The first step of knowledge synthesis involves the CIS to establish a foundation for the development of the adoption guide. The CIS is selected over other knowledge synthesis methods due to the complex and multifaceted nature of the existing literature on technology adoption in health care. CIS is specifically designed to interpret and analyze diverse evidence and generate new theoretical understanding and conceptual frameworks [25]. The iterative, flexible, and dynamic approach allows for more interpretive analysis that is crucial for examining multifaceted adoption challenges [26]. The CIS process involves formulating the review question, “How do existing evidence perspectives inform the adoption of wearable telemonitoring systems for chronic disease management in hospital settings?,” which will serve as a compass to guide the inquiry. The second step consists of a comprehensive literature search across multiple databases—including PubMed, JMIR, IEEE Xplore, Cochrane, and Google Scholar—as well as conference proceedings and dissertation databases, to identify a sampling frame of potentially relevant papers [26]. This study uses the following search terms: adoption or implementation, wearable, telemonitoring or remote monitoring, chronic disease, and their derivatives as keywords. The Boolean operators “AND” and “OR” will be used to combine these terms (Multimedia Appendix 1).

Eligibility criteria are defined by content, timeframe, context, study design, and language. Papers that investigate wearable or telemonitoring systems for chronic disease management and those that address implementation or adoption challenges, facilitators, or guidelines for adopting telemonitoring technology are included (content). Concerning the timeframe, papers are eligible if published in 2015 or after, as this period captures the rise of wearable technology. The context is research undertaken in the health care setting, including hospitals. Finally, papers written in English or Indonesian using all study designs are eligible.

The third step is selecting papers purposively based on their likely relevance to offer key conceptual insights and identify the initial range of concepts related to telemonitoring adoption in chronic diseases. This initial sampling is then followed by theoretical sampling, an iterative process that operates concurrently with theory generation, where researchers dynamically refine questions and literature searches based on the emerging themes until theoretical saturation is reached [25-27]. Empirical studies concerning adoption or use of wearable telemonitoring systems in chronic diseases are selected to capture the actual implementation phenomena. We exclude articles that merely focus on the technological aspects, such as the development of devices or systems, without their implementation in patients.

The fourth step is quality appraisal adopting a critical orientation, in a broader sense of critique. Rather than applying rigid critical appraisal, where each study is judged against a predefined set of methodological standards, this approach involves deep questions about how the literature constructs problems, assumptions, and influences proposed solutions, while applying a low threshold for inclusion to maximize the contribution of papers at the conceptual level and exclude fatally flawed papers. A study is defined as fatally flawed if it fails to meet the methodological standards; specifically, it lacks clearly stated research objectives, employs an inappropriate research design for the objectives, fails to provide a clear research process, presents insufficient data to support its interpretation and conclusions, or uses inadequate and unclear methods of analysis [25]. The goal of appraisal focuses on generating key concepts relevant to the adoption of telemonitoring systems using wearable devices, thus applying flexible relevance boundaries in order to be as inclusive and comprehensive as possible.

The fifth step is data extraction. This process is a deeply interpretive and iterative process [26]. The process begins with a detailed inspection of the included papers, focusing on identifying the main arguments or themes and key conceptual contributions of each paper [25]. Investigators note the emergent themes relating to the adoption of telemonitoring technology in hospital settings, including title, objectives, publication year, country of research, study design, sample size, population, study duration, and expected outcome, which are extracted from the articles. Subsequently, findings are identified and categorized to answer the synthesis questions, which may involve findings stated or deduced based on implications presented in the literature and investigators’ insights using relevant theoretical frameworks, involving interpretive judgment beyond summarization [28]. The analysis then aims to generate a new theoretical framework or conceptual model referred to as a synthesizing argument. The key analytical technique draws heavily on techniques from primary qualitative research, particularly grounded theory, in an iterative and dynamic manner [29]. Multidisciplinary review is encouraged to introduce checks and balances and guard against a single interpretive perspective [25]. The extracted data are then sorted and categorized into themes in line with the modified NASSS and TFA frameworks. The NASSS framework serves as the primary organizing framework to map findings into the domains (ie, technology, condition, adopter, or organization), whereas TFA serves as a complementary analytical framework to further explore the adopter and/or organizational domains, as this protocol focuses on adoption in a hospital setting with 4 organizational tiers [15].

Finally, formulation of synthesis, where the concepts are constantly compared with the data to identify the relationships among them, aims to develop a synthesizing argument in the form of a coherent theoretical framework, including a network of constructs and the relationships between them. The process of formulating synthesis is logged to document reasoning or changes as investigators correlate each finding.

CIS is an interpretative methodology that was inspired by qualitative research. Dixon-Woods et al [25] recommend multidisciplinary review teams to allow a broad range of perspectives to be incorporated throughout the synthesis process in CIS. This collaborative approach establishes a vital system of “checks and balances” that prevents framing the analysis according to a single perspective. Such collaboration is beneficial during the iterative stages of refining the “compass” research question and identifying relevant literature from adjacent fields. Engaging a multidisciplinary team of authors serves as a model for verifying the reliability of the interpretive findings and is often considered a precondition for producing high-impact research.

This protocol comprises 3 authors with multidisciplinary backgrounds to represent medical doctor practitioners (JS), health technology researcher (JS), system and technology development (WS), health care service (AU), and qualitative and implementation research (AU). Throughout the process, database searching, screening, data extraction, and quality appraisal tasks will be conducted primarily by JS as part of doctoral research. Systematic searching, screening, extraction, and preliminary findings will be thoroughly discussed with WS and AU for critical review, verification, and correction. Disagreements among the team will be resolved through consensus. The process of key analytical decisions and how disagreements are resolved will be documented.

Phase 2 aims to delve deeper and explore the complexities of telemonitoring adoption in hospital settings. This involves the process of contextualizing guidelines by translating CIS findings into actionable items. These discussions will be structured to explore adoption factors across the technological requirements, implementation procedures, and institutional considerations in correspondence to the modified NASSS framework (technology, organization, and adopters) to produce documented recommendations with actionable statements. The revealed concerns of adopters are noted for subsequent analysis through the TFA constructs. Data collection will involve semistructured individual interviews with stakeholders from 4 different hospitals with telemedicine services available. Within each hospital, informants will be selected through intensity sampling, a purposive strategy that selects information-rich cases manifesting the phenomenon of interest intensely but not extremely [30]. To capture diverse roles across hierarchical levels and professional groups, participants are recruited across 3 organizational tiers: strategic, middle, clinical, and supporting tiers. The strategic tier is represented by the hospital director or vice director of medical service (n=1), and the middle tier will be those who manage team members and understand the operational workflows, resource capabilities, and budgetary considerations, that is, head of the medical service division (n=1) and head of the IT department (n=1). The clinical tier refers to health care professionals delivering clinical services, represented by doctors and nurses (n=1‐2 persons each), while administrative staff (n=1‐2 persons) and information system support (n=1‐2 persons) will represent the supporting tier. Therefore, the tentative number of participants is 8 to 11 participants per hospital, or in total 32 to 44 participants, subject to saturation. This number may vary across the 4 hospitals as some hospitals may apply leaner organizational structures.

Data collection will continue until thematic saturation is achieved. Saturation will be assessed first by within-tier group saturation, where data will be evaluated by stakeholder tiers to specify operational perspectives related to their roles. Second, cross-group saturation will be evaluated across the entire dataset as a whole to identify full insights into adoption barriers, facilitators, and acceptability concerns. Saturation will be monitored iteratively and concurrently during data collection. As interviews progress across the sites, the research team will continuously review the new codes emerging within each tier. When 2 consecutive interviews within a tier yield no new conceptual codes, saturation for the tier is considered achieved [31]. The decision regarding saturation will be made through discussions with coauthors (AU and WS). This dual approach to considering saturation, that is, within- and cross-tier groups, aligns with the methodological model of information power proposed by Malterud et al [32], where qualitative sample size sufficiency is determined by the depth of information a sample holds rather than merely the number of informants.

