Abstract
Purpose
The recent increase in mental health care providers offering telehealth may improve access by reducing travel costs, particularly for those in rural areas. This paper seeks to understand how mental health care utilization changed from 2019 to 2023 for rural and nonrural areas.
Methods
This study uses data from the National Health Interview Survey for 2019 and 2021‐2023 (n = 118,652). To adjust the utilization rates for sociodemographic factors, a probit model with survey weights is used. For each year, the adjusted percentage of rural and nonrural populations receiving any mental health care in the past year is reported. Changes in demographic characteristics of those receiving care are also discussed.
Findings
This study finds that both rural and nonrural populations experienced large increases in adjusted mental health care utilization rates from 2019 to 2023. Utilization rates in rural areas grew 3.72 percentage points, from 9.35% (95% CI: 8.85‐9.84) in 2019 to 13.07% (95% CI: 12.86‐13.28) in 2023. For nonrural areas, utilization rates increased 4.40 percentage points, from 12.06% (95% CI: 11.79‐12.74) in 2019 to 16.46% (95% CI: 15.82‐17.11) in 2023.
Conclusion
Rural populations utilize less mental health care than nonrural populations, but both saw a substantial increase from 2019 to 2023. While telehealth may particularly benefit rural areas, these data suggest a larger increase in utilization for nonrural respondents. Future work is needed to better understand remote mental health care and rural populations.
Keywords: COVID‐19, mental health, rural health, telehealth
INTRODUCTION
Rural populations in the United States have historically had lower mental health care utilization compared to their nonrural counterparts. 1 , 2 , 3 This disparity in utilization is multifactorial. Limited availability of providers and high travel costs associated with in‐person appointments for rural patients are cited as major bariers. 4 , 5 An estimated 65% of rural counties do not have a practicing psychiatrist. 6 This often makes far travel to appointments in neighboring counties a requirement for those seeking care. Furthermore, stigma toward mental health care is considered stronger in rural areas, 7 , 8 which may impact care‐seeking behavior. Uninsurance rates are also higher in rural areas than in nonrural areas in the United States. 9
The COVID‐19 pandemic transformed many aspects of mental health care, particularly with the expansion of remote appointments. The amount of mental health care providers with telehealth offerings, which includes both audio and video appointments, has expanded rapidly in the wake of the COVID‐19 pandemic. 10 , 11 In September 2022, 88.1% of mental health care facilities offered telehealth appointments, compared to just 39.4% in April 2019. 10 While all populations may benefit from telehealth access, the reductions in travel costs are likely more significant for people in rural areas. As a result, the expansion of telehealth could lead to a larger increase in access, and thus mental health care utilization, for those in rural parts of the United States. In fact, before the quick expansion of telehealth in 2020, the rate of telehealth utilization in rural areas was higher than in nonrural areas and was increasing. 12 , 13 The presence of telehealth as an option for mental health care may also change the composition of those receiving care. For example, younger populations report a higher willingness to use telehealth pre‐pandemic. 14
This study does not measure the use of telehealth directly. Rather, it documents changes in overall self‐reported mental health care utilization rates and composition in rural and nonrural areas from 2019 to 2023, a period that coincided with expanded access to telehealth. Causal channels are not explicitly explored in this study, though discussions focus on the possible impact of telehealth expansion in 2020.
METHODS
Data
This study uses data from the National Health Interview Survey 15 (NHIS) to assess trends in mental health care utilization between 2019 and 2023. To avoid any complications presented by lockdowns, problems with survey data, and other COVID‐19 pandemic‐related issues, data from 2020 are excluded. The NHIS is a large, annual, nationally representative survey that collects information on a variety of health‐related topics. It is, to my knowledge, the largest publicly available data source that includes measures of mental health care utilization, rural status, and demographic characteristics. The NHIS data uses the 2013 National Center for Health Statistics Urban‐Rural Classification Scheme. For this study, nonmetropolitan areas are considered rural, while all others (Large Metropolitan, Fringe Large Metropolitan, and Medium and Small Metropolitan) are considered nonrural. The NHIS did not report rural status before 2019, making the longest sample period possible with this data source 2019‐2023. After restricting the data to 2019, 2021‐2023, and adults 18 or older, the data contains 118,652 observations, 18,187 in rural areas and 100,465 in nonrural areas. The NHIS has questions about mental health counseling/therapy use as well as psychotropic medications, resulting in 2 binary measures. To examine changes in utilization rates, I focus on the percentage of respondents who report receiving any mental health care (defined by receiving either mental health counseling/therapy or the use of medication for mental health issues), followed by the 2 separate types.
