Since 2014, the Institute of Data and Analytics (IDA) at the University of Alabama has had a consulting contract with Alabama Medicaid, in which the IDA supports Alabama Medicaid’s analytics group. To fulfill this contract, the IDA has maintained a secure copy of all Alabama Medicaid enrollment and claims data since 2010. These data include specific billing/claim information for all enrolled in Alabama’s Medicaid program. Annually, approximately 1 (25%) of every 4 Alabamians has some form of Medicaid coverage during the year. Medicaid insurance services are made available to low-income populations, as well as those who are disabled, at no cost to participants.

Data from the procedure and provider tables within the Medicaid database were extracted that contained corresponding billing codes indicating actions had taken place over telemedicine. Claims information within the Medicaid database can lag behind for up to 6 months since facilities and practitioners may not always bill in a timely manner. However, over 80% of claims are received within a month from the time of service. The variables extracted included the year and month of the service, the county where the service took place (provider location), the gender of the patient, the age range of the patient corresponding to US Census age groups, provider/procedure information, modality of service (in person vs telehealth), and total billing costs paid. Procedures were further grouped by type of procedure and group of procedure, and providers were further grouped by type of provider as well as the specialty of the provider. These variables were aggregated, and calculations of counts of unique persons, counts of claims, and total cost of claims for each level of aggregation were calculated. The data represent a complete sampling of the Alabama Medicaid records from March 1, 2019, through June 30, 2021. The data were extracted from the August 2021 data backup of the Alabama Medicaid data, which included transactions through July 2021 and represented approximately 80% of all claims through June of that year. In total, 94,254,599 patient claims covered by Alabama Medicaid at some point during the analysis period are represented in the telehealth utilization data, along with a total of about US $5 billion across 738 procedure types from 27 provider categories across 97 specialties.

The COVID-19 tracking reports were downloaded from the COVID Tracking Project [11], which collects and publishes the most comprehensive data about COVID-19 in the United States. We also collected the COVID-19 vaccination progress report for Alabama from the USAFacts website [12]. Lastly, a classification of areas as being rural or urban at the county level was obtained from the Alabama Rural Health Association [13].

The cumulative time frame of analysis was from March 1, 2019, through June 30, 2021, as shown in Figure 1. Alabama announced the first known cases of coronavirus in March 2020; therefore, the time frame being considered as pre-COVID-19 was the period from March 1, 2019, to February 29, 2020. Broad COVID-19 vaccine availability in Alabama occurred during March 2021, so the time frame from March 1, 2021, to June 30, 2021 was analyzed separately as a mass vaccination condition. The core pandemic time frame was characterized by March 1, 2020, through February 28, 2021. This allowed for 12 months of data prior to COVID-19, 12 months of data during the pandemic, and 4 months of analysis of the time frame when vaccines were widely available.

We compared the characteristics of patients who utilized and did not utilize the telemedicine service in each period. For each telemedicine usage modality, chi-square tests were conducted to assess the distribution of services and patient characteristics (eg, age group, gender, race, level of rurality, and service provider type) across different study periods. We then calculated the rate of telemedicine usage for the various characteristics for each month in the study periods.

Institutional board approval was sought and obtained for this project from the University of Alabama's ethics board (UA IRB protocol #21-05-4661). The full study protocol is available through the institution or by contacting the authors. Additionally, the research was vetted and approved by Alabama Medicaid’s internal research review process.

The data set contained 94,254,599 Medicaid claims from March 1, 2019, through June 30, 2021, in the state of Alabama. For each period, we reported the rate of telemedicine usage, which increased from 0.12% (n=55,613) in the prepandemic period to 3.43% (n=1,141,282) and 1.85% (n=254,807) in the subsequent 2 periods, respectively, as seen in Table 1. In addition, the second period (ie, the first year after the onset of the pandemic) saw the lowest average monthly Medicaid claims (2,774,171 claims per month) but the highest average monthly telemedicine Medicaid claims (95,107 claims per month) among the 3 study periods. Such an opposite trend was partly due to the statewide stay-at-home orders in that period.

Figure 2 provides an overview of monthly changes in new COVID-19 confirmed cases, vaccination progress, and telemedicine usage in Alabama. As can be seen in the figure, the nontelemedicine claims and total claims dramatically dropped since March 2020. Both lines reached the valley when the state issued a stay-at-home order in April 2020, while at the same time, telemedicine usage reached its peak point. The monthly claims of telemedicine usage gradually decreased during the months after April 2020, but the number of monthly claims was still at a relatively high level at the end of our study period (ie, June 2021). This is an indication that telemedicine usage is still popular, even after the rollout of mass vaccination and the relaxation of the COVID-19–related executive orders. Interestingly, the monthly new confirmed cases peaked in December 2020, but neither telemedicine nor nontelemedicine claims exhibited any obvious changes during that period. This is possibly due to people having become accustomed to the new norm brought by the pandemic.

