Percentage of total visits delivered by telemedicine is based on visits during the COVID-19 pandemic (defined here as March 18 to July 14, 2020). Quintiles based on the distributions across the 2800 US counties in our sample.
eAppendix. Data Sources and Study Sample
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Patel SY, Rose S, Barnett ML, Huskamp HA, Uscher-Pines L, Mehrotra A. Community Factors Associated With Telemedicine Use During the COVID-19 Pandemic. JAMA Netw Open. 2021;4(5):e2110330. doi:10.1001/jamanetworkopen.2021.10330
While telemedicine use grew rapidly during the COVID-19 pandemic, there was substantial geographic variation in uptake.1,2 What drives this variation is unclear. To understand drivers of telemedicine use, we examined the association of county-level telemedicine use with community factors among individuals with commercial or Medicare Advantage insurance.
Using OptumLabs Data Warehouse deidentified medical claims, we captured all outpatient visits, in person and telemedicine, from January 1 to July 14, 2020. We defined outpatient visits using Medicare’s list of Current Procedural Terminology (CPT) codes eligible for telemedicine.3 Telemedicine visits were those with modifier codes GT, GQ, or 95, or CPT codes 99441 to 99443 with the remainder defined as in-person visits.
After excluding counties with fewer than 100 members enrolled from July 2019 through July 2020, we assessed county-level telemedicine use during the pandemic (defined as March 18 to July 14, 2020) using the percentage of total visits delivered via telemedicine. We measured county-level population density (number of people per square mile) as a proxy for rurality, race, income, broadband availability, and per capita number of hospital beds, community health centers, physicians, and advanced practice registered nurses using data from the US Census and Area Health Resource Files.4 County-level prepandemic telemedicine use was measured as the percentage of total visits delivered via telemedicine from January 1 to March 17, 2020, and cumulative COVID-19 cases per capita were measured over 2 periods, March 18 to April 16, 2020, and March 18 to July 14, 2020.
To measure the association between these county characteristics and percentage point differences in telemedicine use, we used targeted maximum likelihood estimation, a machine learning procedure used in prior clinical literature.5 We divided county characteristic measures into tertiles for meaningful interpretation except for those where there were positivity violations. We addressed violations by dichotomizing variables (see eAppendix in the Supplement). The Harvard Medical School institutional review board exempted this study from review because it involved the study of data recorded such that participants cannot be reasonably identified. This study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for cross-sectional studies.
Data analyses were performed using R statistical software version 3.6.3 (R Project for Statistical Computing) with the SuperLearner and tmle packages.5 These analyses were performed on September 17, 2020.
Across 2800 counties with at least 100 enrollees, the mean (SD) and median (interquartile range [IQR]) percentage of total visits delivered via telemedicine were 19.6% (8.3%) and 18.6% (13.3%-24.7%), respectively. The median (IQR) percentage of telemedicine visits in the 5 quintiles of counties was 10.0% (8.0%-11.2%) for quintile 1, 14.3% (13.3%-15.4%) for quintile 2, 18.6% (17.6%-19.8%) for quintile 3, 23.2% (22.1%-24.9%) for quintile 4, and 31.0% (28.5%-35.2%) for quintile 5. (Figure).
We observed less telemedicine use in counties with lower median income (difference of below vs above median: 2.5%; 95% CI, 2.2%-2.7%), lower population density (difference of lowest vs highest tertile: 7.5%; 95% CI, 7.2%-7.8%), less broadband availability (difference of below vs above median: 1.6%; 95% CI, 1.3%-1.8%), less prepandemic telemedicine use (difference of highest vs lowest tertile: 2.6%; 95% CI, 2.2%-2.9%), and fewer COVID-19 cases per capita in the first 30 days of pandemic (difference of highest vs lowest tertile: 2.8%, 95% CI, 2.5%-3.2%) (Table). We did not observe an association between telemedicine use and COVID-19 incidence over a longer period through July 14.
Through July 14, 2020, we observed substantial variation across counties in telemedicine use. Our results support concerns that rural and lower-income communities may be left behind in the shift to telemedicine use. To ensure telemedicine is accessible by all people in the US, interventions such as increased broadband investment in rural areas6 or greater reimbursement in disadvantaged communities may be needed.
Another key factor associated with less telemedicine use during the pandemic was prepandemic telemedicine use. Health care practitioners with the necessary technology and familiarity with technology may have been able to quickly increase telemedicine use.
Telemedicine adoption was also associated with incidence of COVID-19 over the first 4 weeks of the pandemic, but not incidence over a longer 16-week period. Possibly initial uncertainty about the virus in higher incidence areas drove uptake, and the urgency to use telehealth waned somewhat over time as communities became acclimated to the pandemic. A key limitation is that results may not generalize to other commercially insured populations and those with Medicare or Medicaid insurance.
Accepted for Publication: March 24, 2021.
Published: May 18, 2021. doi:10.1001/jamanetworkopen.2021.10330
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Patel SY et al. JAMA Network Open.
Corresponding Author: Ateev Mehrotra, MD, MPH, Department of Health Care Policy, Harvard Medical School, 180A Longwood Ave, Boston, MA 02115 (firstname.lastname@example.org).
Author Contributions: Dr Patel had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Patel, Rose, Barnett, Mehrotra.
Acquisition, analysis, or interpretation of data: Patel, Rose, Huskamp, Uscher-Pines, Mehrotra.
Drafting of the manuscript: Patel.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Patel, Rose, Barnett.
Obtained funding: Uscher-Pines, Mehrotra.
Administrative, technical, or material support: Patel, Barnett, Huskamp, Mehrotra.
Supervision: Patel, Rose, Huskamp, Uscher-Pines.
Conflict of Interest Disclosures: None reported.
Funding/Support: This project was supported by the Commonwealth Fund, National Institute on Aging of the National Institutes of Health (K23 AG058806-01) and the National Institute of Mental Health (R01 MH112829-01, T32MH019733).
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank Rebecca Shyu, University of Missouri, for contributing to data cleaning, data analysis, and manuscript preparation efforts. She was not compensated.
Create a personal account or sign in to: