Health Information Exchange (HIE): A literature review, assimilation pattern and a proposed classification for a new policy approach

Journal of Biomedical Informatics 64 (2016) 74–86

Contents lists available at ScienceDirect

Journal of Biomedical Informatics journal homepage: www.elsevier.com/locate/yjbin

Health Information Exchange (HIE): A literature review, assimilation pattern and a proposed classification for a new policy approach Pouyan Esmaeilzadeh, PhD a, Murali Sambasivan, PhD b,c,⇑ a

Department of Information Systems and Business Analytics, College of Business, Florida International University, Miami, FL 33199, United States Taylor’s Business School, Taylor’s University Lakeside Campus, Malaysia c Victoria University, Melbourne, Australia b

a r t i c l e

i n f o

Article history: Received 1 February 2016 Revised 12 August 2016 Accepted 15 September 2016 Available online 17 September 2016 Keywords: Health Information Exchange Adoption Assimilation Implementation Institutionalization

a b s t r a c t Objectives: Literature shows existence of barriers to Healthcare Information Exchange (HIE) assimilation process. A number of studies have considered assimilation of HIE as a whole phenomenon without regard to its multifaceted nature. Thus, the pattern of HIE assimilation in healthcare providers has not been clearly studied due to the effects of contingency factors on different assimilation phases. This study is aimed at defining HIE assimilation phases, recognizing assimilation pattern, and proposing a classification to highlight unique issues associated with HIE assimilation. Methods: A literature review of existing studies related to HIE efforts from 2005 was undertaken. Four electronic research databases (PubMed, Web of Science, CINAHL, and Academic Search Premiere) were searched for articles addressing different phases of HIE assimilation process. Results: Two hundred and fifty-four articles were initially selected. Out of 254, 44 studies met the inclusion criteria and were reviewed. The assimilation of HIE is a complicated and a multi-staged process. Our findings indicated that HIE assimilation process consisted of four main phases: initiation, organizational adoption decision, implementation and institutionalization. The data helped us recognize the assimilation pattern of HIE in healthcare organizations. Conclusions: The results provide useful theoretical implications for research by defining HIE assimilation pattern. The findings of the study also have practical implications for policy makers. The findings show the importance of raising national awareness of HIE potential benefits, financial incentive programs, use of standard guidelines, implementation of certified technology, technical assistance, training programs and trust between healthcare providers. The study highlights deficiencies in the current policy using the literature and identifies the ‘‘pattern” as an indication for a new policy approach. Ó 2016 Elsevier Inc. All rights reserved.

Contents 1. 2.

3.

4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Eligibility criteria. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Screening and classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Selection of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Characteristics of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Main findings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Literature synthesis: HIE assimilation phases and assimilation pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. First category: policies to support healthcare organizations in the initiation phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Second category: policies to support healthcare organizations in the adoption decision phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

⇑ Corresponding author at: Taylor’s Business School, Taylor’s University Lakeside Campus, Malaysia. E-mail addresses: [email protected] (P. Esmaeilzadeh), [email protected] (M. Sambasivan). http://dx.doi.org/10.1016/j.jbi.2016.09.011 1532-0464/Ó 2016 Elsevier Inc. All rights reserved.

75 75 75 75 76 76 76 76 77 77 82 82 83

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

5. 6.

4.3. Third category: policies to support healthcare organizations in the implementation phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Fourth category: policies to support healthcare organizations in the institutionalization phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Health Information Exchange (HIE) is the electronic transfer of patient data and health information between healthcare providers. HIE, as a part of health care restructuring program, can be used to facilitate sharing of medical information between healthcare service providers. Evidence shows that the interest among healthcare service providers to share through HIE system is growing rapidly [1]. A few countries (England, Netherlands, Finland and the USA) are developing and advancing their regional and national HIE initiatives [2]. Sharing clinical data can potentially improve patient safety, care coordination, quality of care and efficiency [3], facilitate public health efforts [4] and reduce mortality and healthcare costs [5]. Multiple models of clinical data exchange are being used nationwide. The direct project model automates point-to-point processes in which a provider sends patient data to a known recipient [6]. In a non-directed exchange model, a central organization is considered as a hub that provides a lookup for providers [7]. In a query-based HIE, patient data are aggregated from multiple healthcare institutions [8]. Another model is patient-centered exchange in which patient data and laboratory results are delivered to the patient to share as required [7,9]. Consistent with the goals of the federal ‘‘Meaningful Use” program, the exchange of electronic data has been promoted among healthcare providers and institutions [10]. Although the financial incentives offered by the federal government are very encouraging, healthcare providers have not yet adopted and used HIE [11]. Evidence shows that healthcare providers are not likely to simply adopt HIE just due to healthcare cost reduction motivation or mandated adoption programs that support HIE initiatives [5]. More studies are needed to better define different phases which constitute HIE assimilation and more supportive policies are required to facilitate this process [12]. The success of both HIE adoption and implementation depends on factors beyond technical issues [13]. Other influential factors such as organizational, operational and social contexts that are relevant to the HIE adoption and implementation should be studied [11]. Previous studies mainly focus on the individual user and network levels of analysis and a few studies explain use of HIE at the organizational level of analysis [5]. The success of a HIE project and its potential benefits are not likely to be achieved without considering HIE at organizational level and examining organizational factors [3,10]. Therefore, organizational factors such as organizational value of HIE, organizational characteristics, organizational awareness and commitment, organizational adoption and implementation strategies, barriers to organizational adoption decision, resource allocation, organizational support, technical support, and training should be more critically highlighted. Evidence shows that adoption efforts and implementation processes are different [11] and diverse determinants affect adoption and implementation phases [14]. Assimilation of HIE is a multifaceted process influenced by a series of interrelated phases rather than a single unified process [15]. Most of the HIE literature has analyzed HIE adoption as a single step regardless of the interconnected processes of investment, implementation, and institutionalization. Politi et al. [15] discuss that there are a limited number of studies that have analyzed patterns of HIE assimilation. The other

75

84 84 85 85 86

gap in the literature is the similarity of previous studies in characterizing HIE assimilation solely based on patterns such as no use, basic use and advanced use [16]. According to Rebuge and Ferreira [17], healthcare practice and the HIE system have a complex and variable nature. Frisse and Holmes [18] state that the benefits of HIE can only be reaped if it is well implemented and integrated into clinicians’ workflow. Deficits in the exchange of health information such as sharing incomplete information may result in a doctor’s delay in identifying health problems and also in diagnosing a wrong care planning that finally leads to injury or death [19,20]. A number of variables such as current policies, technical issues, market conditions, and hospital characteristics may still block hospitals from participating in HIE [21]. Two of the key words in the informatics literature are ‘‘adoption” and ‘‘assimilation” and they are used interchangeably. As noted by Brierley [22], there has been no general consensus on the definition of adoption. Since IT adoption is a complex and stage-based process, recent studies have described adoption process in terms of assimilation phases [23]. A large number of studies has been conducted in various settings to explain technology adoption process with regard to sequence of phases [24]. There is a dearth of research that presents a complete model to show HIE assimilation pattern in healthcare settings. It is argued that an aggregated measure resulting from assimilation phases can better explain adoption process [25]. Based on this reasoning, in this paper we use ‘‘assimilation” phases instead of ‘‘adoption” process to better articulate the complex nature of HIE and all related factors affecting HIE at various levels of analysis. This study attempts to review the existing literature to define HIE assimilation phases. This research is aimed at recognizing the pattern of HIE assimilation by proposing a new classification comprising of evolutionary phases. We also categorize healthcare organizations based on their strategic decisions to assimilate HIE. Better understanding of assimilation phases and pattern can help policy makers recognize the reasons why so many healthcare providers have failed to fully integrate HIE into their day-to-day practices. 2. Methods 2.1. Eligibility criteria We considered existing theoretical and empirical studies related to HIE assimilation process in various healthcare settings. All retrieved studies published in the refereed journals from the year 2005 and in English language, were included in the review. We limited our search to the last 11 years since we observed that many studies from 2005 onwards used the concept of HIE and interoperability and sufficiently discussed issues related to HIE assimilation process. Studies that were editorials, commentaries, opinion papers or articles without an abstract were excluded from further consideration. 2.2. Search strategy The aim of this study was to undertake a literature review of existing studies relating to HIE assimilation process. To identify

76

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

the right set of key words and databases, the authors used the help of a health librarian to complete a review of the HIE literature. To meet the objective, the studies were mainly searched in four electronic research databases of PubMed, Web of Science, CINAHL, and Academic Search Premiere. The main keywords used for searching articles were ‘‘Health Information Exchange”, ‘‘HIE initiatives”, ‘‘HIE adoption”, ‘‘HIE implementation”, ‘‘HIE usage”, and ‘‘HIE assimilation”. We continued searching until no new studies were found in light of the selection criteria. The journals that were studied include: JAMIA, Journal of Medical Systems, International Journal of Medical Informatics, BMC Journal of Medical Informatics and Decision Making, Journal of Biomedical Informatics, Applied Clinical Informatics, Healthcare, Information Systems, Social Science & Medicine, Health Affairs, New England Journal of Medicine, and JAMA.

