without significant capital reserves or the ability to borrow cheaply [
External stakeholders that control resources important to the health care organization can exert significant influence. For instance, a health care organization that receives a significant amount of revenue from the CMS will be influenced more by incentives provided by the CMS than an organization that receives a significant cash flow from private third parties. The relative influence of various external stakeholders may be captured by an analysis of the structure of the market in which a health care organization operates.
Stakeholders have varying interests with regard to the capabilities and effects of EHR components depending upon their relationship with the health care organization. Private payers have both SR and LR interests in the EHR. In the SR, their focus is on minimizing expenditures. Because the health care organization would pass on the implementation costs through higher contract costs, payers would not be equal in the SR. In addition, the disruption of EHR implementation could potentially affect care processes and therefore increase claims. Payers would be interested in the LR benefits of the EHR: potential cost savings, better disease management, and increased safety. However, the SR interests of the private payers might overshadow the LR benefits of the EHR. Public payers enable care of the indigent and elderly. As part of the United States Department of Health and Human Services (HHS), the CMS is highly interested in disease management, public health, safety, and research, and it may value these LR capabilities of the EHR more than the SR costs. The CMS, as part of HHS, would also favor the EHR because it supports the Presidential directive to promote the establishment of the Nationwide Health Information Network that links electronic patient records through health information exchanges.
Providers and patients value face time with each other. During EHR implementation, providers might spend less time in communication with patients. Providers must adapt their processes and clinic-to-administrative schedules. Any disruption or action that is perceived as deleterious to this relationship could result in a negative reaction to EHR implementation. As a result, physicians might oppose EHR adoption, or they might simply support the EHR solution with the shortest implementation time or least administrative burden. Patients might not like the reduced face time with the provider, but they might be attracted to EHR components such as e-prescribing, e-results, personal health records, and email access to the provider. These desirable features are available to the patient when the health care organization chooses to adopt various portions of the CPOE component to the EHR.
The articles chosen for final inclusion were read once more to make a list of variables. The variables from the studies were listed as internal and external. There were significant commonalities in the variables used, so they were combined in the model.
As depicted in
The dependent variable was not consistent: seven references used EHR adoption [
As depicted in
As previously stated, there was overlap between the sources/theories. There were four internal forces and seven external forces identified through multiple works by three authors [
This framework captures both internal and external factors that influence the adoption of the EHR. The positive (+) and negative (−) signs in the model describe the relationship identified by the associated authors. For instance, Gin et al [
Conceptual model of factors associated with adoption of the EHR.
The main findings of this study were that nine studies identified external factors and 16 studies identified internal factors associated with the adoption of EHR. These factors were depicted in a conceptual model to describe relationships to EHR adoption.
The conceptual model for EHR adoption illustrates a framework within which both administrators and policy makers can work to understand the levers that exert significant influence in the adoption of EHR. The extensive literature review conducted by this study builds a unique model from which empirical studies can be designed.
Identifying relationships between the adoption factors and adoption of the EHR becomes significant because it identifies levers that will produce a desired action. For instance, if a hospital has a majority of senior providers, perhaps from the Baby Boom generation, the administrators become aware of the additional effort that needs to go into user acceptance. A hospital that has a majority of new providers will not need to expend the resources on user acceptance, because studies already show a penchant for technology in younger generations. Similar inferences on other factors of adoption could be made, and some would require additional study.
For instance, the literature on workflow impact is split. There seems to be evidence that the presence of the EHR both enhances and complicates the providers’ workflow. This observation clearly begs additional questions. Were subjects for the data in different phases of adoption of the EHR? Was the hospital that responded negatively in the middle of an EHR implementation? Logically, a large implementation of any technology will become disruptive to the organization. Several studies could emerge from this relationship alone.
Empirical models could easily be designed to further investigate specific relationships between the IVs and DVs. The set of studies on CPOE was interesting. Although there were some overlaps with adoption of the EHR, there were also studies that only looked at CPOE. There does not seem to be an abundance of evidence in the literature about CPOE, and yet the AHA regularly collects data on six different versions of CPOE: laboratory, radiology, pharmacy, nursing, physician notes, and consults. There was no data to be found about the use or efficacy on CPOE consults. It might be interesting to determine the reason for this paucity of data.
The EHR adoption conceptual model associates internal and external factors with the adoption of the EHR, but it is primarily based on an extensive literature review. So far, it is not empirically tested. However, data are available to test the theory. Because the findings of our study are descriptive in nature, we do not opine on appropriate medical use of the information.
Caution should be identified with the interaction of variables. Some variables will most likely confound or mask the effects of others. For instance, is there a direct relationship between the number of beds of a hospital and the number of full-time equivalents? There are staffing models that would most likely answer that question. If there is a similar relationship, then one of these variables should be eliminated in favor of the stronger one. Otherwise, the effects of the weaker variable will be masked by the other. A false conclusion could easily be identified concerning the masked variable.
A majority of references for this study were from the United States, with one exception from Hong Kong. The internal validity of this study is strongest within the US health care sector. The conceptual model might be limited outside the United States because of the nature of competition between hospitals. The analysis of 83 articles identified studies that used similar methods: survey or secondary data analysis. Many authors analyzed data from the AHA, a well-established data set in the United States. These data are self-reported, which comes with limited bias.
This study also identified overlap between studies in terms of variables. One interpretation of this overlap could be that the variables and associated studies are highly reliable. The key word/phrase searches described in the Methods section identifies the databases queried and results given. Other researchers should be able to duplicate or update this conceptual model going forward.
American Hospital Association
American Recovery and Reinvestment Act
Center for Medicare and Medicaid Services
computerized provider order entry
dependent variables
electronic health record
electronic medical record
the Department of Health and Human Services
Health Information Management Systems Society
health information system
health information technology
Health Information Technology for Economic and Clinical Health
information technology
independent variables
long run
multihospital system
short run
Technology Acceptance Model
The topic of research and conceptual model in this original article are from a chapter of the PhD dissertation of Kruse C DeShazo J, and Kim F sat on the dissertation committee, which helped direct the research and develop the model. Fulton L helped Kruse C extract portions of the dissertation worthy of publication.
None declared.