This retrospective cohort study was conducted at Kaiser Permanente Southern California (KPSC), an integrated health care system that provides comprehensive health care to over 4.7 million racially, ethnically, and socioeconomically diverse members at its 15 hospitals and 234 medical offices [13]. The prepaid health plan provides strong motivation for members to use services at KPSC facilities. KPSC’s electronic health record system stores information about all aspects of the care provided to members, including demographics, medical encounters, vaccinations, pharmacy utilization, membership history, and claims. The study population consisted of individuals of all ages who were KPSC members from January 5, 2020, to October 31, 2020 (week 1 to week 43 in 2020) and the corresponding period in 2019.

We identified outpatient visits and telehealth visits using electronic health record data and claims data. The outpatient visits included all in-person outpatient visits, while telehealth visits included telephone appointment visits and video visits that were conducted synchronously using real-time telephone or live video-audio interaction. These telehealth visits were billable and had a diagnosis or procedure code. Less than 6% of outpatient and telehealth encounters in KPSC were from claims [14]. The study was reviewed and approved by the KPSC Institutional Review Board.

We first described the characteristics of KPSC members enrolled in January 2020. The weekly (Sunday to Saturday) rates of outpatient visits, telehealth visits, and combined outpatient and telehealth visits were calculated during the study period by age (0-5 years, 6-17 years, 18-44 years, 45-64 years, and ≥65 years), sex, self-reported race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, non-Hispanic Asian, and other/unknown), and neighborhood-level median household income (<US $40,000, US $40,000-$79,999, and ≥US $80,000). In calculating a weekly rate, the numerator was the visit count of each visit type and the denominator was person-years of membership during a given week. The trend of visit rates during the pandemic year by subgroups was plotted separately for outpatient, telehealth, and combined outpatient and telehealth visits.

To assess the impact of the COVID-19 pandemic on the visit rate, we selected the following three periods in 2020: week 1-9 (January 5 to March 7, 2020) for the prepandemic period, week 12-16 (March 22 to April 25, 2020) for the early pandemic period immediately after California’s stay-at-home order, and week 40-43 (October 4 to October 31, 2020) for the later pandemic period after reopening. Three corresponding periods in 2019 were also identified and used to control for seasonality. We estimated the ratio of the visit rate during the pandemic period to the visit rate during the prepandemic period in Poisson regression models with the visit count as the dependent variable and natural logarithm of person-years of membership as an offset for each subgroup, adjusting for other subgroup variables and seasonality. The robust variance estimator was used to obtain standard errors of parameter estimates in Poisson regression models. After the adjusted rate ratio was estimated from Poisson regression models, the percentage change was calculated as follows: (adjusted rate ratio − 1) × 100%. We examined if the changes in visit rates differed by age, sex, race/ethnicity, and household income by comparing their 95% CIs.

Among the 4.56 million KPSC members enrolled in January 2020, 15.0% (n=682,947) were ≥65 years old, 51.5% (n=2,345,020) were female, 39.4% (n=1,795,994) were Hispanic, and 7.7% (n=350,721) lived in census tract areas with median household income <US $40,000 (Table 1).

Across all subgroups, outpatient visit rates had a steep drop-off after the start of the pandemic, reaching their lowest levels by early April 2020 (week 14), and then, they gradually increased but remained lower than prepandemic levels (Figure 1). Among age groups, the ≥65 years population had the smallest decrease in outpatient visits (adjusted percentage change: −29.6%, 95% CI −30.5% to −28.6%), with 6.25 visits per person-year during the later pandemic period versus 8.92 visits per person-year during the prepandemic period in 2020 (Table 2). Children aged 6-17 years had the largest decrease (−50.2%, 95% CI −51.6% to −48.7%) in outpatient visits, with 1.92 visits per person-year during the later pandemic period versus 3.86 visits per person-year during the prepandemic period in 2020. The decreases in outpatient visits did not differ statistically by sex, race/ethnicity, and household income groups, ranging from −83.8% to −80.7% when comparing the early pandemic period to the prepandemic period in 2020 and ranging from −38.8% to −35.9% when comparing the later pandemic period to the prepandemic period in 2020.

Across all subgroups, except children aged 6-17 years, telehealth visit rates increased sharply after the start of the pandemic and then slowly decreased (Figure 2). The telehealth visit rates in children aged 6-17 years remained stable after an initial increase during the pandemic. Comparing the early pandemic period to the prepandemic period in 2020, across age groups, the telehealth visits increased the most among children aged 0-5 years (475.5%, 95% CI 446.2%-506.4%) and the least among adults aged ≥65 years (236.6%, 95% CI 228.8%-244.5%) (Table 3). Telehealth visits increased significantly more among males (293.9%, 95% CI 283.4%-304.7%) than females (260.0%, 95% CI 252.6%-267.6%). Across racial/ethnic groups, Hispanic individuals had the largest increase in telehealth visits (295.5%, 95% CI 275.5%-316.5%). Across income levels, telehealth visits in the low-income group increased the most (313.5%, 95% CI 294.8%-333.1%). The patterns were similar when comparing the later pandemic period to the prepandemic period in 2020, but with smaller differences.