2.4. Selection of studies Through database searching, 254 articles published during and after 2005 were retrieved. There were 131 duplicates and nonEnglish articles that were removed resulting in 123 articles. The titles and abstracts of these 123 papers were screened and 46 papers were excluded based on the initial exclusion criteria (no or not relevant abstracts and not relevant settings). The selected papers (77 studies) were reviewed in full and assessed for eligibility. To obtain the final set of papers, 33 papers were further excluded with reasons such as having irrelevant focus of study and offering too-general discussions with no clear theoretical and practical contributions. Finally, 44 papers were used in a qualitative synthesis and a summary of included papers are indicated in Table 2. Fig. 1 depicts the study selection flow process.

2.3. Screening and classification 3. Results Initially, one author reviewed all the retrieved article titles and abstracts to exclude the manuscripts that did not meet the inclusion characteristics. Then, articles that directly discussed issues associated with HIE and were consistent with the aim of the review, were read by both authors in their entirety to be included in conducting this review. We recognized that all included articles could be grouped into four main categories that emerged from the literature review. Therefore, to better screen the manuscripts, the included articles were categorized into four main phases based on their abstracts. The four phases were: HIE initiation efforts, HIE adoption decision, HIE implementation and HIE institutionalization. The factors affecting each phase were extracted from the articles and were transferred into a Microsoft Office Excel spreadsheet. This method allowed the authors to better analyze and describe HIE assimilation phases that were retrieved from the reported studies.

3.1. Characteristics of studies Forty-four studies published between 2005 (January 2005) and 2016 (April 2016) met the inclusion criteria. The main focus of seven articles (16%) directly referred to HIE initiation efforts and awareness. Overall, six articles (13.5%) focused mainly on HIE adoption decision phase including barriers and facilitators affecting healthcare organizations to make an organizational adoption decision. Six articles (13.5%) were mostly related to HIE implementation phase including workflow, set-up plan and installation. A total of 25 studies (almost 57%) primarily addressed issues regarding actual usage and institutionalization of HIE in healthcare organizations. This result provides two important implications. First, the majority of the existing studies focused on only one phase of HIE assimilation phase which is institutionalization. Second, it

Fig. 1. Selection of studies for the review.

Fontaine et al. (2010) [53]

USA

Systematic literature review

N/A

Lack of trust among hospitals and between hospitals and other healthcare organizations is a result of fear which reflects if sensitive patient data and clinical information are shared it will give competitors a competitive advantage. (HIE institutionalization) The main HIE benefits are improved access to test results and other data from outside the practice and decreased staff time for handling referrals and claims processing. Barriers are cost, privacy and liability concerns, organizational characteristics, and technical barriers. A positive return on investment has not been documented. (HIE initiation efforts and institutionalization) HIE stakeholders Conceptual 7

8

USA

Kaelber and Bates (2007) [20] Shapiro et al. (2007) [36] Kern and Kaushal (2007) [48] Grossman et al. (2008) [57] 4

6

Marchibroda (2007) [2] 3

5

Grossman (2006) [40] 2

USA

The HEAL NY grantees based on HEAL NY program

Emergency physicians

Quantitative analysis (questionnaire) Conceptual

Method

Walker et al. (2005) [3] 1

Country Reference #

Table 2 Review of included studies.

As stated by Politi et al. [15], HIE initiatives depend on a set of interrelated variables at different levels (such as strategy, policies, technology, network, user participation, operational and organizational leadership) which work together to improve healthcare delivery. To synthesize the previous literature, the proposed assimilation phases encompass studies conducted at all levels of analysis. Since the concept of HIE is multidimensional, there is a need for an insightful review of literature in order to define assimilation process that reflects complex nature of HIE. According to Thompson [26], assimilation process includes progressive sequences from initiation to development, feedback, and adjustment of technology to become an integral part of an organization workflow. Another research describes the assimilation process through four stages of evaluation, initiation, implementation and routinization [27]. In the IS literature, assimilation stages are mainly described as pre-adoption, adoption-decision and postadoption [28]. Rai et al. [29] define a technology assimilation life cycle that consists of seven stages: awareness, interest, evaluation, commitment, limited deployment, partial deployment and general deployment. Gallivan [30] explains assimilation based on three phases: authority adoption decision, organizational assimilation process, and organizational acceptance as well as its consequences. According to Avgar et al. [31], Healthcare Information Technology (HIT) assimilation consists of three evolutionary stages: investment, implementation and actual usage. Literature has identified unitary-sequence and multiplesequence as the two patterns for innovation assimilation process

Targeted population

3.3. Literature synthesis: HIE assimilation phases and assimilation pattern

USA

Salient points and related phase(s)

In order to reflect a synthesis of the 44 selected papers and explain how these papers lead to reported results and discussions, the extracted information from review of included articles are presented in this section. Table 2 provides an overview of the authors, publication year, method, targeted population, and key points of all the 44 articles. This summary highlights the main points of each study relating to the four phases identified by authors which are HIE initiation efforts, HIE adoption decision, HIE implementation and HIE institutionalization.

Conceptual

3.2. Main findings

USA

shows the importance of actual usage of HIE resulting from the full participation of healthcare providers in exchange activities. The majority of the reported studies were conducted in the United States (86%), 7% in European countries (such as Finland, Israel and Switzerland), 5% in Canada and the remaining 2% in the Asian context (South Korea). Table 1 shows the methodology used in the included studies. Twenty-three studies applied a qualitative study design ranging from literature reviews to conceptual papers. Twenty papers described their results using a quantitative study design.

The most important challenges facing HIE initiatives and organizations is measuring the value of services that result from the HIE to various stakeholders groups (such as providers, payers, and employers). Then, those value assessments should be indicated in business plans which promote and assure sustainability for these initiatives. (HIE initiation efforts) One of the most important benefits of HIE is improved patient safety. Up to 18% of the patient safety errors and as many as 70% of adverse drug events could be eliminated if the right information about the right patient is available at the right time. (HIE initiation efforts) Most respondents were not aware of HIE prior to this study. 85% of respondents reported it was difficult or very difficult to obtain external data, taking an average of 66 min, 72% said that their attempts fail half of the time. (HIE initiation efforts) The programs to facilitate interoperable health IT should specifically address financial sustainability, technical architecture, and practice transformation. (HIE adoption decision)

20 44

State, regional and community-based HIE organizations Patients

Survey Total

USA

Quantitative

Hospital chief information officers, or CIOs

10 4 7 1 2

Qualitative analysis (semi-structured interviews) Conceptual

Conceptual paper Literature review Interview Case study Direct observation and follow-up Interviews

USA

Qualitative

The returns of a fully standardized HIE and interoperability could be a net value of $77.8 billion per year once fully implemented. On the other hand, non-standardized HIE offers smaller positive financial returns. The clinical impact and organizational value of HIE would likely add further value. This offers a compelling business case for national implementation of fully standardized HIE. (HIE implementation) Provider and health plan competition and adversarial relationships between providers and plans are viewed as major barriers to community-wide clinical data sharing. (HIE institutionalization)

Number of studies

HIE organizations

Method

Qualitative analysis (expert interviews)

Study design

USA

Table 1 Study designs of included studies.

