Internet-based self-care programs (ISPs) belong to a category of health information technology, which consist of personalized, self-guided interventions delivered electronically to improve knowledge, awareness, or behavior change for mental or behavioral health conditions [1]. ISPs present interactive therapeutic material in rich formats such as audio, video, and text. Participants use ISPs at a pace and in a setting of their choosing, while being provided varying levels of support [2,3]. Evidence-based ISPs have shown efficacy for the treatment of many common conditions including depression, anxiety, substance use, and insomnia [4,5].

The potential health service benefits of ISPs are pronounced, as they may mitigate chronic and systemic problems associated with access to mental health services [6]. Access to services is a challenge even for integrated health care systems such as Veterans Health Administration (VHA), which are accountable to provide a coordinated range of services to a defined population [7,8]. One solution may lie in the implementation of ISPs, the benefits of which include reduced travel barriers, improved access to evidence-based therapies, and reduced stigma [9]. ISPs may also increase self-care and add to health care system productivity [10,11]. Within the VHA, ISPs may especially benefit patients where travel time and distance to VHA facilities or the lack of trained providers are limiting factors [12].

Despite the advantages of ISPs and evidence of general acceptability, ISP use in US health care systems, including the VHA, is not well documented and presumed low [13,14]. Although ISP use in the United Kingdom, Europe, and Australia is more routine and to some extent institutionalized in policy, there have been notable difficulties with ISP implementation at the point of care with patient and providers in the primary care settings [15-17]. For instance, a recent major pragmatic trial implementing ISPs for depression in primary care sites across the United Kingdom resulted in relatively low use of the programs by patients [18].

The lack of ISP use in US health care settings and the relative difficulty of implementation in other countries suggest the need for comprehensive strategies to address ISP implementation. Such strategies are dependent on the determinants of practice for a given setting—defined as the human and system factors, which determine to what extent and manner interventions are used. Determinants of practice have been referred to elsewhere by alternative terms such as barriers, obstacles, enablers, and facilitators [19]. For instance, in a study implementing evidence-based practices for the treatment of chronic diseases in primary care, the use of checklists by providers and self-management among patients were identified as important determinants of practice [20]. However, unlike other interventions, determinants of practice for ISPs rely on the acceptance and use of both technology and self-care by patients, providers, staff, and administrators [21].

The identification of determinants of practice for ISP use will be a key step in the development of effective strategies for ISP implementation [22]. The objective of this study was to identify determinants of practice for ISP implementation perceived by VHA primary care and integrated mental health care providers and administrators. The identification of determinants will aid in the development of comprehensive strategies for ISP implementation.

VHA primary care provider and administrator perspectives on ISP implementation were explored through qualitative analysis of semistructured interviews. The institutional review board of the VA Connecticut Healthcare System preapproved this study.

To identify determinants encountered in care, participants were primary care providers, primary care mental health providers, clinic/facility administrators, and VHA system administrators who were known to be using or have used ISPs in patient care, interested in using ISP in primary care settings, or otherwise knowledgeable on the subject of ISP use. VHA does not systematically track or record patient or provider use of ISPs, and as yet, there is no validated way to search VHA electronic health records for evidence of use. Therefore, a provider sampling strategy based on in-depth investigator knowledge of the VHA system was used. Investigators identified an initial group of VHA key informants in ISP use. This group was expanded by asking interviewees for referrals to additional key informants. Recruitment was through e-mail contact that included a project description. Physicians, psychologists, nurse practitioners, nurse administrators, and social workers with both clinical and administrative responsibilities were recruited. A total of 20 interviews were conducted; 100% of those approached agreed to participate.

The principal investigator (EH) conducted semistructured interviews using a guide that first defined ISPs, giving examples of VHA-developed ISPs and other programs if needed, such as Moving Forward, PTSD Coach, and Sleep Healthy Using the Internet [23-25]. Determinants were then explored using 9 core question areas, with subquestions to explore specific topics that emerged. Question areas were derived from constructs (performance expectancy, effort expectancy, social influence, and facilitating conditions) from the Unified Theory of Acceptance and Use of Technology, a theory that describes factors associated with technology use [21]. Interviews were audio recorded and conducted over the phone, lasting between 45 and 90 min.

Interviews were transcribed verbatim. Thematic analysis, a foundational approach to the analysis of in-depth interviews, was used to analyze interview content through a 5-step group process: (1) familiarization with transcribed data; (2) generation of initial codes; (3) collating codes into themes; (4) reviewing, discussing, and modifying themes in relation to interview extracts; and (5) defining and developing a typology of themes [26]. Codes were first generated by individual investigators, and then discussed as a group, referencing the interview content to develop a final set of codes. Discrepancies in coding were addressed through group discussions led by the principal investigator (EH). Similarly, themes were then developed from codes using a group discussion process. Codes within themes were reviewed using the entire dataset to identify examples and counterfactuals. Authors EH, LB, and EP participated in steps 1 through 3, whereas all authors participated in steps 4 and 5. Themes were organized according to socio-ecological levels within the integrated health care system: patients, providers, and the organization [27]. This manuscript describes patient- and provider-level themes associated with determinants of practice for ISP use at the point of care. Determinants active at the VHA organizational level will be described in future work. Data were managed and analyzed using Atlas/ti. version 4.2 software (ATLAS.ti Scientific Software Development GmbH).

Of the 20 individuals participating, more males than females were interviewed, while the training, background, and primary work locations of interviewees were relatively balanced (Table 1). Interviews began with questions concerning interviewees’ and their colleagues’ experience with using ISPs in clinical practice. ISP use among interviewees was relatively low. Although 16 (80%, 16/20) interviewees had used an ISP or discussed ISPs with patients at some point in practice, only 6 (30%, 6/20) were currently using them. One primary care provider stated:

Despite this level of use, all 20 interviewees (100%, 20/20) made statements in support of ISP use in primary care. One primary care provider stated:

The aim of this study was to explore primary care provider and administrator perspectives on ISP implementation to identify important perceived determinants of practice for ISP use. We identified 10 themes for determinants of practice at both patient and provider levels. An understanding of these determinants is vital to the development of strategies for ISP implementation in primary care. Comprehensive implementation strategies for ISPs have yet to be clearly defined or rigorously tested. Our investigation of determinants suggests functions for such an ISP implementation strategy, which we feel can be performed by a clinical intermediary (Table 2).

There is little information on rates of ISP use among US providers or within US health care institutions, and we presume this use to be low. Thus, the relatively infrequent use of ISPs reported by this group of interviewees is not too surprising. Nonetheless, interviewees were generally supportive of ISP use, in spite of identifying a number of determinants of practice that can be considered barriers to ISP use. The support for ISPs identified in this study is consistent with the general acceptance of and support for health information technology among providers and patients [13,29]. Support for ISPs relative to actual use may suggest an emerging tipping point for the increased use of ISPs in the United States, as has occurred in other countries. For example, a 2008 survey of mental health providers in the United Kingdom revealed that 99% used ISPs in care and 39% had been trained to use them [30]. At a minimum, our findings indicate that provider and patient attitudes toward ISPs are not strong barriers to implementation.

Perceived patient-level determinants of ISP use including literacy, internet access, and age are well documented, as in a recent analysis of interviews with VHA service users who stopped using an ISP for insomnia before program completion [31-34]. However, conflicting evidence suggests that internet use and the acceptability of ISPs among VHA service users are on par with those of the general population [35,36]. Given these differences and the relative difficulty of modifying the above determinants, compared with other patient-level determinants identified here and in similar studies (ie, interest/motivation, patient expectations, and ISP compatibility with patients), other determinants may be more amenable to mitigation and suitable for incorporation into strategies used to implement ISPs (see discussion below and Table 2) [37].

Two themes at the provider level—familiarity with available programs and change to the traditional delivery of care—stand out as key in this study and the few others that have evaluated provider perspectives on ISP use [38,39]. ISP familiarity as a determinant of practice suggests that educating providers about available ISPs and having them experience the platform should be a component of a comprehensive ISP implementation approach. Educational strategies to implement ISPs in health care settings must be flexible to account for varying levels of baseline ISP familiarity among providers. The likely downstream effects of increased provider familiarity with ISPs are the changes more ISP use could bring to the traditional delivery of care [40]. Providers must become more comfortable recommending not only technology use but also self-care to patients. Similarly, a system of stepped care must be in place so that patients who are not appropriate for or fail initial attempts at ISP treatment are identified and referred to other forms of care (eg, more provider interactions) [16].

The determinants of practice identified here and those from other studies of ISP use outside of primary care, mental health, and VHA settings suggest the need for comprehensive strategies for ISP implementation [31,38,39]. Given the current level of support for ISPs identified among providers in this study, the function of such a strategy should not rely solely on educating providers about ISPs to gain support. Rather, key strategy functions should be more nuanced and include increasing provider familiarity with specific programs, addressing competing demands, providing internet access, and modulating patient expectations, among others (Table 2). Also identified among these determinants is the need for human contact during ISP use. Such contact has already shown evidence for improving adoption and outcomes of ISPs. Support from a human may help address many of the other ISP implementation strategy functions identified here [41,42]. We propose that a clinical intermediary, an individual different from the primary provider, can deliver this human contact. As suggested by interviewees in this study, such a clinical role could be performed by peer specialists, nurses, care managers, and other clinicians trained to facilitate ISP use by engaging in the important implementation strategy functions described in Table 2. Future research should test the use of a clinical intermediary and identify the content, timing, and other specifics of these strategy functions as well as the training needed for this role. In addition, the materials needed to train a clinical intermediary could be spread health system-wide as part of future ISP dissemination efforts.

This study has several limitations. Interviews involved a relatively small group of individuals, and the frequency of ISP use within this group was low despite our objective to recruit ISP users. As such, these findings may represent only an initial indication of the important perceived determinants of practice for ISP use, and may be strengthened by a larger sample size or the inclusion of more VHA providers who regularly use ISP in care if they exist. In addition, the purposeful inclusion of individuals who have chosen not to use ISPs in care may reveal yet more critical determinants of practice. Moreover, it must also be noted that determinants were developed from the perspectives of VHA providers and administrators. We did not verify these determinants using quantifiable means. However, interviewees had a breadth of relevant backgrounds and adequately represented ISP stakeholders within VHA. They, in turn, discussed a range of determinants of practice that influence ISP use. In addition, interviews with health care consumers may reveal yet more determinants. Organizational-level determinants of practice that might suggest enterprise-wide dissemination and implementation strategies or policy changes were not discussed but will be examined in future work. Finally, the VHA is a unique, integrated, and nationwide health care system that provides care for US Veterans. Thus, our results may not be representative of providers or patients in other health care systems.

Developing and testing implementation strategies to facilitate the use of ISPs in primary care is a critical area in need of investigation. As an initial step, this study identified patient- and provider-level determinants of practice for the use of ISPs in the context of outpatient primary care settings in the VHA. An ISP implementation strategy that addresses the determinants identified here, including the use of a clinical intermediary to support both patients and providers, may be an effective way forward. These findings, as well as those from similar studies, can be used to inform the development of ISP implementation strategies, which can be further described and tested in future work.