The NHIS is administered throughout the full year 16 and all measures are recorded regarding the past 12 months. Thus, the data reported in this study pertains to a wider range of possible dates than other data sources, such as claims data.
Analysis
To document utilization rates by rural status and year, I used a probit model to adjust the utilization rates for sociodemographic factors. Probit models use 4 years of NHIS data (2019, 2021, 2022, and 2023) as 2020 data is excluded as discussed previously. The dependent variable in the primary model is a binary indicator equal to 1 if the respondent reported receiving any mental health care (either counseling/therapy or medication) in the past 12 months, and 0 otherwise. In addition to the primary model estimating any mental health care use, I estimate separate probit models for 2 subcomponents: mental health counseling/therapy and use of medication for mental health issues. Each model uses a binary dependent variable coded as 1 if the respondent reported the relevant service in the past 12 months, and 0 otherwise, and the same specification as the main model. I further augment the probit model to include an interaction term between rural status and each year in addition to the main effects. These models are estimated to confirm results from the margins command in previous probit models and gauge the statistical significance.
The controls used in all models include a continuous measure of age as well as binary measures of sex, race/ethnicity, region, year of survey, and insurance coverage. The year of survey is included as a set of binary dummy variables, with 2019 as the reference group, and no observations for 2020. Survey weights provided by the NHIS are used in all models. A probit model was selected to estimate the binary outcomes of utilization due to its flexibility and widely used approach while accounting for nonlinear relationships. 17 This approach is commonly used for binary dependent variables in health services research, especially when estimating adjusted probabilities. Adjusted utilization rates were estimated by the margins command following the probit models. This approach provides predicted probabilities for each year and rural status combination while adjusting for all covariates.
The composition of those receiving mental health care is next analyzed. Specifically, I report the demographic makeup of those receiving any mental health care for the first year of data, 2019, and the most recent, 2023, for both rural and nonrural populations. Here, I report 2019 and 2023 data for a cleaner and easier comparison. Demographic variables considered include sex, age ranges (18‐29, 30‐44, 45+), and race/ethnicity. Observations weighted with survey weights provided by the NHIS. All statistical analyses were done using Stata‐18.
RESULTS
Table 1 reports results from probit models, followed by Figure 1, which shows adjusted utilization rates by year and rural status. In rural areas, adjusted utilization rates for any mental health care increased from 9.35% (95% CI: 8.85‐9.84) in 2019 to 13.07% (95% CI: 12.86‐13.28) in 2023, representing a 3.72 percentage point (P<.001). In nonrural areas, utilization rates increased from 12.06% (95% CI: 11.79‐12.74) in 2019 to 16.46% (95% CI: 15.82‐17.11) in 2023, a 4.40 percentage point increase (P<.001). The first graph in Figure 1 shows these changes.
TABLE 1.
Probit results.
