Practice-Level Variation in Telemedicine Use in a Pediatric Primary Care Network During the COVID-19 Pandemic: Retrospective Analysis and Survey Study

Background: Telehealth, the delivery of health care through telecommunication technology, has potential to address multiple health system concerns. Despite this potential, only 15% of pediatric primary care clinicians reported using telemedicine as of 2016, with the majority identifying inadequate payment for these services as the largest barrier to their adoption. The COVID-19 pandemic led to rapid changes in payment and regulations surrounding telehealth, enabling its integration into primary care pediatrics. Objective: Due to limited use of telemedicine in primary care pediatrics prior to the COVID-19 pandemic, much is unknown about the role of telemedicine in pediatric primary care. To address this gap in knowledge, we examined the association between practice-level telemedicine use within a large pediatric primary care network and practice characteristics, telemedicine visit diagnoses, in-person visit volumes, child-level variations in telemedicine use, and clinician attitudes toward telemedicine. Methods: We analyzed electronic health record data from 45 primary care practices and administered a clinician survey to practice clinicians. Practices were stratified into tertiles based on rates of telemedicine use (low, intermediate, high) per 1000 patients per week during a two-week period (April 19 to May 2, 2020). By practice tertile, we compared (1) practice characteristics, (2) telemedicine visit diagnoses, (3) rates of in-person visits to the office, urgent care, and the emergency department, (4) child-level variation in telemedicine use, and (5) clinician attitudes toward telemedicine across these practices. Results: Across pediatric primary care practices, telemedicine visit rates ranged from 5 to 23 telemedicine visits per 1000 patients per week. Across all tertiles, the most frequent telemedicine visit diagnoses were mental health (28%-36% of visits) and dermatologic (15%-28%). Compared to low telemedicine use practices, high telemedicine use practices had fewer in-person office visits (10 vs 16 visits per 1000 patients per week, P=.005) but more total encounters overall (in-office and telemedicine: 28 vs 22 visits per 1000 patients per week, P=.006). Telemedicine use varied with child age, race and ethnicity, and recent preventive care; however, no significant interactions existed between these characteristics and practice-level telemedicine use. Finally, clinician attitudes regarding the usability and impact of telemedicine did not vary significantly across tertiles. Conclusions: Across a network of pediatric practices, we identified significant practice-level variation in telemedicine use, with increased use associated with more varied telemedicine diagnoses, fewer in-person office visits, and increased overall primary care encounter volume. Thus, in the context of the pandemic, when underutilization of primary care was prevalent, higher practice-level telemedicine use supported pediatric primary care encounter volume closer to usual rates. Child-level telemedicine use differed by child age, race and ethnicity, and recent preventive care, building upon prior concerns about differences in access to telemedicine. However, increased practice-level use of telemedicine services was not associated with reduced or increased differences in use, suggesting that further work is needed to promote equitable access to primary care telemedicine.