This scoping review was conducted using the methodological framework developed by Arskey and O’Malley [16]. The Arksey and O’Malley framework is particularly suited to address broad research questions and can help map the current literature, extract key concepts and themes, and identify gaps. The Arksey and O’Malley framework has several steps including (1) identifying the broad research question, (2) study selection using inclusion or exclusion criteria on the basis of familiarity with the topic of interest, (3) sorting the extracted data from studies into themes and patterns, and (4) collating key themes and issues [16]. The primary research question guiding this review was the following: What are the barriers and facilitators of digital health adoption for hypertension management?

Structured literature searches were conducted using 3 databases to identify relevant studies from 2008 to 2017: PubMed or Medical Literature Analysis and Retrieval System Online, Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Excerpta Medica database (EMBASE). Medical subject headings (MeSH) and selected keywords were searched using Boolean operator OR and these groups were combined using another Boolean operator AND. Keywords used include (1) hypertension (MeSH), hypertensi, (2) mobile applications (MeSH), mobile device, (3) electronic health records (MeSH), personal electronic health record, (4) decision support systems, decision support, (5) remote monitoring (MeSH), (6) providers (MeSH), clinician. The detailed search strategies for PubMed have been provided as an example (see Multimedia Appendix 1). At first, 2 reviewers, with subject matter and methodological expertise, independently reviewed all abstracts identified by the searches and conflicts were resolved by a third reviewer. Then, 2 reviewers screened the full texts to select the final studies to be included in the review. Cohen kappa test revealed an agreement score of 0.75 between the reviewers. Per Landis and Koch, this agreement score could be categorized as substantial agreement between the reviewers [17].

All articles retrieved were screened using the following inclusion criteria: (1) reported on adoption barriers and/or facilitators of digital health solutions, as defined by the AMA, that were provider- or patient-related, (2) focused on hypertension management, (3) published in English, and (4) published between 2008 and 2017. Studies were excluded if they (1) did not report on barriers or facilitators of digital health, (2) described barriers or facilitators exclusively for nonclinical staff such as pharmacists, (3) were editorials or reviews for editorials, epidemiological studies, and protocols, (4) provided insights on acute management of hypertension in perioperative or intensive care settings, or (5) if full texts were unavailable. The authors also conducted a gray literature search (including conference proceedings) through a Web search engine. In addition, 4 articles were handpicked on the basis of the same inclusion criteria used for articles selected via literature databases.

The selected papers were reviewed to extract relevant data. A data extraction template developed by the authors was used to extract key information and concepts from the included studies and the template included the following constructs: the geography, study design, program setting, disease conditions (in addition to hypertension), study objectives, sample description, sample size, digital health category, design features of the intervention, clinical outcomes, cost outcomes, patient experience, provider experience, patient-related barriers and facilitators, and provider-related barriers and facilitators. Descriptive and inductive thematic analyses were conducted for identifying major themes pertaining to barriers and facilitators of DHT adoption. For the analysis of the text passages from the included articles, the inductive thematic analysis was conducted as described by Braun and Clarke [18]. We developed our own a priori framework to categorize barriers into the following 4 categories: (1) provider-related facilitators, (2) provider-related barriers, (3) patient-related barriers, (4) and patient-related facilitators. This analytic process involved reading and rereading of the selected papers, systematically identifying and naming the unit of meaning with codes (words or sets of words that provide a meaning label), and then searching for patterns in the data and organizing the data (smaller themes or codes) into larger themes representing the main ideas and their relationships. Themes were then reviewed by the team and representative data elements were selected to demonstrate the salient themes. At first, 2 investigators (RP and NF) independently performed the initial coding of the first transcript. This coding was then reviewed by the third reviewer (AC). The codes were then reviewed and discussed with the team including senior researchers in the field, providers, and other subject matter experts. Later, 2 reviewers (RP and NF) then recoded all papers, integrating feedback from the team into the coding structure. A final codebook was created using Microsoft Office Excel (version 1808) on the basis of the consensus of the 3 investigators (RP, NF, and AC). During this process, any discrepancies in coding were discussed and resolved among all investigators. Furthermore, any questions about meaning and interpretation of themes were discussed among the team members and resolved through consensus.

A total of 2299 titles and abstracts from PubMed, CINAHL, EMBASE, and 4 handpicked articles from the supplementary gray literature search were assessed for eligibility after removing duplicates (see Figure 1). Of these, 2165 references did not meet the inclusion criteria. After assessing 134 studies in full text, 98 studies were excluded (full texts were either unavailable or studies did not fulfill the inclusion criteria). A total of 36 studies satisfied the inclusion criteria, including the 4 handpicked articles. The articles included in this review were published between 2008 and 2017, with a majority (n=30) published after 2010. Studies were published across the following countries: United States (n=21), United Kingdom (n=4), Canada (n=3), Finland (n=1), Sweden (n=1), Italy (n=1), Taiwan (n=1), Malaysia (n=1), South Korea (n=1), Kenya (n=1), and Germany (n=1). DHTs included in this review were classified into categories as defined by the AMA: remote monitoring for efficiency (n=6), remote monitoring and management for improved care (n=19), clinical decision support (n=6), patient engagement (n=4), televisits or virtual visits (n=6), point-of-care(n=2), and tools providing consumer access to clinical data (n=1). Most studies were conducted in a primary care setting (n=30). A plurality of studies included qualitative assessments (n=15). Quantitative methodologies included randomized controlled trials (RCTs; n=14), nonrandomized trials (n=2), usability pilots (n=2), and pre and poststudies (n=2). In addition, 1 white paper was also included in this review. Multimedia Appendix 2 displays a summary of the studies included in this review. The results of the thematic analysis have been categorized as provider- and patient-related facilitators and barriers as detailed below. Tables 1 and 2 summarize all the themes.

