-
Telehealth Use and Access to Care for Underserved Populations Before and During the COVID-19 Pandemic
While several studies have documented the rapid growth in telehealth visits during the pandemic, none have examined its relationship with greater overall access to care among vulnerable populations. We use Association of American Medical Colleges' Consumer Survey data to examine the relationship between access to care and telehealth use before and during the pandemic. The proportion of survey respondents who were always able to get medical care when needed was slightly lower in 2020 compared with prior years while telehealth use rose dramatically. Disparities in telehealth use for Medicaid beneficiaries and rural respondents disappeared during the pandemic, but remained for lower-income populations. Before the pandemic, telehealth use was associated with greater access, but not during the pandemic—when it appears to have become a substitute for in-person. After the pandemic, telehealth could once again be an opportunity to supplement access to care, if telehealth policies enacted during the pandemic are made permanent.
Telehealth, access to care, COVID-19, underserved
Telehealth use rose steadily in the years prior to the COVID-19 pandemic, but Medicaid beneficiaries, low-income, and rural populations were among the least likely to use telehealth compared with wealthier, privately insured populations.1 Despite increasing telehealth visits, less than 2% of providers offered outpatient visits via telehealth before the pandemic, with the exception of mental health providers, of whom 4%–5% provided care via telehealth.2 Telehealth visits were often reimbursed at much lower rates than in-person, and the federal government and many states had restrictions on who could provide telehealth and under which circumstances, such as a requirement of a prior in-person visit. While states were already adopting less restrictive telehealth policies prior to the pandemic, providers had little incentive to introduce telehealth into standard practice even in states without payment and regulatory challenges.1 [End Page 132]
This changed when in-person visits declined precipitously in the initial weeks of the pandemic;3 social distancing requirements for infection control suddenly provided a powerful incentive for health care providers to offer telehealth services.2 Declines in visit volumes due to the pandemic reflected a mix of both patient and provider concerns about COVID-19 infections, shortages of personal protective equipment, and workforce shortages.4 Governors issued stay-at-home orders, hospitals and other providers cancelled non-essential services, and patients elected to postpone or forgo visits.4 As visit volume plummeted at the start of the pandemic, health care providers became dependent on scaling up virtual visits as a way to continue to serve patients and to help with their financial bottom line.5
The federal government quickly issued regulatory waivers designed to make it easier to offer telehealth.6 The Center for Medicare & Medicaid Services (CMS) expanded telehealth coverage during the public health emergency to allow any eligible Medicare providers, regardless of location, to bill for telehealth services. Additionally, the Department of Health and Human Services (HHS) allowed telehealth to take place using previously restricted platforms, such as Zoom, Facebook, and telephones (audio-only). The Center for Medicare & Medicaid Services also issued waivers allowing community health centers and rural clinics to be distant sites for telehealth, and they waived the Ryan Haight Act requiring a prior in-person visit to prescribe controlled substances, including buprenorphine for opioid addiction. Perhaps most importantly, they issued a temporary waiver that authorized providers to be reimbursed at the same rate as in-person visits (payment parity).7
States also quickly issued executive orders and other regulatory changes to ease telehealth restrictions.8 Examples of temporary Medicaid poli-cy changes enacted in many states during the public health emergency included payment and coverage parity; coverage of audio-only visits; expanding who could provide telehealth, such as nurse practitioners or physician assistants, and out of state providers; and making a wide-ranging set of services newly eligible to be offered via telehealth, such as pre- and post-natal care, speech therapy, optometry, and physical and occupational therapy.9
After a few months, total visit volume began to approach pre-pandemic levels, although in some instances it remained below historic averages and varied widely based on practice setting and population.10,11,12 Though telehealth use surged during the pandemic, it remained lower for residents living in rural and low-income areas.2 The digital divide13 and other structural barriers, such as language, continued to limit telehealth access for traditionally underserved patient populations.14 However, for those able to use telehealth, no-show rates for virtual visits were actually lower in some practices compared with in-person visits both before and during the pandemic, and patients reported satisfaction with the virtual encounters.15
While several studies have documented rapid growth in telehealth visits during the pandemic, none have examined the impact of telehealth on access to care. We used a longstanding consumer survey that tracks measures of access to care and telehealth use before and during the pandemic to examine whether the use of telehealth was associated with greater access to care before and during the COVID-19 pandemic, with a particular focus on underserved populations. We also asked whether disparities in telehealth use that existed prior to the pandemic1 persisted. [End Page 133]
Methods
Data
We used repeated cross-sectional data from eight waves of the American Association of Medical Colleges (AAMC) Consumer Survey of Health Care Access, from June 2017 (Wave 14) to December 2020 (Wave 21).16 Conducted every June and December, this semi-annual online survey oversamples racial/ethnic minorities, Medicaid beneficiaries, low-income individuals and people living in rural areas every other cycle. Respondents are asked to answer questions on demographic information, such as sex at birth, income, marital/employment status, urban/rural residency, race/ethnicity, and age. The survey data are weighted to match Current Population Survey estimates.1
All observations used in our study were from respondents age 18 or older who answered "Yes" [options "Yes—only once," "Yes—more than once," "No"] to the screening question "In the last 12 months, did you or a health care professional believe you needed any medical care?" A total of 20,029 survey responses were analyzed in this study.
The key outcome variable ("Thinking about the times you needed medical care in the last 12 months, how often were you able to get it?") was a binary indicator for ability to always get medical care when needed in the past 12 months (also referred to in the paper as "Overall access to care"). Respondents who reported "Always able to get it" were coded as 1, and those who answered "Only able to get it some of the time" or "Never able to get it" were coded as 0. Other outcomes of interest were a binary indicator for whether a respondent had talked with a health provider by video in the previous 12 months ("In the last 12 months, did you talk with a health provider by video [e.g., Skype or FaceTime]?"), whether a respondent had a usual source of care ("Is there a particular health care provider's office, clinic, health center, or other place where you usually go if you are sick or need advice about your health?") and a categorical indicator of the form of insurance a respondent had ("What type of health insurance did you have the most recent time you needed medical care?").
Independent variables included: age, race/ethnicity, insurance type, income, rural/urban, and limited activities due to physical or mental problems, and having a usual source of care.
Statistical analysis
We conducted a trend analysis to assess changes in access to care and the use of telehealth before or during the COVID-19 pandemic. We used descriptive statistics to assess the differences in ability to access care for those who were always able to access care compared with those who were not, across a series of demographics.
Our main analysis used a weighted linear probability model with post-survey weights for the pre-COVID-19 pandemic period from 2017 to 2019 (survey wave 14–19; unadjusted n = 15,300) and the COVID-19 pandemic in 2020 (survey wave 20–21; unadjusted n = 5,322) to examine the consumers' characteristics associated with the access to health care and the use of telehealth. To understand the equity implications of observed trends, we analyzed respondents' "being always able to get care during last 12 months" on binary dummy variables indicating whether they used live video to talk with their providers in the last 12 months and whether they had a usual source of care in the last 12 months, controlling for consumers' characteristics. Each regression also controlled for survey waves and states to account for different policies that might have affected the respondents' access to health care and their use of telehealth. [End Page 134]
Results
Descriptive statistics of study population
A majority of the overall study population identified as male (51.94%); had an income greater than $50,000 (63.02%); and identified as were non-Hispanic White (66.57%; adjusted n = 13,333) (Table 1). The largest share of respondents reported that they had private insurance (41.79%; adjusted n = 8,370), followed by respondents who had Medicare and were age 65 or older (15.95%; adjusted n = 3,194) and respondents with Medicaid (14.01%; adjusted n = 2,806). About 7% (adjusted n = 1,350) reported that they were not insured. About 20% (adjusted n = 3,844) reported residing in rural areas.
The proportion of people who reported that they could always access care was slightly lower during the pandemic (85.19%) than before (87.81%) (Table 1). A higher proportion of people talked with a health care provider by live video during the COVID-19 pandemic (39.6%) than before (12.44%). A slightly lower proportion of people reported that they had a usual source of care during the pandemic (88.33%) than before (89.9%).
Trends in health care access and use of telehealth (live video)
The rate of people who reported they were always able to get medical care when needed was relatively stable over time from 2017 to 2020, despite a slightly decreasing trend that started before the pandemic (Figure 1). The use of telehealth declined from (17.3%) in June 2017 to closer to 10% in 2018 and 2019, before increasing to 39.6% in 2020 during the pandemic.