(continued on next page)

77

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

78

Table 2 (continued) Reference

Country

Method

Targeted population

Salient points and related phase(s)

9

Sicotte and Paré (2010) [39]

Canada

HIE implementation team members and users

10

Ross et al. (2010) [51]

USA

Qualitative analysis (semi-structured interviews) Case study approach

11

Vest (2010) [11]

USA

12

Vest and Gamm (2010) [62]

13

HIE implementation needs to be a fundamentally multidimensional process. There are numerous factors, such as cultural, financial, technical, political or organizational factors, affecting the change process resulting from HIE projects. (HIE implementation) The greatest facilitator of HIE project is technical assistance and support during and after implementation. Building workflows and information systems to be integrated with the HIE services is another barrier to HIE implementation. (HIE implementation and institutionalization) Adoption and implementation of HIE are different phenomena, thus, many factors associated with an adoption, are not related to implementation and vice versa. Only network membership increases the chance of HIE implementation, whereas competition between hospitals reduces this chance significantly. (HIE institutionalization) Barriers and challenges related to HIE are beyond technology. Technological progress does not automatically fix the problems in healthcare information sharing. HIE requires collaboration among competitors and the healthcare industry has difficulty with this prospect. Competition between providers negatively affect centralized data repositories and network approach to data exchange. Long-term financial uncertainties pose enough risk to upset even the most technologically advanced effort. (HIE initiation efforts and HIE institutionalization) What conceptualizations of usage will best reflect the objectives of HIE? The current literature on HIE should develop a robust measurement of HIE usage under Meaningful Use requirements. (HIE institutionalization)

Small-to-medium sized primary care practices Hospitals

USA

Quantitative analysis (The HIMSS Analytics Database and The AHA survey) Conceptual

Vest and Jasperson (2010) [52] Buntin et al. (2010) [60]

USA

Literature review

N/A

USA

Conceptual

N/A

15

Patel et al. (2011) [45]

USA

Quantitative study (survey)

Physicians

16

Rudin et al. (2011) [9]

USA

Qualitative analysis (Interviews)

Clinician -users and HIE staff

17

Vest et al. (2011) [19]

USA

Quantitative analysis (patient dataset)

Emergency Department encounters among patients less than 18 years old

18

Korst et al. (2011) [38] Dimitropoulos et al. (2011) [41] Adler-Milstein et al. (2011) [56] Unertl et al. (2012) [1]

USA

Quantitative analysis (questionnaire) Quantitative analysis (telephone survey

Hospitals

14

19

20

21

USA

HIE stakeholders

English-speaking adults

HIE initiatives (if fully used) can create essential foundation for restructuring health care delivery and for achieving the key goals of improving health care quality; reducing costs; and increasing access through better methods of storing, analyzing, and sharing health information. (HIE institutionalization) Potential barriers to adopting or using HIE included start-up costs and resources to select and implement a system. A majority reported that technical assistance and financial incentives to use or purchase health IT systems would positively influence their adoption and use of HIE. (HIE adoption decision) Clinicians are motivated to access the HIE by perceived improvements in care quality and time savings, but their motivation can be affected by some factors such as gaps in data, workflow issues and usability issues. Data contribution to the HIE also varies by billing concerns and time constraints. Clinicians, EHR and HIE product vendors and trainers should work toward integrating HIE into clinical workflows. (HIE institutionalization) Based on the patterns of HIE system screens accessed by users, each encounter is classified as: no system usage, basic system usage, or novel system usage The pattern shows that the odds of basic system usage were lower in the face of time constraints and for patients who had not been to that location in the previous 12 months. (HIE institutionalization) Organizational readiness and leadership are important factors affecting organizational participation in health information exchange and achievement of successful inter-organizational collaboration. (HIE institutionalization) Around 70% of respondents were concerned about HIE privacy, 75% were concerned about HIE security. Addressing the specific privacy and security concerns of individuals will be critical to ensuring widespread consumer participation in HIE. (HIE institutionalization) Only 10.7% of US hospitals engaged in HIE with unaffiliated providers. Nonprofit hospitals or those with a larger market share are more likely to participate in exchange activity. (HIE institutionalization)

USA

Quantitative analysis (survey)

Hospitals

USA

Qualitative analysis (direct observation, informal interviews) Conceptual

Hospitals and ambulatory clinics

The success of both HIE adoption and implementation is a function of factors beyond technology. (HIE adoption decision and implementation)

The Office of the National Coordinator for Health Information Technology,

The high cost of exchange is due to the lack of widely adopted standards, failure to use existing standards, and flexibility in the way that standards are implemented. ‘Meaningful Use’ requirements covers critical aspects of HIE, including sharing important information with other providers and patients and reporting quality information and public health results. (HIE implementation and institutionalization) The results indicate the organizational commitment to engage in HIE does not guarantee the usage of the information systems Usage of the information systems that are the result of HIE efforts is frequently very low. These findings recall for the development, operation and evaluation of HIE efforts. (HIE institutionalization) As more organizations share data with one another on a point-to-point basis, measuring the marginal contribution of each external data source and the overall value of HIE will become even more debatable. The financial impact of HIE on emergency department care shows that HIE access is associated with an annual cost savings of $1.9 million. Net of annual operating costs, HIE access reduces overall costs by $1.07 million. Hospital admission reductions is 97.6% of total cost reductions. (HIE institutionalization)

22

Williams et al. (2012) [6]

USA

23

Vest et al. (2012) [16]

USA

Quantitative analysis (patient dataset)

Patients in primary care clinics

24

Frisse et al. (2012) [4]

USA

Quantitative analysis (Tennessee Hospital Association database)

Emergency departments

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

#

Table 2 (continued) Reference

Country

Method

Targeted population

Salient points and related phase(s)

25

Ancker et al. (2012) [7]

USA

Quantitative analysis (telephone survey)

Adult New York State residents

26

Campion Jr. et al. (2012) [8]

USA

27

Avgar et al. (2012) [31]

USA

Quantitative analysis (a cross-sectional study in three communities) Conceptual

Non-clinical staff in outpatient settings, and inpatient physicians HIT stakeholders

28

Adler-Milstein and Jha (2012) [58] Ozkaynak and Brennan (2013) [61] Furukawa (2013) [55]

USA

Conceptual

HIE stakeholders

USA

Qualitative analysis (direct observation and short interviews) Quantitative analysis (national survey)

Clinicians

Consumers are supportive of different architecture of HIE Policy and outreach pertaining to HIE may be most effective if it clarifies the roles and responsibilities of large businesses involved in different aspects of the exchange, and privacy and security controls. (HIE institutionalization) The important feature of query-based HIE systems is patient summary data displayed by default. User role, practice site type, and patient consent workflow may affect patterns of query-based HIE web portal system usage. (HIE institutionalization) Three central, evolutionary stages related to the health IT adoption process: 1. The decision to invest in health IT, 2. The implementation or deployment process, and 3. Institutionalization of the technology. Different organizational barriers emerge over different stages of the health IT adoption process. (HIE initiation efforts, implementation and institutionalization) Physicians and hospitals are concerned about the competitive advantages of sharing their patient data, which may make it easier for patients to seek care from rival institutions. Hospitals see clinical data as ‘‘a key strategic asset” that tie physicians and patients to their organization. (HIE institutionalization) The overall usage rate of the HIE is low. Physicians use the HIE system for patients with specific characteristics (typically adults with chronic (flank, back, leg, etc.) pain as their chief complaints). In order to fully benefit from HIE, understanding organizational and social context during the HIE design and implementation is required. (HIE institutionalization) In 2012, 51% of hospitals exchanged clinical information with unaffiliated ambulatory care providers, but only 36% exchanged information with other hospitals outside the organization. Furthermore, more than half of hospitals exchanged laboratory results or radiology reports, but only about one-third of them exchanged clinical care summaries or medication lists with outside providers. (HIE institutionalization) Although most quickly actors agree on the goals and potential benefits of the HIE system, there are many obstacles and mostly non-technical. The transparency coming from a streamlined exchange of information may improve the continuity, quality and efficiency of care. This transparency can also reveal and challenge habits and practices of care professionals and of citizens. This tension must be removed, and trust must be fostered amongst stakeholders. (HIE institutionalization) HIE is a difficult undertaking due to political and economic reasons. Policies will have to address the shortcomings of HIE models to ensure information is effectively shared between providers to maximize health benefits. (HIE adoption decision) Just 10% of respondents indicated that their organizations were formally engaged in HIE activities, and 49% were unaware of organizational involvement in HIE. Respondents expressed a desire for better decision support, paperless reporting methods, and situational awareness of community outbreaks. (HIE initiation efforts) Adoption of HIE is a multifaceted process and focuses on a collection of interrelated stages rather than on a single unified step. Most of the HIE literature have analyzed HIE adoption as a single step regardless of the interconnected processes of investment, implementation and institutionalization. There are limited number of studies that analyzed patterns of HIE adoption. (HIE adoption decision) Patient value (Coordination of care, patient portal, duplicate testing, and population health interventions), ‘Meaningful Use’ program (HIE as a means to obtain ‘Meaningful Use’) and reduced expense relative to alternatives are facilitator to HIE adoption decision. (HIE adoption decision) The role of direction setting in the form of technological standard setting, privacy standard setting, and funding is better to be assigned to the federal government in order to facilitate HIE implementation. (HIE implementation and institutionalization) Larger hospital systems are more likely to exchange electronic patient information internally, but are less likely to exchange patient information externally with other hospitals. This pattern is shaped due to a commercial cost to sharing data with other hospitals. This contrast between a willingness to share data internally and a lack of willingness to share data externally reflects a trend for larger hospital systems to create ‘information silos. It is also unclear what the best steps are for policymakers to take to ensure that information exchange happens. (HIE institutionalization) Only 30% of hospitals in U.S. are engaged in actual clinical and health information exchange with other providers outside of their practice. For-profit hospitals are far less likely to engage in HIE than non-profit hospitals. Hospitals with a larger market share are also more likely to engage in exchange. Hospitals in less competitive markets are more likely to practice in data exchange. (HIE institutionalization) The majority of primary care physicians use an electronic regional HIE system rather than paper or fax as a primary means of cross-organizational HIE. Findings show that experience with an integrated regional HIE system is more positive than that with other types or regional HIE systems such as web distribution model or master index model. (HIE institutionalization) Actual Use of an HIE resulted in reduced use of hospital resources, noteworthy cost savings, decreased length of stay, and improved quality of care. (HIE institutionalization)