The primary investigator (JS), a male doctoral student with a medical doctor practitioner background and health technology researcher, will conduct the face-to-face semistructured interviews. Having an understanding of the technical infrastructure will enable JS to conduct probing during interviews. However, to prevent the interviewer’s potential biases, neutrality will be maintained using a reflexive diary throughout data collection. The research process will be supervised by a senior female researcher with long experience as a qualitative researcher (AU). Prior to the interview, JS will introduce the study in a meeting facilitated by the participating hospitals, followed by approaching the participants individually via face-to-face meetings to create appointments for the interview. The interview will be conducted in the hospital setting for about 60 to 90 minutes and audio-recorded, with supplementary notes taken by the interviewer. The interview guide will be developed based on the CIS findings and pilot-tested for clarity before the formal data collection. The semistructured interview will explore adoption factors across the technological requirements, implementation procedures, and institutional considerations in correspondence to the modified NASSS framework domains, while any revealed concerns of adopters are analyzed through the TFA constructs.

Data analysis will follow a thematic approach, combining inductive and deductive methods. An initial deductive codebook will be developed a priori based on the modified NASSS framework domains (condition, technology, adopter, and organization) and the TFA constructs. This codebook will provide a starting point for coding while remaining open to themes emerging from the data. All interviews are transcribed verbatim, deidentified, and organized by hospital and organizational tiers. NVivo software will be used for organization, retrieval, and cross-referencing of coded data across participants. All authors will familiarize themselves with the transcripts.

The following data analysis will be carried out. First, inductive open coding will be conducted on the transcripts to identify meaning units and emergent codes grounded in participants’ perspectives. Second, the codes will then be assigned according to the predetermined categories deduced from the NASSS-TFA codebook. Those within the adopter and organization categories will be further grouped using the TFA constructs. Third, codes emerging from the inductive analysis will be reconciled with the deductive codebook. New categories may still emerge from inductive codes that do not fit within the NASSS-TFA domains. As categories should be mutually exclusive, all categories will then be reviewed until no categories overlap. Decisions made during the reconciliation process will be documented.

Themes are refined through an iterative peer debriefing process between authors. Analysis specifically maps barriers, facilitators, and opportunities for wearable telemonitoring system adoption to the modified NASSS framework (technology, adopter system, health/care organization, and condition) to contextualize findings. Within the organization domain, TFA serves as a deductive coding and interpretative lens. The inductively derived themes are then compared against TFA constructs as an organizing template to categorize underlying adoption barriers, facilitators, and opportunities. Consistent with the tier-actor–based approach, the analysis will prioritize cross-stakeholder comparison to identify “who does what” in the adoption of telemonitoring systems. This cross-stakeholder analysis will inform the tiered structure of the guideline to address the specific responsibilities, concerns, and acceptability mechanisms relevant to each organizational tier. The final synthesized themes, illustrated with representative quotes, will inform the guideline’s development by comparing the existing matrix results from phase 1 and identifying gaps; new role-specific recommendations will be formulated, or local context conflicts with global findings will then be enhanced using the same standardized syntax established in phase 1.

The final phase of this study employs a modified NGT, a structured consensus-building process and controlled small-group process ideal for synthesizing expert judgments to develop clinical practice guidelines [33-37]. The process is reported following the ACCORD (Accurate Consensus Reporting Document) reporting guideline for consensus methods [38]. NGT is selected over alternative consensus methods because its structured format prevents single individuals from dominating the discussion. Various perspectives of experts allow a multidisciplinary panel of experts to transparently assess and reconcile differing assumptions within the forum [37,39]. Moreover, NGT allows for listing issues during the development and deployment of eHealth applications. It is considered an adequate technique for conceptual validation, serving to verify the potential suitability, sustainability, and practicality for real-life decision-making [40,41].