| Panel 1. Outcome: any mental health care | ||||
|---|---|---|---|---|
| Coefficient | Standard error | P‐value | Marginal effect | |
| Rural | −0.157 | 0.009 | .000 | −0.033 |
| 2021 | 0.103 | 0.001 | .000 | 0.020 |
| 2022 | 0.182 | 0.012 | .000 | 0.037 |
| 2023 | 0.209 | 0.021 | .000 | 0.043 |
| Male | −0.331 | 0.006 | .000 | −0.070 |
| Age | −0.001 | 0.002 | .714 | −0.003 |
| Insured | 0.410 | 0.009 | .000 | 0.087 |
| Midwest | −0.023 | 0.027 | .407 | −0.005 |
| South | −0.090 | 0.023 | .000 | −0.019 |
| West | 0.012 | 0.005 | .008 | 0.003 |
| White non‐Hispanic | 0.405 | 0.016 | .000 | 0.081 |
| Black non‐Hispanic | 0.127 | 0.001 | .000 | 0.022 |
| Asian non‐Hispanic | −0.261 | 0.014 | .000 | −0.035 |
| AIAN non‐Hispanic | 0.213 | 0.099 | .031 | 0.038 |
| Multiple races | 0.347 | 0.033 | .000 | 0.067 |
| Constant | −1.290 | 0.007 | .000 | |
| Pseudo R 2 | 0.064 | |||
| Panel 2. Outcome: mental health counseling/therapy | ||||
|---|---|---|---|---|
| Coefficient | Standard error | P‐value | Marginal effect | |
| Rural | −0.185 | 0.008 | .000 | −0.034 |
| 2021 | 0.091 | 0.002 | .000 | 0.015 |
| 2022 | 0.175 | 0.012 | .000 | 0.031 |
| 2023 | 0.203 | 0.025 | .000 | 0.036 |
| Male | −0.291 | 0.000 | .000 | −0.053 |
| Age | 0.002 | 0.002 | .225 | −0.002 |
| Insured | 0.389 | 0.009 | .000 | 0.071 |
| Midwest | −0.053 | 0.031 | .085 | −0.010 |
| South | −0.127 | 0.027 | .000 | −0.023 |
| West | 0.007 | 0.008 | .364 | 0.001 |
| White non‐Hispanic | 0.348 | 0.018 | .000 | 0.059 |
| Black non‐Hispanic | 0.142 | 0.005 | .000 | 0.021 |
| Asian non‐Hispanic | −0.236 | 0.015 | .000 | −0.027 |
| AIAN non‐Hispanic | 0.253 | 0.058 | .000 | 0.041 |
| Multiple races | 0.335 | 0.032 | .000 | 0.057 |
| Constant | −1.399 | 0.012 | .000 | |
| Pseudo R 2 | 0.063 | |||
| Panel 3. Outcome: mental health medication | ||||
|---|---|---|---|---|
| Coefficient | Standard error | P‐value | Marginal effect | |
| Rural | −0.052 | 0.003 | .000 | −0.003 |
| 2021 | 0.042 | 0.001 | .000 | 0.002 |
| 2022 | 0.116 | 0.019 | .000 | 0.007 |
| 2023 | 0.095 | 0.004 | .000 | 0.005 |
| Male | −0.107 | 0.004 | .000 | −0.006 |
| Age | −0.021 | 0.001 | .000 | −0.001 |
| Insured | 0.249 | 0.001 | .000 | 0.014 |
| Midwest | 0.103 | 0.009 | .000 | 0.006 |
| South | 0.099 | 0.004 | .000 | 0.005 |
| West | 0.076 | 0.012 | .000 | 0.004 |
| White non‐Hispanic | 0.314 | 0.013 | .000 | 0.017 |
| Black non‐Hispanic | 0.039 | 0.019 | .042 | 0.002 |
| Asian non‐Hispanic | −0.244 | 0.014 | .000 | −0.007 |
| AIAN non‐Hispanic | 0.041 | 0.004 | .000 | 0.002 |
| Multiple races | 0.163 | 0.033 | .000 | 0.008 |
| Constant | −1.813 | 0.032 | .000 | |
| Pseudo R 2 | 0.033 | |||
FIGURE 1.
Adjusted mental health care utilization by rural status.
Note: Utilization rates are adjusted using a probit model that controls for socioeconomic variables. 95% confidence intervals are shown with brackets.
The interacted model findings suggest similar results. This model indicates that overall utilization increased between 2019 and 2023, with the probability of receiving any mental health care rising by approximately 4.86 percentage points over the period (P<.001). However, the interaction terms reveal that this increase was significantly smaller among rural respondents. Specifically, the 2023 interaction term, which has nonrural respondents as the comparison group, is negative and statistically significant (marginal effect = –0.0258, P = .003), suggesting that rural residents experienced a more modest gain in utilization relative to their nonrural counterparts. These results are roughly aligned with results from the margins command shown above.
