Use and Content of Primary Care Office-Based vs Telemedicine Care Visits During the COVID-19 Pandemic in the US

Key Points Question Is there a quantifiable association between the coronavirus disease 2019 (COVID-19) pandemic and the volume, type, and content of primary care encounters in the US? Findings In this cross-sectional analysis of the US National Disease and Therapeutic Index audit of more than 125.8 million primary care visits in the 10 calendar quarters between quarter 1 of 2018 and quarter 2 of 2020, primary care visits decreased by 21.4% during the second quarter of 2020 compared with the average quarterly visit volume of the second quarters of 2018 and 2019. Evaluations of blood pressure and cholesterol levels decreased owing to fewer total visits and less frequent assessment during telemedicine encounters. Meaning The COVID-19 pandemic was associated with changes in the structure of primary care delivery during the second quarter of 2020, with the content of telemedicine visits differing from that of office-based encounters.


+ Invited Commentary + Supplemental content Introduction
Since February 2020, the coronavirus disease 2019 (COVID- 19) pandemic has been associated with more than 4.4 million cases and 150 000 deaths in the US, as well as widespread social and economic changes. 1 While the ultimate health care system impacts of the pandemic remain uncertain, many early health care consequences associated with the pandemic have been noted, ranging from postponement of elective care to permanent clinic and hospital closures. 2,3 Early reports suggested that substantial increases could be expected in the delivery of telemedicine, or remote clinical services, during the first few months of the pandemic in the US, 4,5 owing to concern regarding the potential for workplace transmission of COVID-19, the implementation of social distancing policies, and the redeployment of health care personnel. A more recent update indicated that the delivery of telemedicine increased during mid-April and has since subsided modestly, although levels remain substantially higher than before the pandemic. 6 These changes, which have been accompanied by changes in federal 7,8 and state 9,10 guidance and reimbursement, have occurred in the context of structural and social factors 11,12 hindering widespread telemedicine adoption.
Investigations of telemedicine during the pandemic, while yielding insights, have generally been based on small or nonrepresentative samples and limited to analyses of the frequency of such encounters rather than descriptions of their content. [4][5][6]13 We quantified national changes in the volume and type of primary care associated with the COVID-19 pandemic. In addition to characterizing blood pressure and cholesterol measurement and initiation or continuation of prescription medicines for hypertension and dyslipidemia, we explored variance in telemedicine use across different patient populations and geographic regions of the US.

Data
We used the IQVIA National Disease and Therapeutic Index to conduct a cross-sectional analysis, focusing on the period from the first quarter (Q1) of 2018 through the second quarter (Q2) of 2020.
The National Disease and Therapeutic Index is a nationally representative audit of outpatient practice in the US. 14,15 Other studies [16][17][18] have compared the National Disease and Therapeutic Index with the National Ambulatory Medical Care Survey, a nationally representative audit conducted by the National Center for Health Statistics, and found that the surveys yielded substantively comparable estimates of outpatient care. Using a 2-stage, stratified sampling design, the National Disease and Therapeutic Index audit is based on a sample derived from the American Medical Association and the American Osteopathic Association. 19 20 We restricted our analysis to primary care visits, defined as those accounted for by the fields of internal medicine, pediatrics, geriatrics, general practice, and family practice. For cases in which we specifically assessed pharmaceutical prescribing (eg, initiations of pharmacologic therapy for specific diseases), we examined treatment visit, defined as a patient encounter for a specific diagnosis in which a pharmacologic treatment was initiated or continued. Except where depicted otherwise, we excluded the approximately 3% to 4% of hospital-based visits and 1% to 2% of visits taking place in other settings (home, nursing home, and unspecified sites of care) each calendar quarter.
Our analysis of telemedicine use across geographic regions included an examination of how such use varied by COVID-19 burden, expressed as the rate of COVID-19 fatalities per 100 000 individuals. To estimate this rate, we summed COVID-19 deaths within each national region as of July 28, 2020, 21 and divided this number by the total population within that region. 22

Statistical Analysis
We used descriptive statistics to perform our analysis. Our main outcomes were visit type, assessment of blood pressure or cholesterol measurement, and initiation or continuation of prescription medicines. First, we extracted the total number of visits between January 1, 2018, and June 30, 2020, and plotted these numbers over time to examine general trends and to assess for inflection points and outliers. Next, we aggregated these visits by calendar quarter. We then characterized the distribution of visits, stratified by encounter type, across visit characteristics of interest, such as patient age, sex, race, and type of insurance. We limited our analysis of telemedicine use by race to Black race and White race given racial disparities in health care. We calculated a Pearson correlation coefficient to examine the association between telemedicine use and COVID-19 burden across 8 geographic regions. We used standardized errors to estimate 95% CIs. Analyses used 2-tailed, unpaired testing. A threshold of P < .05 was used to establish statistical significance.
Statistical analyses were conducted using Stata software, version 15 (StataCorp LLC).