29

30

USA

Hospitals

31

Geissbuhler (2013) [12]

Switzerland

Conceptual

HIE stakeholders

32

Vest et al. (2013) [21] Dixon et al. (2013) [35]

USA

Qualitative analysis (interviews) Quantitative analysis (Online survey)

National health informatics policy experts Hospital Chief Information Officers

33

USA

34

Politi et al. (2014) [15]

Israel

Quantitative analysis (patient dataset)

Adult patients

35

Yeager et al. (2014) [46]

USA

Qualitative study (indepth interviews)

Health care stakeholders

36

Vest et al. (2014) [50]

USA

Quantitative analysis (interviews)

HIT experts

37

Miller and Tucker (2014) [5]

USA

Quantitative analysis (the American Hospital Association database)

Hospitals

38

Adler-Milstein and Jha (2014) [10]

USA

Hospitals

39

Hyppönen et al. (2014) [13]

Finland

Quantitative analysis (American Hospital Association’s IT Supplement database) Quantitative analysis, web-based questionnaire

40

Carr et al. (2014) [44]

USA

Quantitative study (survey)

Physicians

Clinicians at Emergency department

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

#

(continued on next page) 79

The most commonly cited barriers to HIE use were incomplete information, inefficient workflow, and reports that the exchanged information did not meet the needs of users. Highest HIE use was in sites with proxy users supporting clinicians. (HIE implementation) N/A Literature review USA

N/A

44

43

Park et al. (2015) [43] Parker at al. (2016) [49] Eden et al. (2016) [37] 42

South Korea USA

Quantitative study (patient dataset) Literature review

HIE clinics

Salient points and related phase(s)

Type of medical information which is shared or received has the following characteristics: Diversity, Breadth, and Volume. Identifying providers with a high Impact but low Interoperability score could assist planners and policy-makers to optimize technology investments intended to electronically share patient information across the continuum of care. (HIE institutionalization) The need for fair allocation of benefits and costs among stakeholders was a critical factor for successful adoption of the HIE technology. (HIE initiation efforts) Little generalizable evidence currently exists regarding benefits attributable to HIEs. (HIE initiation efforts)

Method

McMurray et al. (2015) [42]

Country Reference

41

Canada

[32]. The first pattern describes the assimilation process as a linear sequence and the second one assumes that the process is not linear and there are some contingency factors that make the process unpredictable and complex [33]. The contingency variables are categorized into three main groups: adopter attributes, innovation attributes and environmental conditions [34]. In the context of HIE, the effects of healthcare organizations’ characteristics such as organization size, structure, culture, and strategies have been reported. HIE-related factors and attributes such as advantage, cost, incentives, standard requirement, complexity, observability, compatibility and environmental conditions such as uncertainty, market conditions and competition have also been discussed in the literature [11]. Thus, HIE assimilation in healthcare organizations can be studied in light of multiple contingency variables according to multiple-sequence pattern. By synthesizing the literature, we have recognized four main phases of HIE assimilation in healthcare organizations. HIE assimilation defines the process within healthcare organizations stretching from initial awareness of the HIE system, to potentially, formal adoption and full-scale deployment. Therefore, the HIE assimilation embodies four phases of initiation, adoption decision, implementation process and institutionalization. Table 3 provides the dimensions of the four main phases. (a) Initiation phase The first phase is initiation that embodies awareness and attitude formation. According to Dixon et al. [35], the HIE assimilation begins with awareness level which discusses whether the healthcare institutions and potential users are aware of the HIE services or not. The frozen zone is where the awareness level is too little or organizations are not aware of HIE models. Several researchers describe that due to lack of awareness, no adoption and implementation takes place [2,20,35]. Therefore, we consider awareness as the main foundation for HIE assimilation process. Shapiro et al. [36] argue that the viewpoints of key organizational decision makers about HIE can change when they are exposed to the potential benefits and gains of HIE initiatives such as quality improvement, cost reduction, and patient care coordination. Once stakeholders become aware of HIE efforts, they may become interested and their attitude to adopt HIE will be formed. At the end of this phase, discussions, investigation and analysis are conducted by organizational decision makers to adopt HIE [21]. HIE assimilation process may stop by the end of this stage after performing a detailed cost-benefit analysis. The first transition point (#1 – Table 3) appears and healthcare organizations that become interested as a result of analysis and those that hold a positive perception about potential benefits of HIE efforts move to the second phase. (b) Adoption decision phase

#

Table 2 (continued)

Designing the HIE ontology

Targeted population

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

N/A

80

At the beginning of this stage, healthcare organizations have no actual plans for adoption and have no strategy for implementation and integration with HIE services. According to Vest et al. [21], the main concern in this phase is the funding issues. Thus, if healthcare organizations are provided with financial incentives they may become motivated (or pushed) to make HIE adoption decision. Eden et al. [37] indicate that the federal incentives, requirements of HITECH Act and organizational readiness can moderate the organizational adoption decision [37]. These factors may encourage healthcare providers to obtain the necessary organizational approvals and required resources to implement HIE. In reality, mandatory adoption cannot simply remove the hindering variables [14]. One possible response to mandatory HIE is resistance to the mandated system by appealing to the legislature. Vest [11] states that even if the adoption decision (phase 2) is mandated, actual usage (phase 4) which results from sharing health information

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

81

Table 3 Dimensions of the main four phases of HIE assimilation.

with other providers outside of the practice cannot be forced. Therefore, as discussed by Korst et al. [38], a group of healthcare providers will move to the next step (transition point #2 – Table 3) by resolving financial, organizational and legal challenges for the implementation of HIE systems and some institutions may not be able to move further due to financial issues, lack of organizational readiness and resource allocation concerns. (c) Implementation phase Once the organizational adoption decision is made, the next phase (implementation process) begins. Several researchers describe that implementation phase includes two dimensions: set-up and execution of implementation plan [3,37]. In the set-up stage, healthcare providers develop a plan for implementation of HIE systems and clearly define the scope and objectives of implementation. According to Sicotte and Paré [39], during this phase, healthcare organizations also predict likely problems, concerns and possible changes required in the organizational structure and process design to facilitate implementation. Then, the developed plan is executed to acquire and install the certified and standard software, hardware, network and systems (such as standard EHR) [6]. At the end of this phase, HIE services are available to potential users in the organizations. Consistent with Frisse et al. [4] and Ancker et al. [7], after the execution step, institutions can go through iterative steps (resource allocation iteration in Fig. 2) and re-allocate required resources (such as providing technical assistance and individual financial incentives) to help healthcare professionals use HIE. Resource allocation may occur based on the needs assessment when more resources are required or some resources appear to be missing in the initial organizational analysis. This implies that institutions do not go through a linear assimilation process [39] and resource allocation occurs multiple times. After implementation and allocating resources to potential users (such as physicians), healthcare providers will move to the next step (transition point #3 – Table 3) which is institutionalization. Implementation phase can also be terminated or become incomplete if the set-up plan fails or implementation process is abandoned. (d) Institutionalization phase The last phase is institutionalization or integration of HIE systems into the healthcare organization’s routine. Several studies indicate that in the integration phase, the actual use of the system begins, unexpected problems are addressed and a common organizational understanding about the system is created [12,19]. Vest et al. [16] state that HIE integration highlights the degree to which healthcare organization is likely to engage in electronic exchange of clinical data inside the organization and with other healthcare