This review contributes to existing literature by highlighting factors that enable or hinder the adoption of digital health solutions from the perspectives of both providers and patients. These results show that the key facilitators of DHT adoption by physicians include integration with clinical workflow 33% (8/24), ease of use 21% (5/24), improvement in patient outcomes 21% (5/24), financial factors 8% (2/20), and organizational support 4% (1/20). Technology usability and technical support 29% (10/35), positive impact on well-being and self-management 26% (9/35), improved patient-provider relationship 24% (8/35), and a reduction of in-office visits 20% (7/35) were most frequently reported facilitators for patients. The most frequently reported barriers to use of DHTs reported by physicians include lack of integration with workflow 36% (10/28), lack of validation of technology 25% (7/28), and lack of usability and support 25% (7/28). Finally, a lack of technology usability 41% (9/22), interference with the patient-provider relationship 23% (5/22), and lack of validation of technology 14% (3/22) were the top barriers reported by patients.

Although these findings highlight some common themes reported in previous work, there are several key differences and contributions from this study. A 2017 study by Mileski et al, examining the facilitators and barriers to implementing telemedicine for hypertension management [13], only focused on telemedicine, whereas our study examined all DHTs from the perspective of both patients and providers. Consistent with Mileski et al, we found that improved outcomes, increased patient knowledge and self-management, and cost savings were important facilitators of DHT adoption. Another systematic literature review by Gagnon et al [15] evaluated the factors influencing adoption of DHTs by health care professionals and some barriers reported in this review, such as the lack of organizational support and lack of reimbursement for providers, these were consistent with our study findings. Furthermore, most of the studies included in the review by Gagnon et al were conducted in large hospitals. In contrast, most studies in our review, 86% (31 out of 36 studies), were conducted in primary care settings. Additionally, Gagnon et al [15] examined DHTs across multiple diseases, whereas our review focused specifically on DHTs for hypertension management.

Multiple conceptual models exist to describe acceptance and usage of technology, such as Rogers diffusion of innovations theory [54], the technology acceptance model [55], and the unified theory of acceptance and use of technology (UTAUT) [56]. These models have been applied to describe the adoption of electronic health records and other forms of DHTs [57]. As a thematic analysis approach was used to identify new or emergent themes, we neither tied our analysis to a preexisting conceptual model nor sought to validate a preexisting conceptual model. However, it is worth noting that the themes that emerged from our analysis align with several of the constructs described in UTAUT. For example, the themes of clinical workflow integration and technology usability relate to the UTAUT construct of effort expectancy. Similarly, the theme of improvement in patient outcomes relates to the UTAUT construct of performance expectancy.

Lack of usability or ease of use was found to be a major barrier for both patients and providers in our review. Furthermore, lack of integration with clinical workflow was an important barrier for physicians. In the light of these findings, it is important that developers of DHTs should aim to improve the experience of both patients and providers through human-centered design thinking principles [58]. Such a process considers the needs and perspectives of all stakeholders during the product development cycle and implementation in a health care setting. With the right design, providers can interact with DHTs more easily to gain valuable insights on their patients’ health, without compromising their existing workflow. In addition, successful implementation of DHTs in the clinical setting demands time and resources; new programs deploying DHTs should assess all the additional resources required for managing and coordinating care of patients to reduce the burden on providers.

Furthermore, providers often require hospital leadership to be supportive of a culture of innovation within their organization while weighing risks and benefits to patients and providers [38]. Therefore, organizational commitment to engaging providers at an early stage of DHT implementation by evaluating provider needs, identifying provider champions for implementing DHTs, and providing adequate training in the hospitals are critical to foster adoption.

Although not a prominent theme in this review, some studies show that the current health care policy and regulatory landscape are increasing pressure on health care organizations to provide lower-cost and higher-quality health care [59,60]. With tightening health care budgets, identifying long-term return on investment (ROI) on DHTs and establishing financial incentives through a clear reimbursement policy for providers are vital factors in increasing provider adoption. Therefore, future studies should incorporate discussions of implementation costs and ROI, in addition to examining clinical outcomes seen as a result of DHTs.

First, as technology and policy are evolving at a rapid pace, certain barriers and facilitators that were identified in older articles may be less relevant today. Nevertheless, some facilitators and barriers are likely to remain constant over time, such as the critical importance of integration of DHTs into clinical workflow and technology usability. Second, reporting barriers and facilitators was not the primary aim of some of the studies included in this review. Thus, a portion of the data was collected from impressions reported in discussion sections of the published studies, which includes interpretations and speculations made by the researchers involved in the studies. Finally, some of the studies included in this review provided little context on barriers and facilitators reported. In such instances, reviewers used their best judgement to determine whether the barriers or facilitators reported were best categorized as provider- or patient-related barriers or facilitators. Regardless of limitations, the themes in this review provide comprehensive evidence that could better inform and strengthen DHT development and implementation.

Our findings suggest that DHT adoption for hypertension is influenced by several important factors such as integration into the clinical workflow, usability, improvements in patient outcomes, and positive impact on the patient-provider relationship. Real-world testing and incorporating feedback from both patients and providers in designing technologies will improve their overall usability. Finally, to fully realize the potential of digitally enabled hypertension management, there is a greater need to validate these technologies to provide patients and providers with reliable and accurate information on both clinical outcomes and cost effectiveness.