We conducted a consultation meeting identifying the following components needed by NHSO: payment method, rationing criteria (ie, eligibility criteria), type of technology or services reimbursed, monitoring and evaluation methods, and fraud prevention measures. We developed a research question to address the breadth and varied components of telemedicine reimbursement models and policy-making requirements: “What are the scope and main findings of existing literature on reimbursement models, benefits, and limitations for telemedicine services used in national and subnational jurisdictions?” Table 1 presents definitions we used to guide the research. Multimedia Appendix 1 shows telemedicine reimbursement components.

We searched 6 databases systematically (ie, PubMed, EMBASE (Ovid), Web of Science, Scopus, EBSCO Global Health, and ECONLIT), using search terms and syntax for telemedicine and reimbursement, adapted to the subject headings and terminology for each database. The search strategy for all 6 databases can be found in Multimedia Appendix 2. It was developed in consultation with a London School of Hygiene and Tropical Medicine librarian to improve specificity. In total, 10 studies from the final search strategy iteration in PubMed were reviewed independently by WI to confirm search strategy breadth and comprehensiveness.

We uploaded search outputs to the Covidence (Veritas Health Information Ltd) literature review software for deduplication and screening. Table 2 shows our eligibility criteria. To increase consistency among researchers, titles or abstracts were independently double-screened by any two researchers among EH-K, PM, YMT, ZNW, KKC, and SVD, until more than 80% agreement in screening decision was achieved. Conflicts were resolved through team discussions until consensus was reached. Full texts were screened using the same approach.

The 6 researchers (EH-K, PM, KKC, YMT, ZNW, and SVD) used Excel (Microsoft Corp) to extract source identifiers, relevant characteristics (eg, publication year, source type, and jurisdiction), and findings. Findings were extracted iteratively into reimbursement components as described in the Multimedia Appendix 1.

To pilot data charting, each of the 6 researchers extracted data from 2 of the 12 randomly selected papers, followed by a thorough discussion to reach consensus. Once data charting began, EH-K checked the first 5 sources charted by each researcher to improve data extraction quality and consistency.

We developed an interview guide based on the research question and reimbursement components from the NHSO consultation (Multimedia Appendix 1). To refine questions, we piloted the guide with one of the authors (EH-K), who has experience providing telemedicine services.

We used purposive and volunteer sampling by first emailing a seed list of 14 stakeholders identified from the Health Intervention and Technology Assessment Program Foundation (HITAP) network, which largely consists of Asian LMICs. To recruit additional stakeholders, we posted an invitation and study description on HITAP’s social media pages, calling for individuals with relevant experience or insights. We defined eligible candidates as telemedicine implementers (ie, implementing telemedicine in their scope of work) or providers with experience or information on telemedicine reimbursement. We received five expressions of interest, for which all candidates were deemed to be eligible. Profiles of the informants are available in Multimedia Appendix 3.

Interviews were conducted in English from August 6 to 23, 2024, over Zoom (Zoom Video Communications Inc). They lasted 45‐60 minutes, were recorded using Zoom, and transcribed using Otter.ai (Otter.ai, Inc) after anonymization. Stakeholders were asked to share links to websites or documents they considered relevant for the research team review (eg, government websites and unpublished documents). Written responses were collected from one stakeholder whom we were not able to interview. Transcripts were cleaned and validated by YMT and ZNW.

We reported review screening using the PRISMA-ScR (Preferred Reporting Items of Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) [18,19]. We summarized the extent, nature, and distribution of eligible sources descriptively. We first synthesized literature findings data, and then interviewed transcripts narratively, using a qualitative descriptive approach guided by the research question and data extraction categories (Multimedia Appendix 1). We reported relevant excerpts and insights from literature and stakeholders in the same categories. To improve internal validity and credibility, we triangulated data from interviews against literature, documents, and websites [20]. All stakeholders provided information for jurisdictions not identified from the review, except for the United States and Canada.

In the United States, Medicare and Medicaid were the reported public insurance programs, with Medicare administered at the federal level and Medicaid at the state level. In other jurisdictions, we found telemedicine publicly reimbursed at the federal level in Germany, Australia, Sweden, Belgium, the United Kingdom, and Switzerland, and at the provincial level in Ontario, Canada. As per stakeholders, the National Health Insurance (NHI) covered telemedicine in Taiwan, and publicly funded telemedicine was available in India and Nepal. In India, the central government funded the capital expenses for the national telemedicine (ie, Sanjeevani) while state governments managed the operational costs. Similarly, in Nepal, telemedicine services were supported through collaborative efforts between national and local governments to expand health care access in remote and underserved regions.

This review found multiple coverage aspects for telemedicine reimbursement that we categorized into telemedicine purposes, health conditions, patients’ nonhealth condition, service provider, participant interactions, interaction modes, and technology, which are described below.

Our reviews and interviews reported that reimbursable telemedicine services are often for diagnosis, treatment, follow-up, patient support, and management, indicating telemedicine’s critical role in direct clinical care [21-43]. Practices for these purposes were often associated with chronic conditions to provide ongoing care and prevention of disease progression, providing accessible care for patients with long-term health needs [21,22,24-26,28,29,33-35,42,43]. For example, the United States reimbursed remote care and referral coordination, medication management by phone call or secure SMS text message, self-care plan management, and other patient-provider communications through phone, text, and patient portals under the chronic care management program [22,33]. Telemedicine for prevention and promotion was reported to be reimbursed for patient self-care management [22] and school health–related programs [28]. Reimbursable telemedicine services for rural emergency departments in the United States also include the arrangement of interhospital transfer when necessary for patients with sepsis [32].

From this review, 10 papers specified patients’ nonhealth eligibility criteria for reimbursement [24,32,34,38,41,42,44-47]. Within these papers, residing in remote or underserved areas as a patient was one of the commonly mentioned eligibility criteria (n=3, 27%) in the United States [38,42,45]. However, the United States has modified its regulations to broaden the geographical scope for telemedicine reimbursement. The Bipartisan Budget Act of 2018 expanded Medicare coverage, allowing patients in urban areas to also access treatment for acute stroke [42]. Medicare also had strict requirements on the types of facilities for which the patients could receive telemedicine, which was that the patients had to be in providers’ offices, hospitals, critical access hospitals, rural health clinics, federally qualified health centers, hospital-based or critical access hospital-based renal dialysis centers, skilled nursing facilities, and community mental health centers when meeting with the distant providers [42].

Finally, eligibility could also extend to the patient’s status as new or existing to the health provider. In Taiwan and India, a prior relationship must have been established between the provider and the patient for subsequent follow-ups to be conducted using telemedicine. In other cases, telemedicine was only available to existing patients receiving care in the context of schizophrenia and related psychotic disorders [24].

A total of 23 (81%) sources reported the health workers who delivered reimbursable telemedicine services [22-37,39,40,42,43,46,48,49]. Only certain health workers were mentioned (one source mentioned more than one health worker), with physicians being the most frequent type that delivered reimbursed services (n=17, 74%) among these sources.

All jurisdictions from the literature and interviews reimbursed telemedicine when the interaction was between providers and patients. In the United States, Canada, and India, telemedicine between providers, such as consultations, referrals, and getting second opinions, was also reimbursed [37,42]. In Canada, provider-to-provider interaction was done through a secure messaging server where primary care doctors or nurses could request opinions from specialists [37]. In the United States, provider-to-provider consultation occurred from a hub to a spoke in remote regions. For example, an experienced emergency physician or nurse provides on-demand consultation to emergency departments in remote hospitals, particularly for patients with uncommon or severe conditions [32]. In India, this interaction meant that providers at the spokes (ie, state-level health facilities) provided assisted teleconsultation to the patients connected to the specialists at the hubs (ie, secondary or tertiary-level health facilities).