providers. Therefore, healthcare providers might choose to participate partially or fully in exchange efforts with regional or unaffiliated providers or defer HIE integration efforts and actual usage [10]. Health care providers that have implemented the HIE system are expected to actually use the HIE services by integrating the system into their organizational activities. According to Miller and Tucker [5], in reality, healthcare providers face unanticipated problems, technical issues and concerns especially related to exchange activities with other providers and may choose not to participate in both internal and external information exchange. During the institutionalization phase, partial use of HIE may emerge. According to Adler-Milstein and Jha [10], this reflects a condition in which a number of healthcare providers may use HIE to exchange health information with only regional institutions as internal exchange and see no advantage from external data exchange with unaffiliated organizations. Competition is another influential factor that leads to interrupted integration and failed actual use of HIE [40]. As stated by Dimitropoulos et al. [41], privacy and security issues are common challenges cited by many HIE organizations and these concerns prevent them from sharing all patient-related information with other health providers. These concerns should be addressed in the implementation phase by providing better security and privacy guidelines [7,41]. Under partial usage, healthcare providers may only share selective types of clinical and patient-level information with select providers outside of their practice as external exchange [42]. The implemented HIE system will be used to its fullest potential if healthcare providers experience no privacy, security, technical, and competition concerns and also fully admit that exchange activities with other providers (whether regional or unaffiliated) are useful, effective, and practical [10]. This phase shows the importance of trusting relationship and competition-free collaboration among healthcare providers pertaining to exchange activities. Once the HIE system is fully used by healthcare organizations on a national scale, it can be integrated with other organizational systems to improve health care delivery and save healthcare costs [43]. Fig. 2 depicts the HIE assimilation pattern within healthcare organizations according to the four main phases. The pattern recognized from the literature indicates the extent of assimilating HIE by healthcare organizations where assimilation extends from initial awareness to full institutionalization. HIE assimilation pattern is concerned with understanding the pattern a HIE project follows as it spreads across a population of potential adopters over time. Therefore, the vertical axis shows the four assimilation phases and the horizontal axis represents time. The following assimilation pattern shows the dimensions of each phase, HIErelated workflow, course of actions (such as analysis and resource allocation), iterative steps along with HIE assimilation failures which are lack of awareness, rejection of HIE, failed implementation, and no actual usage.

82

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

Fig. 2. HIE assimilation pattern within healthcare organizations.

4. Discussion This study describes the various phases of HIE assimilation in detail and explains the policies that are required to facilitate HIE initiatives. Fig. 2 shows that the assimilation pattern identified in the literature is different from existing theories of IT adoption such as Diffusion of Innovation. The existing theories are very general and cannot be used to articulate HIE assimilation due to the presence of multiple unique contingency variables and issues such as policy interventions that may vary based on the type of organization (for profit and not-for-profit organizations), size (large and small) and market conditions. A new assimilation pattern needs to highlight these unique issues associated with HIE which will be missing if existing theories are used. We need to emphasize that the main audience of this review is HIE policy makers but the results and generated knowledge are useful for all HIE stakeholders such as HIE organizations, leaders of health care institutions, organizational designers and HIE vendors. Our results show that HIE assimilation has been studied too narrowly. HIE assimilation is instead more multifaceted and must include consideration of other factors such as awareness, ROI, incentives, competition, size, organizational structure, training, and type of healthcare organizations at key decision points along the identified assimilation pattern. Policy makers need to phase down HIE initiatives and devise strategies matching each phase.

Better policy can thus be designed around the key decision points along the assimilation pathway if they are better understood. Table 4 depicts assimilation phases along with unique factors affecting each phase and implications for HIE policy makers. We also need to articulate drivers, barriers, healthcare organizations’ decisions and strategies in each phase to better describe HIE assimilation pattern. To reflect implications from the synthesis of this review, we have categorized healthcare providers according to their positions and strategies along the recognized HIE assimilation pattern. The proposed categorization displays five possible paths which can be taken by healthcare organizations. The classification (Fig. 3) that is developed based on the literature review can highlight the deficiencies in the current policies related to each assimilation phase. The following four sections discuss in more detail the four assimilation phases (Table 4) and related policies to support healthcare organizations in each phase. 4.1. First category: policies to support healthcare organizations in the initiation phase Awareness is fundamental for a successful HIE assimilation. Several researchers mention that many healthcare organizations are still not aware of HIE services and benefits [35,36,44]. Thus, there is no chance for the remaining assimilation phases (adoption decision, implementation and institutionalization) to occur. This

Table 4 Assimilation phases, influential factors and implications. Assimilation phases

Influential factors

Implications and recommendations to policy makers

Initiation phase

Awareness, interests, benefits, costs, expected ROI, value added

Adoption decision phase Implementation phase

Funding, resource allocation concerns, type of organizations

Providing more visible benefits (quality, safety, efficiency) and costs of participation in HIE to deliver a transparent cost-benefit model, Providing robust evidence to prove profitability and sustainability of HIE business model, Presenting clinical, financial, organizational value, Holding conferences about HIE to expose organizational decision makers to the benefits of HIE Financial incentive programs, Federal mandates

Institutionalization phase

Potential users’ acceptance, individual incentives, organizational integration, competition, collaboration concerns, organization size

Changes in organizational structure, workflow, and routines. Network execution issues, Installation challenges, lack of standards, privacy and security issues

Providing infrastructures for implementation, Execution support programs, Providing maintenance support, Business process reengineering support, Better data security safeguards and privacy controls, Clarifying widely adopted standards, guidelines and formats Providing technical assistance, Supplying additional individual incentives, Providing training programs for potential users (healthcare professionals), Developing trusting relationship and alliances to encourage interorganizational exchange

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

83

Fig. 3. Strategies of healthcare organizations for the HIE assimilation pattern.

circumstance is shown as the first path in Fig. 3 which is called (0-0-0-0 Path). If healthcare organizations follow this path the result is No awareness (AW = 0), No adoption planning and decision (AD = 0), followed by No implementation (IM = 0) and finally, No actual usage (AU = 0). Integrative marketing programs and more informative HIE conferences should be conducted to raise awareness of various HIE stakeholders. As discussed by Park et al. [43], in this phase, healthcare organizations assess the general value of services that emerge from HIE to various stakeholders such as providers, patients, payers, and employers. Then, they focus on a cost-benefit model to project potential values of HIE. In line with Kaelber and Bates [20], the most critical evaluation questions the impact of health information technology and HIE services on quality, safety, and efficiency of healthcare services. According to Marchibroda [2], HIE faces a long time horizon before what may be termed functionality, profitability, and/or sustainability are achieved. Therefore, achieving ROI takes long for healthcare providers and addressing the clinical, financial, and organizational value of inter-organizational relationships resulting from HIE is a complex task. As a result, there are two categories of healthcare organizations in the initiation phase. The first category will find HIE very beneficial due to the assessment and cost-benefit analysis. Thus, they will move to the second phase which is organizational adoption decision phase. The second category of healthcare organizations will find HIE services costly or not beneficial and they get stuck in the initiation phase with many questionable values until a facilitator (such as financial incentives) yields a convincing ROI. This type of healthcare organization will also move to the second phase. To support healthcare organizations in the first phase, policy makers need to provide more detailed information about possible values (organizational efficiency and effectiveness) and costs of

participation in HIE to promote widespread adoption and use. On the basis of robust evidence and reasoning, healthcare organizations are more likely to recognize the profitability and sustainability of HIE business model. To increase awareness, various conferences about HIE can be held to highlight clinical, financial and organizational benefits and in order to help organizational decision makers make the adoption decision. 4.2. Second category: policies to support healthcare organizations in the adoption decision phase Making an adoption decision is a function of collecting information, allocating resources and devising strategies required to apply a radical organizational change. In this phase, there are two categories of healthcare organizations. According to Patel et al. [45], the first category consists of organizations that are stuck in costbenefit analysis based on their evaluation showing that participation in HIE initiative will have no added value due to doubts about patient value, stakeholder involvement, and sustainability of regional HIE systems. Therefore, these organizations will be disengaged from the assimilation process regardless of the potential benefits of HIE, guidelines, financial incentives and requirements. This circumstance results in the second possible path (1-0-0-0 Path) as shown in Fig. 3. This path implies that healthcare organizations are aware of regional or national HIE services (AW = 1) but costbenefit analysis and value evaluation are not very convincing. Thus, No adoption plan is devised (AD = 0) and No implementation plan is developed (IM = 0) and finally, No actual usage occurs (AU = 0). The second category includes healthcare organizations which have no doubt about the benefits of sharing clinical information with other healthcare providers [45,46]. One of the main benefits