JAMA Network Open | Health Policy
of 2020 and an average of 57.3% of telemedicine visits for Q1 and Q2 of 2020.    One reason for this finding was that assessment of blood pressure and cholesterol was statistically significantly less common among telemedicine than among office-based visits. For example, of 58.7 million (95% CI, 55.3-62.1 million) Q2 2020 office-based visits, 69.7% had a blood pressure recorded compared with 9.6% of telemedicine visits during the same time period

Discussion
While the COVID-19 pandemic has impacted health care delivery in many ways, little is known regarding how the volume, site, and content of primary care in the US has changed. We used a nationally representative audit of outpatient care to characterize primary care delivery in the US   between 2018 and Q2 of 2020. The pandemic has been associated with a more than 25% decrease in primary care volume, which has been offset in part by increases in the delivery of telemedicine, which accounted for 35.28% of encounters during the second quarter of 2020. Despite the increased use of telemedicine, its uptake has varied across the continental US and has not been correlated, at a regional level, with COVID-19 burden. Overall, the pandemic has been associated with marked reductions in the primary care assessment of cardiovascular risk factors such as blood pressure and cholesterol levels, owing to decreased total visit volume and less frequent assessment during telemedicine visits than during office-based visits. These findings are notable because little is known about the association between primary care delivery and the COVID-19 pandemic and because the pandemic has generated interest in telemedicine as a means to safely deliver primary care.
Our analysis was based on an assessment at a single point, and the degree to which the COVID-19 pandemic may be associated with permanent increases in the use of telemedicine remains to be seen. Historically, limited reimbursement, interstate licensure requirements, and patient and clinician factors have slowed the uptake and adoption of telemedicine. 12,23 In response to the pandemic, US federal and state agencies and other stakeholders have modified policies and procedures, such as the Centers for Medicare & Medicaid Services provision of telehealth waivers for providers, 24 to allow greater use of telemedicine to support remote clinical encounters. 25, 26 We did not find a correlation between regional COVID-19 burden and telemedicine adoption, suggesting that other factors may account for regional differences in the uptake of this mode of health care. In addition, whether the federal and state rules and regulations that have been modified will be made permanent and whether the current embrace of telemedicine by patients and clinicians will endure remain unknown. 27 If substantial primary care volume continues to be delivered using telemedicine, a focus on the content and quality of such encounters is inevitable. 28 Despite findings in a systematic review of 86 articles demonstrating the feasibility and acceptance of telemedicine for use in primary care, to our knowledge, relatively few rigorous comparisons of clinical outcomes in office-based vs telemedicine encounters have been performed. 29 Our finding that such visits were less likely to include blood pressure or cholesterol assessments underscores the limitation of telemedicine, at least in its current form, for an important component of primary care prevention and chronic disease management.
Middle-aged individuals and those who were commercially insured were more likely to adopt telemedicine during the pandemic than their counterparts with other or no insurance. This difference may be due in part to the perceived elective or deferrable nature of visits among children 30 and greater familiarity with telemedicine technology among middle-aged than among older adults. 31 We did not find substantial differences in telemedicine use by payer type, and, contrary to our expectations and evidence of a digital divide, 32 we did not find evidence of a racial disparity in telemedicine use when examining the frequency of telemedicine encounters as a proportion of a patient visits among Black versus White individuals.

Limitations
This study has limitations. First, both the COVID-19 pandemic and health system response continue to evolve, and our analyses reflect the provision of care at a single point in time. Second, as with any outpatient audit, the data that we used were subject to measurement error, although prior analyses have yielded estimates comparable to those from the National Ambulatory Medical Care Survey.
Third, factors such as patient race are complex multidimensional constructs, and our assessment of race in this context provides a limited window through which to understand how telemedicine adoption may vary across different populations. 33,34 Fourth, our data did not allow us to examine more granular geographic associations between COVID-19 burden and telemedicine adoption. Last, we considered the assessment of 2 important cardiovascular risk factors, but many other dimensions of primary care might also be compared between office-based and telemedicine encounters, and we