A total of 21 (68%) sources described the temporal modality reimbursed for telemedicine, with one source describing more than one modality [22-28,30-32,35,37,38,40-45,47,49]. Most of the interactions took place in real-time (n=19, 90%), using audio, video, or other equipment, much like a face-to-face visit, followed by store-and-forward, where information (ie, transmitting a digital image from one provider to another) was used later for diagnosis or treatment (n=6, 28%), and remote monitoring (n=6, 28%) where the provider had ongoing access to patients’ data to monitor trends or facilitate interventions. Medicare defined telemedicine as real-time interactive audio-visual communication between the provider and the patient, which meant store-and-forward and audio-only telemedicine would not be eligible for reimbursement [42].

A total of 19 (61%) sources described the technology in reimbursable telemedicine service, with more than one technology described in one source [22,26,27,29-33,35,37,38,40,42-44,48,50,51]. The most frequent technologies used were videoconferencing (n=15, 79%) followed by audio-only (ie, phone calls; n=14, 74%), software apps (n=3, 16%), messaging services (n=3, 16%), website (n=1, 5%), and the physical device (ie, implant device and insulin pumps; n=1, 5%). According to an interviewee, Taiwan allowed the use of commercial videoconferencing platforms and mobile apps, while India only provided free telemedicine services through eSanjeevani. In the province of Ontario, Canada, the Ontario Telehealth Network used a website-based platform to connect providers with patients from their home or community-based settings to conduct clinical visits (Table 4 and Multimedia Appendix 4).

The payment method is classified based on the unit of reimbursement (eg, payment per service item or fee-for-service [FFS]) [52], except for two sources that reported prospective payment systems (ie, based on relation to providers’ income and cost for providing service) [52] without providing more details [35,47]. A summary of payment methods, the jurisdictions that used these methods, the health conditions they addressed, their pros and cons, and the technologies eligible for reimbursement under each method is provided in Table 5 and Multimedia Appendix 5.

The most frequently reported method was the FFS model (n=19, 59%) used in the United States, Canada, Belgium, Estonia, Germany, the Netherlands, and Switzerland [22-24,29-34,36-38,41,43,45,46,48,50,51]. The FFS model refers to paying providers for each service delivered, such as consultation, tests, or procedures [50]. A similar model, payment per contact, was used in Sweden to incentivize the service volume [44]. The FFS payment method was said to encourage the roll-out of telemedicine because providers were incentivized to deliver more telemedicine, as per the stakeholders in Nepal and Taiwan.

A stakeholder, along with our review, reported the downside of FFS, noting that it could encourage the unnecessary overutilization of telemedicine. This overutilization could result in a financial burden for the payer without providing additional health benefits for patients [22,30].

The capitation method was reported in 2 (6%) sources [29,38], where providers receive per capita payment for the registered population during a specific period for a fixed reimbursement amount [29]. A similar model to capitation, payment per episode (ie, occurrence of symptoms of an illness) model, is reported to be used in the United States by home health agencies’ telemedicine services [49]. Providers need to manage patients’ episodes within the reimbursement formula and can increase their reimbursement amount with more patients [49].

The case-based method was reported to be used in some European countries for inpatient services [34]. The diagnostic-related groups system was reported to be used in these countries to classify the diagnoses [34]. A similar model to case-based is activity-based funding that was reported to be used in Australia [26] and Denmark [34]. In this model, providers are paid for hospital activities (eg, phone calls by pharmacists).

The blended method, where payers use more than one payment method for a service, was reported in the United States [40] and some European countries [34]. In the United States, the providers of the Comprehensive Primary Care Plus program received payment through FFS, a monthly prospective risk-adjusted per-beneficiary care management fee, and an annual prospective performance-based incentive payment to account for nontraditional visits such as telemedicine [40]. In Denmark, Estonia, Germany, and the Netherlands, primary care services received payment through a combination of FFS and capitation, while a combination of capitation, cost reduction, and outcomes-based incentives was applied for primary care services in Italy [34].

Another payment method is value-based, which refers to a payment method that ties the amount paid to the provider to the results they deliver. The review reported several types of value-based payment models: Accountable Care Organization (ACO) [21,28,43,48,50], quality-adjusted outcome capitation [34,50], and bundled payments [21,27,34,42,48]. Although three sources did not identify the type [25,34,39]. Four sources described, broadly, the use of value-based payment for telemedicine, which is said to reduce unnecessary treatments, promote the efficient use of resources, and lead to long-term cost savings [21,25,43,48].

