Institutional Barriers to the Successful Implementation of Telemedicine for Type 1 Diabetes Care Free

The T1DX-QI is a multicenter quality improvement collaborative focused on improving outcomes for patients with type 1 diabetes and has performed ongoing work in the area of telemedicine adoption since the start of the pandemic in March 2020. We assessed overall telemedicine rates (visits that occurred using video-conferencing software in lieu of a medical visit) for centers through electronic health record (EHR) data and clinic surveys. For at least 15 centers (12 pediatric and 3 adult care clinics) who had EHR data, we captured the overall proportion of total visits conducted as telemedicine encounters and calculated the percent change in monthly telemedicine rates from the start of the pandemic through September 2021. In addition, a center-level survey was administered from September to November 2021 and was completed by 33 centers (inclusive of the 15 centers with EHR data), which reported telemedicine rates, telemedicine goals, and key drivers for supporting telemedicine in the clinical delivery system.

We generated and assigned a capacity score to each center based on the following survey responses, which we identified as key components for supporting a telemedicine visits: having a defined telemedicine workflow (yes = 1, no = 0), having staff support assigned to telemedicine visits as for regular clinic visits (yes = 1, no = 0), having support for uploading data to the EHR that did not require providers’ involvement (automated EHR uploading or staff uploading = 1, provider uploading = 0), and having an institutional telemedicine goal regarding the proportion of visits seen via telemedicine (yes = 1, no = 0). We calculated a total capacity score by adding scores for all four capacity elements.

Finally, we compared center-level capacity scores by overall telemedicine rates across centers. Because of the limited number of centers, we elected not to perform statistical analyses, but rather provided summary statistics. This study was approved by the Western Institutional Review Board, and participating clinics also obtained their own approvals per institutional policy.

For all centers that contributed data from the EHR, the combined monthly rate of telemedicine visits was 37.6% (n = 2,294) in September 2020, and this rate decreased to 20.2% (n = 1,180) by September 2021, which represents an overall decline of 48.5%. Figure 1 shows the monthly proportion of telemedicine visits, the monthly number of telemedicine visits, and the overall percent change in telemedicine visits for each center. All centers except for one experienced a consistent decline in telemedicine visits over the period. Despite this consistent decline, telemedicine rates varied widely among centers, ranging from 2.9 to 63.6%.

Of the 33 centers surveyed, 24 were pediatric and 9 were adult care providers. The majority of centers reported conducting 11–25% of visits as telemedicine encounters in the fall of 2021, with a greater proportion of centers performing more (≥26%) than fewer (0–10%) (Figure 2A). Figures 2B–E show that the majority of centers had a workflow, support for data uploading, and a telemedicine goal, but a minority had assigned staff dedicated to telemedicine. Only 12% of centers reported having all four capacity components, but a majority of centers had a telemedicine capacity score of at least two components (Figure 2F). Figure 2G shows the average individual and total telemedicine capacity scores according to the telemedicine rates reported by the centers. Generally, the lower the average score was, the lower the rate of telemedicine was, although a statistical test was not conducted.

Table 1 shows the breakdown of telemedicine rates and telemedicine capacity components by pediatric and adult care centers based on survey responses. The majority of centers provided institutional support for participation of multidisciplinary providers in clinics for both adult and pediatric centers, although the deployment of these individuals differed by clinic type; in pediatric clinics, the interaction of these providers was mixed, integrated with the provider telemedicine visit, or occurring separately, but the majority of adult clinics did not have integrated visits.

Centers did report a number of barriers impeding the sustainability of telemedicine over the long term, including limited Internet access for patients (94%); health disparities, including poverty, geographical isolation, and limited access to services as assessed by the provider (91%); limited access to patient data (88%); limited access to laboratory test results (64%); reimbursement issues (48%); and licensing laws (42%). Other reported barriers included patients not having the required technology and patient preferences.

This is one of the first studies to describe telemedicine rates for individuals with type 1 diabetes for a large and geographically diverse sample of endocrinology centers up to 18 months after the start of the COVID-19 pandemic. Given the initial near-total adoption of telemedicine during the peak of the pandemic (95.2%), it was expected that rates would drop significantly as patients would likely seek in-person care after a long period of only virtual contact with their providers and care teams. Our data show a telemedicine rate of 20% by September 2021, which is higher than the pre-pandemic level of <1% (1). However, our capacity analysis shows that key institutional barriers remained for many centers, even 18 months after the start of the pandemic, and these barriers affect the long-term sustainability of telemedicine.