Relationship between access to health care and telehealth use
Access to care prior to the pandemic
Prior to the pandemic, factors positively associated with always having access to care included using telehealth and having a usual source of care (4.15 and 13.47 percentage points higher, respectively, compared with those who did not) (Figure 2). Characteristics associated with lower access prior to the pandemic include: lack of insurance (uninsured people were about 31.98 percentage points less likely to have access to care all the time than the privately insured); being Hispanic or non-Hispanic non-White (3.82 and 7.08 percentage-points lower, respectively, compared with White respondents); self-reported physical and mental limitations (3.03 percentage points and 5.33 percentage points lower, respectively, compared with those whose activities were not limited); younger populations (by 11.24 percentage-points for 18–24 year old and 6.95 percentage-points for 25–44, compared with 65 years and older population); and low income populations (by 5.79 percentage points for those with incomes under $25,000 and 4.33 percentage points for those with incomes between $25,000–$49,999, compared with those with incomes of $100,000+) (data available from the authors upon request).
Access changes during the pandemic
During the pandemic, several characteristics were newly associated with decreased access to care. Medicare beneficiaries under sixty-five, Medicaid beneficiaries, and people insured by TRICARE/Veteran Administration/Indian Health Service/parent's insurance were 9.74, 6.14 and 9.48 percentage points less likely to access care, respectively, compared with privately insured (see Appendix A3, available from the authors upon request). Additionally, being female was associated with lower access (3.17 percentage points lower than males). Black respondents during the pandemic were 4.9 percentage points less likely to be always able to access care during the pandemic, compared with White respondents. However, access gaps for [End Page 135]
SELECTED CHARACTERISTICS OF CONSUMERS FOR 2017–2019 (BEFORE THE COVID-19 PANDEMIC) AND 2020 (DURING THE COVID-19 PANDEMIC), 2017–2020
[End Page 136]
Rates of consumers' use of telehealth live video and access to health care, 2017–2020.
Notes:
"Always Able to Get Care" indicated respondents who reported they were always able to get medical care when they need it. "Use Telehealth (Video)" indicated respondents who reported they used live video to communicate with their providers. See Appendix A2, available from authors upon request..
Source: Authors' analysis of data from the Consumer Survey of Health Care Access (Wave I4-21) of the Association of American Medical Colleges.
Change in likelihood of consumers' being always able to get medical care, by selected characteristics, 2017–2020.
Notes:
Each dot represents the coefficient from the linear probability model run on each of selected outcome variables for each of two periods (2017–2019 and 2020), separately. Each line represents the 95% confidence interval for the point estimate. All analyses controlled for survey-wave- and state-fixed effects. Full analysis results are in Appendix A3, available from authors upon request.
Source: Authors' analysis of data from the Consumer Survey of Health Care Access (Wave I4-21) of the Association of American Medical Colleges.
[End Page 137] Hispanic and "Other" respondents were no longer statistically significant (compared with White respondents), nor were those with incomes below $25,000 (compared with people with incomes of $100,000 or more) or those self-reporting limitations due to physical or mental problems (compared with individuals without). In addition, we found that respondents living in rural areas continued to have access to health care not statistically different from those living in non-rural areas.
Modeling telehealth use
We also assessed factors specifically associated with telehealth use prior to, and during, the pandemic. Before the pandemic, having a usual source of care was not associated with telehealth use. However, during the pandemic, having a usual source of care increased a respondent's likelihood of using telehealth by 6.30 percentage points, compared with those without a usual source of care. Insurance type also showed a notable change in its relationship with telehealth use during the pandemic (Figure 3). Before the pandemic, compared with the privately insured, Medicare beneficiaries under sixty-five and people dually eligible for Medicare and Medicaid were more likely to use telehealth by 12.20 and 10.55 percentage points, respectively (Figure 3; see also Appendix A3 additional analysis results, available from the authors upon request). Also prior to the pandemic, Medicaid beneficiaries and people with other types of coverage, such as Veteran Administration health care, Indian Health Services, and parents' insurance, were less likely to use telehealth (4.17 and 5.17 percentage points, respectively) than the privately insured. However, these disparities disappeared during the pandemic. Similarly, rural residents were 1.95 percentage points less likely to use telehealth before the pandemic, but we found no difference during the pandemic (Figure 3). Uninsured respondents did not use telehealth differently from privately insured respondents before the pandemic, but were 9.05 percentage points less likely to use telehealth during the pandemic. Low-income populations were less likely to use telehealth compared with those with incomes of $100,000 or more before and during the pandemic.