84

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

resulting from HIE is patient value where a patient portal (as a key piece of HIE) provides value to patients by facilitating access to their own medical information and documentation [7]. Therefore, this category of healthcare organizations has no concern about using the HIE to improve quality of the healthcare services as well as care coordination for patients [46]. Consistent with DesRoches [47], the most important barrier in this phase is funding concerns. Start-up costs, technical infrastructure payment and uncertain return of investment are considered as significant variables affecting HIE organizational adoption decision and planning [48]. As stated by Parker et al. [49], the main question that reflects uncertainty during this stage is: how can the potential financial costs and benefits be distributed among stakeholders? Due to financial concerns, for-profit hospitals are less likely to adopt HIE than not-for-profit and public hospitals [10]. Therefore, to help healthcare organizations overcome the barriers in this phase, new policies need to clarify how incentive programs should be designed and allocated and how federal mandates should be enforced to support sustainability of HIE business model. Financial incentives from government agencies and policy makers can relieve monetary concerns of HIE adoption process by funding required infrastructures and organizational facilities as well as subsidizing upgrades of hardware and software [45]. 4.3. Third category: policies to support healthcare organizations in the implementation phase When an organizational adoption decision is made, the healthcare organizations are moving to the next phase which is the implementation. According to Vest et al. [50], in the implementation phase, healthcare organizations execute the set-up plan and make the system available for all potential users. Implementing HIE can refer to providing the infrastructure for the electronic exchange of health information. The most common challenge in this phase is creating an alignment between the current organizational practices with the HIE’s underlying workflow. As stated by Yeager et al. [46], limited IT manpower and competition or mismatch with other organizational projects such as CPOE and EHR are some HIE implementation barriers. Therefore, the main barrier to HIE implementation is technical which reflects the need to incorporate HIE into existing management systems and EHRs. Workflow issues are another important barriers. According to Ross et al. [51], though some practices are open to reengineering workflows with HIE services, most practices instead would like HIE to complement their existing workflows. Concerns about the HIE’s central repository model, specifically its usability, data security, and data quality can be raised in this phase [41]. The data security concerns may be alleviated in case of using other HIE models such as a direct messaging or query model from one EHR to another EHR. To help healthcare organizations in this phase, new policies need to implement robust data security safeguards and privacy controls. In line with Ancker et al. [7], using a patient-centered model of care has been on the rise among healthcare providers to concentrate on patients’ needs. To adhere to the patient-centered model, healthcare organizations need to implement certified EMR systems that are linked to their community’s HIE service. Campion Jr. et al. [8] describe the emergence of workflow barriers from initial difficulties such as automating patient status (opted-in/opted-out) consistent with hospital registration. The cost of building interfaces, the patient portal and analytics are the common concerns mainly to smaller hospitals [46]. Therefore, it is likely that healthcare providers become unsuccessful in the implementation of HIE due to a poor process re-engineering which results in lack of integration between existing workflows and HIE initiatives. New policies should focus on systematic guidelines to provide technical

support for execution and maintenance of certified and standard technology which enables HIE. When the HIE system is installed and becomes ready to use, the next step (institutionalization phase) begins. 4.4. Fourth category: policies to support healthcare organizations in the institutionalization phase Investment in HIE and also the implementation phase cannot guarantee that its full potential will be utilized. According to Unertl et al. [1], healthcare organizations should be encouraged to reap the potential benefits of HIE by participating in exchange activities, improving care coordination and learning from other organizations. In the institutionalization phase, healthcare providers can be divided into three categories based on their strategies about exchange activities. The first category consists of the healthcare organizations that have acquired the HIE system but have not yet used it formally according to standards. This category can be considered as a good example to prove that adoption and implementation of HIE (may be due to HITECH Act and the incentive program) do not necessarily translate to successful usage [11]. As stated by Ross et al. [51], organizations that have not developed a technical support infrastructure and training cannot gain the benefits of HIE. This category of healthcare providers prefer not to engage in HIE due to the unsolved technical and non-technical concerns [10]. Thus, to help healthcare organizations bypass these issues in this phase, new policies should provide technical assistance, provide additional individual incentives, and offer training programs for healthcare professionals. In line with Vest and Jasperson [52], using information systems occurs at various levels of analysis such as individual level when an individual physician or nurse is seen as the end user when considering HIE. According to Fontaine [53], perspectives and attitudes of physicians and nurses toward participation in exchange activities and developing HIE are still not well studied. One of the main challenges affecting the institutionalization phase is lack of resource and technical assistance [51]. Physicians and staff of healthcare institutions may not be fully aware of HIE technical details to solve technical issues. Besides, HIE software which is developed by vendors may have various guidelines, standards, and features that support interoperability but also hinder healthcare organizations from actual usage of the system [54]. Therefore, physicians are not willing to spend the extra time using records through the HIE without having access to some resources such as technical assistance, additional training or individual incentives (resource allocation iteration). Nontechnical concerns refer to market conditions which are characterized by market competition and lack of critical mass within a market [46]. Under this circumstance, organizational decision makers have concerns about losing their market share by sharing patient information with other organizations. On the other hand, lack of sufficient participation in the HIE within market networks may lead to the belief that using the HIE for decision-making will not be useful. Therefore, this category of healthcare providers assesses little value in using the HIE when a critical mass of providers are not inputting patient information into the HIE. As shown in Fig. 3, this circumstance indicates the third path which is called (1-1-1-0 Path). This path implies that healthcare organizations are aware of HIE added values (AW = 1), organizational adoption planning is made (AD = 1), the implementation and installation are completed (IM = 1) but finally, the HIE system remains unused (AU = 0). New policies should design additional individual incentives to encourage healthcare professionals to use an implemented HIE. Hospitals’ levels of engagement in exchange activity vary by the type and organizational affiliation of the provider with which the information is exchanged and also type of clinical information

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

exchanged [55]. According to Adler-Milstein et al. [56], only a small minority of US hospitals participate in electronic exchange of clinical data with unaffiliated providers. The stakeholder willingness to engage in the exchanges is affected by uncertainty about who benefits from HIE [57]. Therefore, in line with Adler-Milstein and Jha [58], the second category consists of healthcare organizations that have implemented HIE system and to some extent agree that information exchanging can improve workflow efficiency and quality of care delivery. This category also voices concerns about losing patients due to exchange activities and becomes less likely to contribute patient data to the HIE [10]. Thus, the healthcare organizations choose to partially engage in exchange activities by restricting exchange of patient data. Despite the availability of nationally recognized standards [59], most of the organizations do not exchange clinical information electronically with other unaffiliated providers. According to Miller and Tucker [5], larger hospitals (which are a part of a larger system) are more likely to exchange electronic patient information internally. However, they become less willing to do external information exchange with other hospitals due to the risk of losing their patients. Developing cooperative relationships and alliances between healthcare providers as well as growing networks to obtain optimal patient care may remove the negative effects of competition and lack of trust between healthcare organizations and facilitate the institutionalization process [11]. Consistent with Grossman et al. [57], hospitals operating in more competitive markets and those with a smaller market share are less engaged in HIE due to the notion that the potential loss of patients that may be facilitated by HIE outweigh the potential benefits. As shown in Fig. 3, this circumstance indicates the fourth path which is called (1-1-1-0.5 Path). This path implies that healthcare organizations are aware of HIE potential gains (AW = 1), organizational adoption decision is made (AD = 1), the implementation phase is completed (IM = 1) and finally, the HIE system is partially used (AU = 0.5). This suggests that stronger policy efforts are required to realize the vision of nationwide data exchange to remove information silos. The third category of healthcare providers in the institutionalization phase are those organizations that have fully used the HIE. According to Buntin et al. [60], these health care providers completely support national HIE policies and are fully convinced that electronic exchange of patient data with other healthcare organizations directly contribute to a high-performing healthcare system. Therefore, they actively engage in a regional HIE effort and external information exchange with independent providers. Furthermore, some organizations which previously selected to use HIE partially may go back and reallocate required resources to shift to a full contribution (resource reallocation iteration – Fig. 2) [10,44]. As depicted in Fig. 3, this circumstance shows the fifth possible path which is called (1-1-1-1 Path). This path implies that healthcare organizations are completely aware of regional and national HIE services (AW = 1), cost-benefit analysis and value evaluation are very convincing, and organizational decision is made (AD = 1), the setup plan and implementation is completed (IM = 1) and finally, the HIE system is fully used (AU = 1). This is the desirable path which leads to optimal healthcare system. New policies should highlight this path to healthcare organizations and remove the barriers in each phase to convince healthcare professionals to fully participate in exchange activities. In line with the three categories of health care organizations in the institutionalization phase, evidence shows that the majority of U.S. hospitals have not yet participated in HIE [10,61]. As stated by Vest and Gamm [62], failure to act on HIE initiatives to improve healthcare efficiency and effectiveness can almost cause decline in competitiveness of healthcare services. This implies that the current policies have not been successful in removing the barriers and reinforcing HIE assimilation and stronger policies are needed.