Given that only four of 33 centers (12%) reported having all four capacity elements, it is clear that more work is needed to improve the operational capacity of centers to perform telemedicine. Only a minority of centers reported having staff partially or fully dedicated to telemedicine. When patients are seen for in-person clinic visits, a medical assistant typically prepares them for the visits by navigating their entry into a room, capturing their vital signs, reviewing medications, and downloading data from their diabetes devices. However, this type of support is not provided for telemedicine visits in many centers. If providers run late or patients have difficulty connecting to a virtual visit or downloading data, both must navigate these obstacles independently, which can lead to delays, lower the quality of telemedicine visits because of a lack of access to diabetes data, and incur financial costs because of missed or canceled visits. Health care staffing models have traditionally been constructed to support in- person care; new models of operational staffing to directly support telemedicine will be critical for the ultimate sustainment of telemedicine activities.

When we first reported telemedicine statistics for the T1DX-QI for September 2020, just one of the 21 centers (<5%) assessed had automated integration of data downloading into the EHR. As of the fall of 2021, just three centers (∼10%) reported having automated downloading, which is an improvement, but still represents only a minority of centers and speaks to the significant technical challenges of integrating diabetes device data into the EHR. In 2020, in 43% of centers, providers were either solely (10%) or partially (33%) responsible for capturing data downloads, but, by 2021, in 36% of centers, providers were either solely (3%) or partially (33%) responsible for this task. Access to diabetes data are the linchpin of telemedicine; therefore, technology to support the integration of data into the EHR and improved telemedicine workflows supporting providers and patients are crucial.

We are unaware of other studies that have captured telemedicine rates for diabetes care in such a large and comprehensive sample of centers for up to 18 months post-pandemic, nor are we aware of studies that have used EHR data to capture longitudinal monthly statistics on telemedicine rates in a cohort of centers. The majority of centers are located in urban settings. The COVID-19 Healthcare Coalition evaluated telemedicine trends between January 2019 and March 2021 using health care claims and reported a decline in the proportion of individuals with diabetes who participated in telemedicine from 27% in the second quarter of 2020 to 13% by March 2021 (2). We recognize that the end date of the current study was still relatively early in the pandemic, and this study did not focus on diabetes centers who see a large population of individuals with type 2 diabetes. Additional studies of telemedicine adoption reported on rates during the early stages of the pandemic (3), were single-center studies (4), focused on populations outside of the United States (5), or simply described the characteristics of telemedicine users without providing specificity to the center-level components of diabetes-related telemedicine care (6).

Because this was a center-level analysis, we could not assess key patient-level barriers to telemedicine such as lack of insurance coverage for telemedicine visits, lack of access to the Internet, lack of access to the online patient portal, lack of access to diabetes devices, or inability to download diabetes data as potential factors that may affect overall telemedicine rates. Furthermore, we did not examine providers’ and patients’ preferences regarding telemedicine, which are also important factors. Haynes et al. (6) surveyed a small sample of patients (n = 53) at a single center who had an in-person rather than a telemedicine visit during the start of the pandemic, which may have been a biased sample in that the study was targeted to individuals without a telemedicine visit. Still, individuals and families identified a variety of factors, including the belief that in-person care was of higher quality than care using telemedicine, lack of familiarity with technology, and a lack of smartphone access. Conversely, Crossen et al. (7) surveyed a panel of patients in the T1D Exchange patient registry and online community regarding their experiences with and opinions about telemedicine care during the pandemic. More than 60% of the 2,235 individuals who responded had used telemedicine; of these individuals, 62% felt telemedicine care was as effective as or was more effective than in-person care, and >80% wished to use telemedicine into the future. However, notably, the most common reason for not using telemedicine was that the providers were not offering it (49%), which indicates the need for further research focused on providers’ preferences and/or clinic resources for telemedicine. Additionally, Tanenbaum et al. (8) found telehealth to be used successfully for diabetes technology onboarding, and parents reported that they believed telehealth should be an option for all families.

Telemedicine rates for type 1 diabetes care decreased substantially from 2020 to 2021 across all participating centers but were still higher than in the pre-pandemic period. Centers with higher rates of telemedicine had a higher average number of center-level capacity components, suggesting the need for improved structural components to ensure telemedicine capacity and sustainment of telemedicine utilization.

The authors thank the Leona M. and Harry B. Helmsley Charitable Trust for funding the T1DX-QI. The authors also thank all team members and centers in the T1DX-QI.

J.M.L. is on the GoodRx medical advisory board and is a consultant for Tandem. O.E. is a Health Equity Advisory Board member for Medtronic Diabetes. F.V. is a consultant for Eli Lilly. No other potential conflicts of interest relevant to this article were reported.

J.M.L. wrote the manuscript and researched data. J.M.L., E.O., and N.N. analyzed the data and reviewed/edited the manuscript. A.M., O.E., M.G., F.S.M., N.R.F., S.A., S.H., M.W., A.N., and F.V. contributed to the discussion and reviewed/edited the manuscript. J.M.L., O.E., and F.V. conceptualized the study. E.O. is the guarantor of this work and, as such, had access to all of the data and take responsibility for the integrity of the data and the accuracy of the data analysis.

Portions of the content of this article were previously presented in abstract form at the American Diabetes Association’s 82nd Scientific Sessions in New Orleans, LA, 3–7 June 2022.