In Table 2 (Multimedia Appendix 1) and Table 3 (Multimedia Appendix 2), we summarize the characteristics of Medicaid claims in the 3 periods (ie, before the pandemic, after the onset of the pandemic, and after the rollout of mass vaccination). Monthly Medicaid claims are reported along with the proportions for the groups within each classification. The sparklines, with the lowest and highest values marked in blue dots, were used to visualize the trend in claim counts over the 3 periods. Chi-square tests were performed to assess the distribution of various characteristics of patients who used telemedicine (age group, gender, race, level of rurality, and service provider type) across the 3 periods. These chi-square tests all resulted in a P value <.001. Applying the correction for multiple comparisons, we calculated the Bonferroni correction α values for each test, and the smallest adjusted α value was 0.05/42=0.0012, suggesting that there were significant shifts in the constitution of telemedicine and nontelemedicine usage in terms of the aforementioned patient/service characteristics classifications. The trend in the total number of Medicaid claims was similar to nontelemedicine Medicaid claims, as the latter made up a large proportion of the former (see Table 1).

For each month within the study time frame, we reported the rate of telemedicine visits, which is defined as the count of telemedicine claims divided by the count of all claims. We then compared the rates across groups for the various characteristics.

The COVID-19 pandemic introduced never-before-seen challenges to modern society. Although not directly planning for a pandemic, telehealth availability had been steadily increasing prepandemic but utilization was low. Barriers to implementing telemedicine include poor access to technical equipment, management, and reimbursement [15]. The top barriers are technology specific and could be overcome through training, change management techniques, and alternating delivery by telemedicine and personal patient-to-provider interaction. However, technology, as the major barrier, can be overcome through training, changing management techniques, and alternating delivery of services through telemedicine and personal patient-physician interactions [16]. The pandemic presented an opportunity to stress-test the underutilized telehealth modality, and the results showed that not only did the technology meet a large proportion of the needs but also a shift in society and providers toward a willingness to continue using the medium was observed. Beyond the baseline period, the COVID-19 pandemic has prompted a significant rise in the use of telemedicine for both urgent and nonurgent medical visits [17].

Although Wood et al [18] reported no significant differences in telehealth completion rates by age, sex, gender, or insurance, we found some differences in different dimensions. A telemedicine usage percentage of 57% in period 2 for 0-17-year-olds indicates a strong willingness of younger individuals to use telehealth services. The use of telemedicine by children varied by age, race, ethnicity, and recent preventive treatment, building upon previous concerns about disparities in telemedicine availability [19]. The extent to which this population directly uses telehealth versus having a parent or guardian facilitate the communications is unknown, and an area where expanded research could provide additional insights.

As a population, females experienced a higher increase in telehealth utilization compared to males. One possible reason for this could be related to the 0-17-year-old population. If mothers use telehealth for their children, then it stands to reason that they might likely do the same for themselves. Alabama Medicaid covers more than half of all births annually in Alabama, so the overall population contains a large number of females and younger persons. According to Patel et al [20], rapid telemedicine expansion can be particularly complex for pediatric patients, but approaches that satisfy privacy, security, and convenience will effectively increase pediatric enrollment capacity for telemedicine. Although telemedicine appears to be feasible and acceptable for clinical patients, questions about confidentiality, quality of care, and health disparities remain unanswered, so clinical guidelines are needed to guide best practices [21].

All races experienced a uniform increase in telehealth utilization, but the Black/African American community experienced the least in terms of percentage of increase. Research into racial disparities in telehealth by Rivera et al [22] in 2021 echoed these results by showing that African Americans are less likely to use telehealth and online services compared to Caucasians. In addition, according to Wegermann et al [23], there is still a gap in the growth of teletherapy relative to others for vulnerable populations, including those who are older, are non-Hispanic Black individuals, or have Medicare/Medicaid health insurance [23]. However, the results of this study are at odds with what Campos-Castillo and Anthony [24] found in that their data showed that African American respondents are more likely than Caucasians to report using telehealth because of the pandemic, particularly when perceiving the pandemic as a minor threat to their own health. Geographic differences could also have significant influences on racial disparities in telehealth utilization.

Although the Hispanic community experienced a large 42-fold increase, this was in part due to the fact that there was such low utilization (49 average monthly) during the prepandemic period. As for the percentage of the modality of visit type, telehealth was the lowest for Hispanics in all months examined except for April 2020, when they briefly passed the Black community by a small margin. As noted in multiple studies, telehealth access and utilization among Hispanics are inhibited by trust barriers, awareness, and eHealth literacy [25].

The willingness of younger patients to use telehealth, coupled with higher utilization rates among women, provides a positive glimpse into the future of telehealth. Racial and ethnic disparities in use due to many factors continue to exist, but the systemwide increase across all races/ethnicities also shows significant potential for sustained use. The types of services that the current set of available technologies can facilitate, however, provide insights into their usage rates.

When examining provider type, behavioral health and psychologists’ services via telemedicine experienced the highest levels during the second phase, with over 60% of both services moving to the modality. Telehealth is well equipped to provide voice and video communications, which are some of the key aspects needed for these types of services. The next highest 2 were mental health services and therapists, a continued indication that the overall set of mental health services is most easily transferred to telemedicine. Dubin et al [26] found an almost 2-fold increase in the use of telemedicine by urologists, indicating that they have the ability to adopt and adapt telemedicine into their practice, but the barriers involved in the telemedicine technology itself still prevent many from taking advantage of it [26].