85

5. Limitations A number of limitations of this study need to be acknowledged. First, caution needs to be taken in interpreting the findings due to methods of selecting included studies. We searched only for articles in English and additional eligible studies might have existed in other languages. Other inclusion criteria and keywords chosen for conducting this review also limited the inclusion of all potentially relevant research. Moreover, only four main electronic research databases were searched for reporting articles. Thus, using other databases may have included additional studies and broader reflection of the current literature. Second, gray literature such as secondary analyses, reports, and dissertations were not reported in this study and only peer-reviewed publications were included. Therefore, there is a potential for introducing possible publication bias. Finally, the heterogeneity among study designs, methods, settings and purposes of reported studies made the synthesis of the reported literature difficult. Various settings of HIE and diverse factors affecting assimilation process in different contexts limited our ability to generalize the findings. Nevertheless, our results can be a reasonable representation of the current literature and serve as an important reference for HIE policymakers.

6. Conclusions Potential cost saving, quality improvements and patient care coordination as a result of HIE will be visible when healthcare providers share electronic patient data with other providers. HIE assimilation is more than organizational adoption decisions and it is beyond mere installation of required hardware, software and infrastructures. Although some healthcare organizations have simply acquired a HIE system, they may decide not to fully engage in sharing clinical information with other healthcare providers outside of their practice. Exchange of patient information is ensured when an implemented HIE system is fully used by providers to gain its potential benefits. Low adoption and utilization of HIE indicate the continued existence of barriers that current policies may be failing to address. The findings contribute a number of theoretical and practical implications for researchers and policy makers to better analyze the HIE assimilation process and pattern in order to facilitate assimilation of HIE projects. A collective knowledge base that includes the network, organization, team, and individual level of analysis will result in a complete understanding of HIE use. Motivators, barriers, potential facilitators and pattern of HIE assimilation should be further supported by empirical data in future studies. Since majority of hospitals are yet to fully engage in HIE, stronger policies are required to ensure exchange of health information within affiliated organizations and between unaffiliated providers. This review serves as a helpful means of highlighting deficiencies in current policy using the literature and the identified ‘‘pattern” as evidence for a new policy approach. The paper focuses on the phases and pattern of assimilation to develop unique insights for policy makers. Current policies related to HIE adoption mostly focus on technical barriers (such as lack of standardized structure and format of exchange activities or lack of resources for implementation). There are many other non-technical variables which may prevent providers from participating in HIE (such as lack of cooperation and trust between providers, lack of incentives for institutionalization and lack of legal framework at a strategic level). Forming alliance between healthcare providers and developing a trusting relationship between healthcare providers can convince healthcare providers that medical records and health information are not a part of their property and they need to exchange data with others to gain collective benefits associated

86

P. Esmaeilzadeh, M. Sambasivan / Journal of Biomedical Informatics 64 (2016) 74–86

with HIE. Policy makers need to go beyond the technical aspects to develop a more analytical model of HIE assimilation.

References [1] K.M. Unertl, K.B. Johnson, N.M. Lorenzi, Health information exchange technology on the front lines of healthcare: workflow factors and patterns of use, J. Am. Med. Inform. Assoc. 19 (3) (2012) 392–400. [2] J.M. Marchibroda, Health information exchange policy and evaluation, J. Biomed. Inform. 40 (6) (2007) 11–16. [3] J. Walker, E. Pan, D. Johnston, J. Adler-Milstein, D.W. Bates, B. Middleton, The value of health care information exchange and interoperability, Health Aff. 24 (2005) 10–18. [4] M.E. Frisse, K.B. Johnson, H. Nian, C.L. Davison, C.S. Gadd, K.M. Unertl, et al., The financial impact of health information exchange on emergency department care, J. Am. Med. Inform. Assoc. 19 (3) (2012) 328–333. [5] A.R. Miller, C.E. Tucker, Health information exchange, system size and information silos, J. Health Econ. 33 (2014) 28–42. [6] C. Williams, F. Mostashari, K. Mertz, E. Hogin, P. Atwal, From the Office of the National Coordinator: the strategy for advancing the exchange of health information, Health Aff. 31 (3) (2012) 527–536. [7] J.S. Ancker, A.M. Edwards, M.C. Miller, R. Kaushal, Consumer perceptions of electronic health information exchange, Am. J. Prev. Med. 43 (1) (2012) 76–80. [8] T.R. Campion Jr., A.M. Edwards, S.B. Johnson, R. Kaushal, Health information exchange system usage patterns in three communities: practice sites, users, patients, and data, Int. J. Med. Inform. 82 (9) (2013) 810–820. [9] R. Rudin, L. Volk, S. Simon, D. Bates, What affects clinicians’ usage of health information exchange, Appl. Clin. Inform. 2 (3) (2011) 250–262. [10] J. Adler-Milstein, A.K. Jha, Health information exchange among U.S. hospitals: who’s in, who’s out, and why?, Healthcare 1 (2) (2014) 26–32 [11] J.R. Vest, More than just a question of technology: factors related to hospitals’ adoption and implementation of health information exchange, Int. J. Med. Inform. 79 (12) (2010) 797–806. [12] A. Geissbuhler, Lessons learned implementing a regional health information exchange in Geneva as a pilot for the Swiss national eHealth strategy, Int. J. Med. Inform. 82 (5) (2013) 118–124. [13] H. Hyppönen, J. Reponen, T. Lääveri, J. Kaipio, User experiences with different regional health information exchange systems in Finland, Int. J. Med. Inform. 83 (1) (2014) 1–18. [14] E. Rogers, Diffusion of Innovations, fourth ed., Free Press, New York, 1995. [15] L. Politi, S. Codish, I. Sagy, L. Fink, Use patterns of health information exchange through a multidimensional lens: conceptual framework and empirical validation, J. Biomed. Inform. 52 (2014) 212–221. [16] J.R. Vest, L.D. Gamm, R.L. Ohsfeldt, H. Zhao, J. Jasperson, Factors associated with health information exchange system usage in a safety-net ambulatory care clinic setting, J. Med. Syst. 36 (4) (2011) 2455–2461. [17] A. Rebuge, D.R. Ferreira, Business process analysis in healthcare environments: a methodology based on process mining, Inform. Syst. 37 (2) (2012) 99–116. [18] M.E. Frisse, R.L. Holmes, Estimated financial savings associated with health information exchange and ambulatory care referral, J. Biomed. Inform. 40 (6) (2007) 27–32. [19] J.R. Vest, J. Jasperson, H. Zhao, L.D. Gamm, R. Ohsfeldt, Use of a health information exchange system in the emergency care of children, BMC Med. Inform. Decis. Mak. 11 (1) (2011) 78. [20] D.C. Kaelber, D.W. Bates, Health information exchange and patient safety, J. Biomed. Inform. 40 (6) (2007) 40–45. [21] J.R. Vest, T.R. Campion Jr., R. Kaushal, et al., Challenges, alternatives, and paths to sustainability for health information exchange efforts, J. Med. Syst. 37 (6) (2013) 1–8. [22] J.A. Brierley, Why the proper definition of the ABC matters: a note, Adv. Manage. Account. 19 (2011) 225–249. [23] M.A. Hameed, S. Counsell, S. Swift, A conceptual model for the process of IT innovation adoption in organizations, J. Eng. Technol. Manage. 29 (3) (2012) 358–390. [24] M. Al-Sayed, D. Dugdale, Activity-based innovations in the UK manufacturing sector: extent, adoption process patterns and contingency factors, Br. Account. Rev. (2015). [25] R.G. Fichman, The role of aggregation in the measurement of IT related organizational innovation, MIS Quart. 25 (4) (2001) 427–455. [26] V.A. Thompson, Bureaucracy and innovation, Adm. Sci. Quart. 10 (1) (1965) 1–20. [27] J. Hage, M. Aiken, Social Change in Complex Organizations, Random House, New York, 1970. [28] L.G. Tornatzky, M. Fleischer, The Process of Technological Innovation, Lexington Books, Lexington, MA, 1990. [29] A. Rai, P. Brown, X. Tang, Organizational assimilation of electronic procurement innovations, J. Manage. Inform. Syst. 26 (1) (2009) 257–296. [30] M.J. Gallivan, Organizational adoption and assimilation of complex technological innovations: development and application of a new framework, Database Adv. Inform. Syst. 32 (3) (2001) 51–85. [31] A.C. Avgar, A.S. Litwin, P.J. Pronovost, Drivers and barriers in health IT adoption: a proposed framework, Appl. Clin. Inform. 3 (4) (2012) 488–500. [32] F. Damanpour, M. Schneider, Initiation, adoption, and implementation of innovation: effects of context, organization, and leaders, Br. J. Manage. 17 (2006) 215–236.