A purposively selected panel of approximately 12 experts will be invited to participate in the NGT through consensus among all authors. To capture a comprehensive range of perspectives, the NGT panel will comprise hospital directors, managers of medical services, finance, and information technology, as well as health service experts from academia. A person is considered an expert if they possess a high level of knowledge or lived experience relevant to the topic, in line with methodological guidance for the expert panel [42]. Hospital directors are eligible if they are currently working or former hospital directors with 5 years of experience at the strategic level and are familiar with telemedicine services. Managers at the managerial level are eligible if they are currently working or formerly worked in hospital management, have 3 years of experience in a decision-making role within a hospital organization in medical services, administration/finance, or IT, and are familiar with telemedicine. A health service expert is a teacher/researcher in health care/hospital management and registered as a lecturer at a recognized university in Indonesia. Panel recruitment will be started with purposive sampling, followed by snowball sampling, where initially selected participants recommend eligible colleagues. Panel participants will be reimbursed for their time.

The NGT process will be implemented using the Indonesian language, led by a steering committee. The steering team consists of the first author (JS), supervised by AU and WS, with research experience in health care (AU) and technology (WS). Prior to commencement, a briefing document will be prepared by JS and disseminated to all expert panel members by email 1 week in advance, as not all participants are familiar with the topic.

The NGT will follow 5 key stages [33]: (1) introduction and explanation, where V1 guideline as a result of knowledge synthesis and qualitative study is a matter to discuss and the core question for improvement is clarified, that is, “We have developed the guideline based on the evidence synthesized (CIS) and stakeholders’ feedback (qualitative study). The materials are not piloted for the NGT prior to the NGT process. What is needed to improve the adoption of wearable telemonitoring guidelines for hospitals to manage chronic conditions?” This step also gives an opportunity to state the aim, which is development of an adoption guideline based on prior phases’ findings, and to clarify the question; (2) silent generation of ideas, where participants individually reflect on and document their suggestions to enhance the guideline; (3) round-robin sharing, where each participant states 1 idea at a time for transcription without discussion; this stage ensures every participant has an equal opportunity; (4) clarification, where the group discusses each idea for understanding and to group similar concepts; and (5) voting by participants. The steering committee will not be included in the decision of consensus.

The ranking stage will involve 2 rounds of private voting using a 9-point Likert scale (1=highly inappropriate, 9=highly appropriate) to evaluate both appropriateness and feasibility of each recommendation. Following established consensus methodologies, a predefined threshold will be used to determine agreement and govern the final Domain-by-Actor matrix. Items achieving a median score of 7 to 9 paired with majority agreement of more than 50% plus 1 expert will be considered “appropriate” and included; items with a median of 1 to 3 will be considered “inappropriate” and discarded; and items with a median of 4 to 6, or those with a median of 7 to 9 that fail to reach more than 50% majority agreement threshold, will be flagged for further discussion and modification in the second voting round, allowing for reevaluation and refinement of the consensus qualitatively [43,44]. If the modified items clear the majority agreement threshold on the second and final vote, they will be included; if they fail to achieve more than 50%, they will be systematically excluded to protect the integrity of the framework. After each round of voting, the steering committee will present the result of the voting round. The research phases of the study design are summarized in Figure 3.

The multiphase research design is chosen to address the development of adoption guidelines, as it builds a foundation based on existing evidence, is contextually grounded by a qualitative study, and is expert-validated. The design follows an exploratory sequential mixed methods approach to develop adoption guidelines. The synthesized foundation through CIS is contextualized in the qualitative study to refine the guideline. All 3 phases are essential, but the qualitative phase holds the highest importance, as it is meant to contextualize global findings into an Indonesian context relevant to the target setting, which will also be the content of expert validation in the NGT. The integration occurs at 3 points. First, after the CIS, the results are used to develop the semistructured interview guide. This ensures data collection by JS, supervised by AU, is grounded in the existing evidence. Second, after the qualitative study, themes are integrated to produce version 1 of the guideline, combining the CIS result guideline into a more contextualized version of the draft (V1); this point is completed by JS and all 3 authors. The third integration point is when the draft guideline after the qualitative study is presented to the NGT panel and is reviewed by an expert panel, refined, and prioritized through consensus. The definitive endpoint of this study is reached when the matrix is populated with role-specific, actionable recommendations synthesized and validated through the 3 study phases.