Next, mental health counseling and medications are examined separately. As shown in the second and third graphs of Figure 1, rural areas had an adjusted utilization rate of mental health counseling of 7.07% (95% CI: 6.65‐7.49) in 2019, which increased 2.98 percentage points to 10.05% (95% CI: 9.78‐10.31) in 2023 (P<.001). Nonrural areas reported a 9.76% (95 CI: 9.74‐10.05) utilization rate of mental health counseling in 2019 and 13.49% (95% CI: 12.85‐14.13) in 2023, a 3.73 percentage point increase (P<.001). For mental health medications, rural areas grew from an adjusted utilization rate of 1.98% (95 CI: 1.89‐2.06) in 2019 to 2.46% (95% CI: 2.41‐2.52) in 2023. This represents an increase of 0.48 percentage points (P<.001). In nonrural areas, the adjusted utilization rate for mental health medication grew from 2.23% (95% CI: 2.17‐2.29) in 2019 to 2.77% (95% CI: 2.74‐2.80) in 2023, an increase of 0.49 percentage points (P<.001).
Table 2 presents the demographic composition of those receiving any mental health care in 2019 and 2023, stratified by rural status. In rural areas, no statistically significant changes were observed in the share of users by sex, age groups, or race/ethnicity. In contrast, nonrural areas saw significant changes. In 2019, 29.2% of those utilizing care were aged 30–44, whereas in 2023, this age group made up 33.8% (P <.001). In 2019, 41.2% of those receiving care were aged 45+, which declined to 34.5% in 2023 (P <.001). In 2019, 28.9% of those receiving care were non‐White, who saw an increase to 31.5% in 2023 (P = .031).
TABLE 2.
Composition of those utilizing mental health care.
| Rural | Chi‐squared test | Nonrural | Chi‐squared test | |||
|---|---|---|---|---|---|---|
| 2019 | 2023 | P‐value | 2019 | 2023 | P‐value | |
| Male | 33.67% | 40.55% | .326 | 37.71% | 37.85% | .429 |
| Female | 66.33% | 59.45% | .326 | 62.29% | 62.15% | .429 |
| Age 18‐29 | 27.81% | 33.00% | .216 | 29.64% | 31.76% | .231 |
| Age 30‐44 | 26.15% | 28.70% | .327 | 29.17% | 33.76% | .000 |
| Age 45+ | 46.03% | 38.30% | .066 | 41.19% | 34.48% | .000 |
| White | 84.61% | 83.43% | .273 | 71.13% | 68.51% | .031 |
| Non‐White | 15.39% | 16.57% | .273 | 28.87% | 31.49% | .031 |
| N | 463 | 467 | 3,045 | 3,672 | ||
Note: Weighted percentages reported are out of those reporting any utilization of mental health care in the past 12 months. Chi‐squared test results are conducted between 2019 and 2023 values for each variable within rural status.
DISCUSSION
The analysis reveals a notable increase in mental health care utilization between 2019 and 2023, as well as a larger share of younger adults receiving care in nonrural areas. Both rural and nonrural areas show growth in terms of higher utilization, as rural areas increased utilization rates by 3.72 percentage points and nonrural areas by 4.40 percentage points. It is worth noting that compared to 2019 rates, these 2 increases are similar in terms of percentage increases. 2023 utilization rates in rural areas were 40% higher compared to 2019 levels, and nonrural areas were 36% higher than their 2019 levels.
One possible explanation for these differences is the gap in internet access. Rural populations report slightly lower rates of broadband access compared to nonrural populations. 18 , 19 While the gap is shrinking, 72% of rural respondents of a 2021 Pew Research Survey reported having access, compared to 77% and 79% of urban and suburban respondents. 18 Despite the expansion of telehealth, rural individuals without internet access may face challenges in utilizing these services.
Several other factors may limit the ability of people in rural areas to utilize telehealth. First, people in nonrural areas are more likely to work from home, 20 particularly as remote work became more common following the COVID‐19 pandemic. 21 This shift could make telehealth a more viable option for individuals in nonrural areas, as they have greater flexibility in their work schedules to attend virtual appointments. As previously mentioned, stigma toward mental health care in rural areas is estimated to be larger than in urban areas, 7 which may limit increases in utilization in rural areas even with telehealth access. Similarly, higher uninsurance rates in rural areas may prevent some from receiving care even when offered telehealth. 9
While the growth from 2019 to 2023 was larger in nonrural areas, it is worth noting that the increase in rural populations was still of a high magnitude. In fact, the 2023 utilization rate in rural areas is higher than the 2019 rate in nonrural areas, indicating a substantial increase in care being delivered.