The ACO method has only been applied in the US setting. In an ACO model, providers are accountable for costs and outcomes for their clinical decision-making [48]. Rambur et al [48] described three categories of ACO: risk-sharing ACOs, full risk-bearing ACOs, and upside-only ACOs. In the risk-sharing scheme, payers and providers benefit from savings on cost containment, receive extra payments if their spending is below a set benchmark, and also share any financial losses [50]. In the full risk-bearing scheme, providers have to manage their care delivery themselves, and revenue to the providers may be in a fixed global budget or a collective cap for a population or service [48]. In the upside-only scheme, providers benefit from savings by managing costs and are not liable for any costs that exceed projections [48].

Bundled payments are used in the United States [21,27,34,42,48] and the Netherlands [34]. Providers receive a lump sum payment to cover all services involved in an episode of patient care [34,48]. One advantage of this payment method was promoting coordinated care and linking reimbursement to quality and cost reduction outcomes, exemplified in the case of tele-intensive care unit, where a monthly bundled payment from Medicare and Medicaid covered the appropriate mix of critical care specialists, subspecialty nurses, pharmacists, and technical support personnel [27].

Most sources described the rate as a numeric value in currency (n=7, 23%), followed by a comparison between the rate of telemedicine and in-person equivalent (n=7, 23%), and some other ways (n=2, 6%). An example of rates in numeric value was exemplified by the cost of an initial assessment conducted by a dermatologist and an ophthalmologist to be 44.45 and 45.85 Canadian dollars, according to the Ontario Schedule of Benefits dated 2021 [37]. Equal payment for telemedicine and in-person visits was described using the term “payment parity,” implemented in many US states and Taiwan, and was said to be able to encourage providers to offer telemedicine [51].

The remaining two papers described rates in some other ways: as components of the rate (professional fee, facility fee, and technical fee) or copayments required for telemedicine visits [42,44]. According to a stakeholder, basing telemedicine reimbursement on relative value units could result in lower payments compared to in-person visits. This was because relative value units primarily accounted for the time and effort involved, and telemedicine services were often perceived as requiring less of both when compared to in-person visits. This could discourage providers from offering telemedicine services, highlighting a potential barrier to its adoption.

The reimbursement pathway refers to the process providers follow to obtain payment for their services, including steps from claim submissions to the final receipt of payment. This review found that the current procedural terminology codes were often required in the claims submission process to link the cost with services provided [22,23,31,33,38,41,42,45,46,50]. In Taiwan, the process began with the patient registering for the telemedicine service and verifying their eligibility for NHI coverage. Once eligibility is confirmed, the patient consults with a qualified health care provider who offers medical advice, diagnoses, and treatment recommendations. The health care provider then documented the service, including the diagnosis, treatment plan, and any prescriptions issued. Afterward, a claim is submitted to the NHI using standardized billing codes. The NHI reviews the claim for accuracy and completeness, and upon approval, disburses payment to the health care provider according to predetermined reimbursement rates.

This review did not find any information pertaining to fraud prevention for telemedicine reimbursement. Our stakeholder from Taiwan mentioned that telemedicine followed the same fraud prevention strategies as other reimbursable services billed to the NHI Administration. These included strict eligibility verification, where health care providers must confirm patient eligibility before delivering services using a comprehensive database to confirm that only eligible individuals receive reimbursement. Claims could undergo both automated checks and manual audits to identify inconsistencies or unusual billing patterns. According to the stakeholder, NHI Administration also used advanced data analytics and artificial intelligence (AI) tools to detect fraud by analyzing billing data and identifying high-risk profiles of providers and patients. However, we could not identify any information in the public domain to corroborate the use of AI in this mechanism. Additionally, a whistleblower protection program encouraged individuals to report fraudulent activities while protecting those who come forward. Providers or patients found engaging in fraud faced penalties, including fines, suspension of NHI reimbursements, and even criminal prosecution.