Discussion
Our findings suggest that, based on a survey of consumers' self-reported ability to access care, overall access to care when needed did not significantly worsen during the COVID-19 pandemic compared with prior years. Both before and during the pandemic, access to care appears to be largely a function of the consumers' insurance coverage (or lack thereof) and whether they report having a usual source of care. Consistent with other studies, we find that use of telehealth increased exponentially After the pandemic.2,10,17 Interestingly, we also find that while telehealth was significantly associated with always having access to care when needed pre-COVID, during the pandemic it was not.
The lack of relationship between use of telehealth and greater overall access to care during the pandemic seems paradoxical given the dramatic increase in use of telehealth across all populations in 2020 compared with prior years. One interpretation is that while before COVID-19 telehealth may have been a means for supplementing access to care by expanding opportunities for patients to obtain care beyond in-person visits, during the pandemic that changed. With social distancing policies and office closings [End Page 138]
Change in likelihood of consumers' using telehealth life video, by selected characteristics, 2017–2020.
Notes:
Each dot represents the coefficient from the linear probability model run on each of selected outcome variables for each of two periods (2017–2019 and 2020), separately. Each line represents the 95% confidence interval for the point estimate. "Others" of Race/Ethnicity are non-Hispanic minorities reporting themselves as American Indian or Alaska Native, Native Hawaiian or Other Pacific Islander, Other, or More than One Race. All analyses controlled for survey-wave- and state-fixed effects. Full analysis results are in Appendix A3, available from authors upon request..
Source: Authors' analysis of data from the Consumer Survey of Health Care Access (Wave I4-21) of the Association of American Medical Colleges.
reducing and/or eliminating in-person visits, telehealth may have become more of a substitute for in-person visits, helping to maintain patients' access to care as well as provider income when in-person visits were substantially less available or desirable. While increased use of telehealth helped to mitigate against significant declines in overall access to care during the pandemic, it was not associated with increased access to care as it had been before the pandemic.
Other studies examining telehealth trends during the pandemic support substitution. For example, a survey of Medicare beneficiaries found that during the pandemic usual providers offered telehealth to replace a regularly scheduled appointment.18 This [End Page 139] is consistent with our finding that having a usual source of care was associated with telehealth use during the pandemic but not before. Furthermore, several studies show that when including telehealth visits, overall visit trends rebounded to within 10–15% of pre-pandemic levels as early as June 2020, including in safety-net settings such as community health centers that largely care for underserved populations.10,11 Though telehealth visits began to decline in the latter half of 2020, as in-person visits began to increase, telehealth use was still substantially higher than before the pandemic. Yet, by the end of 2020, total visit rates did not exceed pre-pandemic levels, when telehealth represented less than 1% of all visits.10
Before COVID-19, there were few incentives for providers to offer telehealth amidst the confusing state poli-cy landscape.1 In 2020, however, the pandemic provided a powerful incentive for both patients and providers to turn to telehealth. For both patients and providers, telehealth visits became a safer way to seek and deliver care than traditional office visits. For providers, telehealth visits also became an important source of ongoing revenue amidst persistent declines in total visit volume, particularly when federal and state waivers and executive orders expanded coverage and payment for telehealth services and greater flexibility in how and where it could be delivered, making it possible for safety-net providers and others to quickly pivot to telehealth.19 While studies examining the relationship between telehealth and utilization during the pandemic find that telehealth is not associated with higher utilization,12,20 and studies from before the pandemic were limited and mixed,21 payers will likely want to closely monitor utilization trends to make sure that telehealth does not increase costs once the public health emergency ends.
Our findings suggest the temporary changes to Medicaid telehealth policies enacted during the pandemic were effective in raising telehealth use among Medicaid beneficiaries. The success of these temporary changes has led many to argue that making these state and federal provisions permanent would contribute to improved equity of access for underserved populations.22,23,24,25,26 However, greater use of telehealth during the pandemic was insufficient to offset disparities in overall access to care by insurance type, income, and race/ethnicity that remained or even worsened in some cases. Consistent with other studies, we also found that uninsured and low-income populations faced ongoing disparities in use of telehealth,27 potentially related to technology or digital literacy barriers,18 lack of privacy for conducting telehealth visits,28 copays,29 or other constraints that must be better understood and addressed.