Despite increases in telehealth visits during the pandemic, across all subgroups, the rate of combined telehealth and outpatient visits still decreased during the early pandemic period and only returned to prepandemic levels for some subgroups (Figure 3). Comparing the later pandemic period to the prepandemic period in 2020, across age groups, the rate of combined outpatient and telehealth visits decreased the most among children aged 0-5 years (−19.7%, 95% CI −21.9% to −17.5%) (Table 4). The combined visit rate among adults aged 18-44 years during the later pandemic period exceeded the prepandemic level by 4.7% (95% CI 2.5%-6.8%). The combined visit rates returned to the prepandemic levels by October 2020 for the following subgroups: Hispanics (percentage change: −1.8%, 95% CI −6.1% to 2.6%), Non-Hispanic Black individuals (percentage change: 1.1%, 95% CI −3.0% to 5.3%), those with income <US $40,000 (percentage change: 0.7%, 95% CI −3.4% to 5.0%), and those with income US $40,000-$79,999 (percentage change: −2.1%, 95% CI −5.8% to 1.8%).

We examined the impact of the COVID-19 pandemic on outpatient and telehealth visits by demographic characteristics and household income levels in a large integrated health care system. We found that increases in telehealth visits during the COVID-19 pandemic varied across age, sex, race/ethnicity, and household income groups, while decreases in outpatient visits during the COVID-19 pandemic were similar across all subgroups, except age groups. The rates of combined outpatient and telehealth visits in the Hispanic, non-Hispanic Black, and low-income groups returned to prepandemic levels by October 2020.

While adults aged ≥65 years had the highest telehealth visit rates before and during the pandemic, this age group had the smallest percentage increase in telehealth visits during the pandemic among all age groups, suggesting that greater barriers existed for the senior population to further expand their use of telehealth services during the pandemic. For example, seniors may have limited access to digital technologies and may have difficulty in using technologies required for telehealth visits. Recent studies among Medicare beneficiaries showed that 26.3% of individuals lacked digital access at home, making it unlikely for them to have video visits with clinicians [15]. Furthermore, a recent study showed that difficulty hearing and problems of speaking or making oneself understood were the top two reasons for hesitancy to use telehealth [16].

Members living in an area with median household income <US $40,000 increased their telehealth visits per person-year from 0.82 during week 1-9 to 3.38 during week 12-16, a higher percentage increase than that for those living in an area with median household income ≥US $80,000 (from 0.92 during week 1-9 to 3.22 during week 12-16). The KPSC service area covers mostly nonrural residents whose access to a telephone is not a barrier for using telehealth services. Following the CDC’s recommendation to delay elective and nonemergent care on March 18, 2020, KPSC offered zero co-payments for telehealth visits, which might have been especially helpful for the low-income population. In addition, telehealth visits provide flexibility in care location, and save time and cost associated with commuting for in-person care.

Hispanic individuals had the greatest percentage increase in telehealth visits across racial/ethnic groups, although their visit rates during the prepandemic and pandemic periods were lower than those of non-Hispanic Black and non-Hispanic White individuals. A study conducted in New York City during the pandemic showed that Hispanic and Black patients had lower odds of using telehealth visits versus office visits than either White or Asian patients, but the authors did not consider the trend of telehealth utilization [17]. A recent study based on survey data showed that Black respondents were more likely than White respondents to report using telehealth because of the pandemic [18]. The fact that a larger proportion of racial/ethnic minorities live in low-income areas, along with KPSC’s zero co-payment policy for telehealth visits, might support telehealth care among Hispanic individuals. In addition, telehealth is accessible for members with limited English proficiency. Spanish is a common language in southern California, spoken by many providers. The removal of Spanish language barriers could have encouraged Hispanic individuals to continue using telehealth care and contributed to larger increases of telehealth care in this population.

The impact of the COVID-19 pandemic on telephone and video visits may differ because the barrier to technology use is higher for video than telephone [12]. We assessed the rate of telehealth visits within KPSC separately by telephone and video visits (Multimedia Appendix 1). The trend of telephone visits was similar to the trend of overall telehealth visits, with the rates increasing sharply after the start of the pandemic and then gradually decreasing. The use of video visits was very low before the pandemic, with only 2.1% of telehealth visits occurring within KPSC being video visits. The video visit rate increased gradually over time after the start of the pandemic, contributing to 24% of telehealth visits by the end of October 2020. The slower and more gradual increase in video visits was likely due to technology barriers, for example, need for a web camera and digital literacy to conduct a video visit. These barriers were more prominent in those aged ≥65 years (Multimedia Appendix 2), with video visits having a slower start compared to that in some other age groups. In addition, it took time for both members and providers to adopt this new approach, with gradual increases in video visits occurring across all subgroups. Given the growing familiarity with new telehealth technologies, even as the COVID-19 pandemic wanes, patients likely will continue to pursue telehealth services beyond pre-COVID-19 levels.

There are some potential limitations in this study. First, while we studied the impact of the pandemic on health care utilization among subpopulations, we did not address the quality of care and population health. Second, we assessed members’ neighborhood-level income based on their address, which is less accurate than individual-level income. Third, because KPSC’s health plan, hospitals, and medical groups are integrated to create a system for coordinated and comprehensive patient care, our patients may have better access to care than the general population. Therefore, these results may not be generalizable to populations in less integrated health care settings. Finally, we focused on relative change (ie, percentage change from baseline) in utilization as part of this study, but we acknowledge that absolute change (ie, difference from baseline) in utilization may be of interest as well. As such, we have presented absolute visit rates in each time period by subgroup to allow for such comparisons to be made.

While the ≥65 years population had the least percentage increase in telehealth visits, the Hispanic and low-income groups had the greatest percentage increase in telehealth utilization, possibly due to the low cost and flexibility of location and time. The use of virtual care in these subgroups reduced exposure to COVID-19, reinforced social distancing protocols, increased the number of patients who could be treated, and potentially mitigated the impact of COVID-19 in vulnerable populations. Furthermore, the use of virtual care might be a promising way to reduce health care disparities even after the COVID-19 pandemic ends.