[33] S. Gopalakrishnan, F. Damanpour, Patterns of generation and adoption of innovation in organizations: contingency models of innovation attributes, J. Eng. Technol. Manage. 1 (1994) 95–116. [34] E. Rogers, Diffusion of Innovations, fifth ed., Free Press, New York, 2003. [35] B.E. Dixon, J.F. Jones, S.J. Grannis, Infection preventionists’ awareness of and engagement in health information exchange to improve public health surveillance, Am. J. Infect. Control 41 (9) (2013) 787–792. [36] J.S. Shapiro, J. Kannry, A.W. Kushniruk, G. Kuperman, et al., Emergency physicians’ perceptions of health information exchange, J. Am. Med. Inform. Assoc. 14 (6) (2007) 700–705. [37] K.B. Eden, A.M. Totten, S.Z. Kassakian, P.N. Gorman, M.S. McDonagh, B. Devine, M. Pappas, M. Daeges, S. Woods, W.R. Hersh, Barriers and facilitators to exchanging health information: a systematic review, Int. J. Med. Inform. 88 (2016) 44–51. [38] L.M. Korst, C.E. Aydin, J.M. Signer, A. Fink, Hospital readiness for health information exchange: development of metrics associated with successful collaboration for quality improvement, Int. J. Med. Inform. 80 (8) (2011) 178–188. [39] C. Sicotte, G. Paré, Success in health information exchange projects: solving the implementation puzzle, Soc. Sci. Med. 70 (8) (2010) 1159–1165. [40] J.M. Grossman, T.S. Bodenheimer, K. McKenzie, Hospital-physician portals: the role of competition in driving clinical data exchange, Health Aff. 25 (6) (2006) 1629–1636. [41] L. Dimitropoulos, V. Patel, S.A. Scheffler, S. Posnack, Public attitudes toward health information exchange: perceived benefits and concerns, Am. J. Manage. Care 17 (12) (2011) 111–116. [42] J. McMurray, L. Zhu, I. McKillop, H. Chen, Ontological modelling of electronic health information exchange, J. Biomed. Inform. 56 (2015) 169–178. [43] H. Park, S.I. Lee, H. Hwang, Y. Kim, E.Y. Heo, J.W. Kim, et al., Can a health information exchange save healthcare costs? Evidence from a pilot program in South Korea, Int. J. Med. Inform. 84 (9) (2015) 658–666. [44] C.M. Carr, C.S. Gilman, D.M. Krywko, H.E. Moore, B.J. Walker, S.H. Saef, Observational study and estimate of cost savings from use of a health information exchange in an academic emergency department, J. Emergency Med. 46 (2) (2014) 250–256. [45] V. Patel, E.L. Abramson, A. Edwards, S. Malhotra, R. Kaushal, Physicians’ potential use and preferences related to health information exchange, Int. J. Med. Inform. 80 (3) (2011) 171–180. [46] V.A. Yeager, D. Walker, E. Cole, A.M. Mora, M.L. Diana, Factors related to health information exchange participation and use, J. Med. Syst. 38 (8) (2014) 1–9. [47] C.M. DesRoches, E.G. Campbell, S.R. Rao, K. Donelan, T.G. Ferris, A. Jha, et al., Electronic health records in ambulatory care—a national survey of physicians, N. Engl. J. Med. 359 (1) (2008) 50–60. [48] L.M. Kern, R. Kaushal, Health information technology and health information exchange in New York State: new initiatives in implementation and evaluation, J. Biomed. Inform. 40 (6) (2007) 17–20. [49] C. Parker, M. Weiner, M. Reeves, Health information exchanges—unfulfilled promise as a data source for clinical research, Int. J. Med. Inform. 87 (2016) 87. 1–9. [50] J.R. Vest, T.R. Campion, L.M. Kern, R. Kaushal, et al., Public and private sector roles in health information technology policy: insights from the implementation and operation of exchange efforts in the United States, Health Policy Technol. 3 (2) (2014) 149–156. [51] S.E. Ross, L.M. Schilling, D.H. Fernald, A.J. Davidson, D.R. West, Health information exchange in small-to-medium sized family medicine practices: motivators, barriers, and potential facilitators of adoption, Int. J. Med. Inform. 79 (2) (2010) 123–129. [52] J.R. Vest, J. Jasperson, What should we measure? Conceptualizing usage in health information exchange, J. Am. Med. Inform. Assoc. 17 (3) (2010) 302– 307. [53] P. Fontaine, S.E. Ross, T. Zink, L.M. Schilling, Systematic review of health information exchange in primary care practices, J. Am. Board Family Med. 23 (5) (2010) 655–670. [54] K.T. Fuji, K.A. Galt, A.B. Serocca, Personal health record use by patients as perceived by ambulatory care physicians in Nebraska and South Dakota: a cross-sectional study, Perspect. Health Inform. Manage./AHIMA, Am. Health Inform. Manage. Assoc. 5 (2008). [55] M.F. Furukawa, V. Patel, D. Charles, M. Swain, F. Mostashari, Hospital electronic health information exchange grew substantially in 2008–12, Health Aff. 32 (8) (2013) 1346–1354. [56] J. Adler-Milstein, C.M. DesRoches, A.K. Jha, Health information exchange among US hospitals, Am. J. Manage. Care 17 (11) (2011) 761–768. [57] J.M. Grossman, K.L. Kushner, E.A. November, et al., Creating Sustainable Local Health Information Exchanges: Can Barriers to Stakeholder Participation Be Overcome?, Center for Studying Health System Change, Washington, DC, 2008 [58] J. Adler-Milstein, A.K. Jha, Sharing clinical data electronically: a critical challenge for fixing the health care system, JAMA 307 (16) (2012) 1695–1696. [59] J.J. Cimino, M.E. Frisse, J. Halamka, L. Sweeney, W. Yasnoff, Consumermediated health information exchanges: the 2012 ACMI debate, J. Biomed. Inform. (48) (2014) 5–15. [60] M.B. Buntin, S.H. Jain, D. Blumenthal, Health information technology: laying the infrastructure for national health reform, Health Aff. 29 (6) (2010) 1214–1219. [61] M. Ozkaynak, P.F. Brennan, Revisiting sociotechnical systems in a case of unreported use of health information exchange system in three hospital emergency departments, J. Eval. Clin. Pract. 19 (2) (2013) 370–373. [62] J.R. Vest, L.D. Gamm, Health information exchange: persistent challenges and new strategies, J. Am. Med. Inform. Assoc. 17 (3) (2010) 288–294.