Limitations
This analysis has several limitations. First, there may have been changes in 2020 outside of access to telehealth that could have influenced mental health care utilization, such as the differential impact from the COVID‐19 pandemic. 22 Additionally, this study uses self‐reported data, which may lead to underreporting for mental health care, given its documented stigma. 7 Next, the NHIS did not report data on rural status before 2019, preventing a longer period from being explored. Accessing multiple years of pre‐2020 data would allow an understanding of trends before the quick expansion of telehealth. The NHIS only offers data on rural status via the 2013 National Center for Health Statistics Urban‐Rural Classification Scheme, which prohibits comparison with other, more nuanced measures. Further, this study does not include data from 2020, restricting any examination of the immediate impact of telehealth expansion. The results reported in this study are not intended to be considered causal with respect to telehealth access.
CONCLUSION
While more research is needed, the data in this simplified setting suggest that telehealth may have contributed to a broad‐based rise in utilization, but do not indicate a larger increase in rural areas. In fact, this work finds evidence of a larger increase in nonrural areas, perhaps indicating important differences between the 2 populations with respect to mental health care. Nevertheless, rural populations did see a stark increase in care, with 2023 utilization rates in rural areas being higher than 2019 rates in nonrural areas. More research is needed, however, to draw any causal conclusions, as the results presented here only document trends. Policy makers interested in improving access to mental health care may look to investments in rural internet connection and mental health education.
CONFLICT OF INTEREST STATEMENT
The author declares no conflicts of interest.
ACKNOWLEDGMENTS
I thank Austin Kines and Shyam Raman for their helpful feedback.
Stanley BJ. Trends in mental health care utilization in rural and nonrural areas, 2019‐2023. J Rural Health. 2025;41:e70074. 10.1111/jrh.70074
REFERENCES
- 1. Kirby JB, Zuvekas SH, Borsky AE, Ngo‐Metzger Q. Rural residents with mental health needs have fewer care visits than urban counterparts. Health Aff. 2019;38(12):2057‐2060. doi: 10.1377/hlthaff.2019.00369 [DOI] [PubMed] [Google Scholar]
- 2. Chen Z, Roy K, Khushalani JS, Puddy RW. Trend in rural‐urban disparities in access to outpatient mental health services among US adults aged 18‐64 with employer‐sponsored insurance: 2005‐2018. J Rural Health. 2022;38(4):788‐794. doi: 10.1111/jrh.12644 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Lambert D, Agger M, Hartley D. Service use of rural and urban Medicaid beneficiaries suffering from depression: the role of supply. J Rural Health. 2008;15(3):344‐355. doi: 10.1111/j.1748-0361.1999.tb00756.x [DOI] [PubMed] [Google Scholar]
- 4. Mseke EP, Jessup B, Barnett T. Impact of distance and/or travel time on healthcare service access in rural and remote areas: a scoping review. J Transp Health. 2024;37(37):101819. doi: 10.1016/j.jth.2024.101819 [DOI] [Google Scholar]
- 5. Mack B, Graves J. Mental Health in Rural Areas. 2022. Accessed August 4, 2025. https://www.ruralhealth.us/getmedia/cf3c3922‐25cb‐49a0‐bb04‐0bad81d634f9/NRHA‐Mental‐health‐in‐rural‐areas‐policy‐brief‐2022.pdf
- 6. Andrilla CHA, Patterson DG, Garberson LA, Coulthard C, Larson EH. Geographic variation in the supply of selected behavioral health providers. Am J Prev Med. 2018;54(6):S199‐S207. doi: 10.1016/j.amepre.2018.01.004 [DOI] [PubMed] [Google Scholar]