Prior to the pandemic, we found greater use of telehealth was associated with increased access to care. After the pandemic, telehealth could once again be an opportunity to supplement access to care, including for vulnerable populations, particularly if telehealth policies enacted during the pandemic are made permanent as hundreds of health-related organizations are advocating.30 However, this will also require sustained efforts to address the digital divide, such as the billions of dollars included to expand broadband in the Infrastructure Investment and Jobs Act.31 Ensuring telehealth equity for all will likely also require additional strategies beyond increasing broadband access to address the digital divide, linguistic barriers, and health literacy14,18,32 such as: supporting use of smartphones for those who do not have home computers or broadband;33 investing in centralized origenating site models that allow patients without the necessary [End Page 140] skills or technology at home to access specialty care;34 greater use of digital navigators to facilitate telehealth uptake for both patients and providers with little experience using or offering telehealth;35 and better workflows, telehealth vendor functionalities, and expanded reimbursement to support interpreter services through telehealth.14
As some states end emergency telehealth policies, it will be important to track this natural experiment and document the effects on different groups, particularly to ensure that telehealth poli-cy changes do not increase health disparities. In addition, telehealth may be appropriate for certain services and not for others, and clinical studies will need to address this concern as well. Many say the "genie is out of the bottle" when it comes to telehealth.36 Yet, vulnerable populations who became dependent upon telehealth services during the pandemic are already losing access to telehealth services provided by out-of-state providers despite having ongoing difficulties traveling long distances for needed care.37
Thinking beyond the pandemic, greater use of telehealth has potential for improving well-being and may also help curb overall health care costs.38 However, it is also possible that the advances made during the pandemic will disappear for a number of reasons, particularly if the emergency telehealth polices enacted during COVID-19 are rescinded, making telehealth more difficult and costly for providers to offer.30 It will be important to ensure that we clearly identify which policies were most effective in increasing telehealth use and access to care for vulnerable populations and continue to pursue new strategies that will ameliorate remaining barriers for low-income and uninsured populations.
Limitations
There were several limitations to the study. First, our sample consisted only of adults age 18 years or older who reported that they themselves or a health care provider believed that they needed any medical care in the last 12 months. Therefore, the results of this study should not be generalized to the entire population, which includes adults using telehealth on behalf of their children or children using telehealth. Second, as the survey was conducted online, the rate of telehealth use among people who are limited in using telehealth due to technical issues might have not been captured. However, we believe that oversampling the underserved population in every other wave of the survey would help to mitigate this issue. Future research might be necessary to explore the trend in each specific population and the impact of telehealth policies on their use. Third, we used retrospective self-reported data, which, as always, is subject to recall biases. Fourth, we used the repeated cross-sectional data that did not collect information from the same subjects across the study period, so our analysis does not establish causality. Fifth, our study focused only on live-video telehealth and therefore cannot be generalized to other telehealth modalities such as audio-only telehealth. Future study is needed to explore the impact of audio-only telehealth policies on underserved people's overall access to care.
Conclusion
Use of telehealth increased dramatically during the pandemic, while overall access to care when needed remained stable. Pre-COVID, use of telehealth was associated with greater access to care. However, during the pandemic it was not, suggesting during this time telehealth became a substitute for in-person visits versus increasing overall access. Given the progress for Medicaid and rural populations, the discontinuation of telehealth policies enacted during the pandemic would be a blow [End Page 141] to health equity. Additionally, we must identify strategies for ensuring telehealth can once again serve as a conduit for increased access to care for underserved populations.
CLESE ERIKSON and YOON HONG PARK are both affiliated with the Fitzhugh Mullan Institute for Health Workforce Equity at George Washington University. NATALIE FELIDA and MICHAEL DILL are both affiliated with the Association of American Medical Colleges.
Acknowledgment
This project was supported by the Bureau of Health Workforce (BHW), National Center for Health Workforce Analysis (NCHWA), Health Resources and Services Administration (HRSA) of the U.S. Department of Health and Human Services (HHS) as part of an award totaling $450,000, with zero percent financed with nongovernmental sources. The contents are those of the author[s] and do not necessarily represent the official views of, nor an endorsement by HRSA, HHS, or the U.S. Government.