Although all services have diminished since their peak in the shutdown period of the second phase of the pandemic, behavioral health, psychologists’ services, and mental health providers showed continued high utilization post-large-scale vaccine availability ranging from 10%-25%. All other services, including therapists, dropped back to below 5% telemedicine usage after vaccine availability. This is a possible indication that although telemedicine works for many services, in-person service is preferable to telemedicine. The study conducted by Smrke et al [27] reported the benefits of telemedicine for patients and clinicians in the long run since it can change cancer treatment delivery, particularly for patients with rare cancers who reside far away from expert centers.

Alhajri et al [28] found that video consultation should be frequently used in remote clinical consultation for acute conditions but that audio consultation is comparable in providing remote follow-up care for patients with chronic conditions and that audio consultation may greatly increase geographically accessible telemedicine services. Targeted efforts may be required during video visits to address patient populations that are older or have lower levels of knowledge [29]. At the same time, smartphone technology can serve as an extension of telemedicine, enabling the future of telehealth practices [30]. According to Orrange et al [31], patients’ satisfaction using telemedicine is affected by their level of confidence in physicians and visit-related factors. The aims of improving access to treatment while avoiding overuse and fraud should be balanced in telemedicine policy, both in terms of regulation and in terms of payment [32].

In terms of rurality, a uniform increase in telehealth utilization was observed across highly rural, moderately rural, and urban counties, with moderately rural counties reaching the highest overall utilization rate. Research conducted by Chu et al [10] documented an increase in the adoption of telemedicine in rural and remote areas, but the use of telemedicine increased in urban and less rural populations during the COVID-19 pandemic. The observed tight correlation of telehealth trends across all types of counties indicates a significantly uniform willingness to use telehealth and overall consistency in the availability of telehealth. Studies such as Breton et al [33] indicate that mobility issues and patients living in remote areas could negatively impact telehealth utilization, but results from this analysis did not find significant access issues, even in the most rural counties. This could be attributed to the more widespread availability of high-speed mobile data services. To preserve the long-term viability of telemedicine programs in the aftermath of the COVID-19 pandemic, persuading third parties to continue to fund these services is necessary [34]. The future of telemedicine will also require addressing access barriers for vulnerable populations, such as people with disabilities, by making significant, long-term changes in technology, regulatory and legislative infrastructure, and customized solutions that meet the unique needs of patients and health systems [35].

There are multiple factors that influence patient and practitioner willingness to use telemedicine, including connectivity, familiarity, and reimbursement policies, among others. The results of this study are likely applicable to states with similar demographics and rurality as Alabama, but broader applicability across the United States is questionable. However, Alabama has a high degree of rurality, poverty, and generally lower metrics in many socioeconomic factors, so if a technology such as telehealth can work in rural Alabama amongst its Medicaid population, then as long as connectivity exists elsewhere, we see no reasons, if policies allow, that similar results cannot be observed in other regions. In addition, it was reported that there were substantial differences in telemedicine completion rates among commercial insurance, Medicare, and Medicaid [36]. We only examined patients with a particular insurance type (ie, Medicaid) in this study, and researchers can also conduct similar analyses on telemedicine adoption of patients with other insurance types. Lastly, the study was descriptive in nature, which did not allow for the controlling of confounding factors. As a result, it is difficult to justify if any of the changes seen in a particular variable (ie, rurality) are not a product of other demographics (ie, if a particular demographic group of people may tend to live in a more rural area). Future studies may build multiple regression models to systematically investigate the factors that could influence telemedicine adoption.

The pandemic has had a jarring and brutal impact in terms of loss of life, economic stress and despair, and many other negative aspects. If a silver lining were to exist, though, from health care access Wand availability perspectives, it would be in the form of telemedicine adoption and utilization. Within the Alabama Medicaid population, telehealth services have proven to be extremely viable and have withstood the large-scale availability of vaccines to continue to be a modality of choice for health care in both rural and urban areas. Telemedicine is likely to continue to play an integral role in health care, and as a first step toward increasing the use of telemedicine, health care systems should focus on improving patient portal usage for better access to telemedicine services [37].

The potential benefits of sustained large-scale utilization of telehealth, especially in rural areas, are quite significant, yet concerns exist. As noted by Shachar et al [38], although telehealth may increase access, safety and privacy concerns are still common among users. In a similar vein, Bokolo [39] called for stakeholders and policymakers to confront the social, organizational, and technological determinants that are barriers to the increased adoption of telehealth. Although all of these concerns are valid and still exist, the pandemic triggered an interesting test of the telehealth systems nationwide, and the response was shown to be promising. Even in a state such as Alabama, with high rurality and high poverty, telehealth has not only shown to be an effective stop-gap measure but also has continued to show increased utilization postvaccine availability. The noted disparities among races with lower utilization rates among Black and Hispanic communities, coupled with the difference in usage amongst urban versus highly rural areas, stand as opportunities for increased focus in the future.