- 7. Schroeder S, Tan CM, Urlacher B, Heitkamp T. The role of rural and urban geography and gender in community stigma around mental illness. Health Educ Behav. 2020;48(1). doi: 10.1177/1090198120974963 [DOI] [PubMed] [Google Scholar]
- 8. Jones AR, Cook TM, Wang J. Rural–urban differences in stigma against depression and agreement with health professionals about treatment. J Affect Disord. 2011;134(1‐3):145‐150. doi: 10.1016/j.jad.2011.05.013 [DOI] [PubMed] [Google Scholar]
- 9. Turrini G, Branham D, Chen L, et al. July 2021 Research Report 1 H Access to Affordable Care in Rural America: Current Trends and Key Challenges . 2021. Accessed August 4, 2025. https://aspe.hhs.gov/sites/default/files/documents/09e40880648376a13756c59028a56bb4/rural‐health‐rr.pdf
- 10. McBain RK, Schuler MS, Qureshi N, et al. Expansion of telehealth availability for mental health care after state‐level policy changes from 2019 to 2022. JAMA Netw Open. 2023;6(6):e2318045. doi: 10.1001/jamanetworkopen.2023.18045 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Cantor J, McBain RK, Kofner A, Hanson R, Stein BD, Yu H. Telehealth adoption by mental health and substance use disorder treatment facilities in the COVID‐19 pandemic. Psychiatr Serv. 2021;73(4):411‐417. doi: 10.1176/appi.ps.202100191 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Vogels E. Some Digital Divides Persist Between Rural, Urban and Suburban America. Pew Research Center; 2021. https://www.pewresearch.org/short‐reads/2021/08/19/some‐digital‐divides‐persist‐between‐rural‐urban‐and‐suburban‐america/ [Google Scholar]
- 13. Peña MT, Lindsay JA, Li R, Deshmukh AA, Swint JM, Morgan RO. Telemental health use is associated with lower health care spending among Medicare beneficiaries with major depression. Med Care. 2024;62(3):132. doi: 10.1097/MLR.0000000000001952 [DOI] [PubMed] [Google Scholar]
- 14. Fischer SH, Predmore Z, Roth E, Uscher‐Pines L, Baird M, Breslau J. Use of and willingness to use video telehealth through the COVID‐19 pandemic. Health Aff. 2022;41(11):1645‐1651. doi: 10.1377/hlthaff.2022.00118 [DOI] [PubMed] [Google Scholar]
- 15. 2019–2023. National Health Interview Survey . National Center for Health Statistics. Centers for Disease Control. Accessed August 5, 2025.https://www.cdc.gov/nchs/nhis/documentation/index.html [Google Scholar]
- 16. CDC . NHIS Methods. National Health Interview Survey. 2024. Accessed August 5, 2025. https://www.cdc.gov/nchs/nhis/about/nhis‐methods.html
- 17. Wooldridge JM. Introductory Econometrics: A Modern Approach. South‐Western Cengage Learning; 2012. [Google Scholar]
- 18. Mehrotra A, Huskamp HA, Souza J, et al. Rapid growth in mental health telemedicine use among rural Medicare beneficiaries, wide variation across states. Health Aff. 2017;36(5):909‐917. doi: 10.1377/hlthaff.2016.1461 [DOI] [PubMed] [Google Scholar]
- 19. Curtis ME, Clingan SE, Guo H, Zhu Y, Mooney LJ, Hser Y. Disparities in digital access among American rural and urban households and implications for telemedicine‐based services. J Rural Health. 2021;38(3):512‐518. doi: 10.1111/jrh.12614 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Paul J. Work from home behaviors among U.S. urban and rural residents. J Rural Stud. 2022;96:101‐111. doi: 10.1016/j.jrurstud.2022.10.017 [DOI] [Google Scholar]
- 21. Brooks MM, Mueller JT, Thiede BC. Rural‐urban differences in the labor‐force impacts of COVID‐19 in the United States. Socius. 2021;7. doi: 10.1177/23780231211022094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Callaghan T, Lueck JA, Trujillo KL, Ferdinand AO. Rural and urban differences in COVID‐19 prevention behaviors. J Rural Health. 2021;37(2):287‐295. doi: 10.1111/jrh.12556 [DOI] [PMC free article] [PubMed] [Google Scholar]