Telehealth Availability for Mental Health Care During and After the COVID-19 Public Health Emergency

This cohort study evaluates changes in availability of telehealth services at outpatient mental health treatment facilities throughout the US during and after the COVID-19 public health emergency.


Introduction
Over the course of the COVID-19 pandemic, telehealth availability expanded rapidly. 1 This change was marked and persistent for mental health care, for which most services do not require in-person physical examinations or diagnostic tests. 2 State and federal policies enacted during the pandemic promoted the telehealth transition by altering Medicare to reimburse for telehealth services 3 and state Medicaid agencies to approve reimbursement for audio-only telehealth, for example. 4 May 11, 2023, the Biden Administration declared the end of the COVID-19 public health emergency (PHE). 5Correspondingly, a range of telehealth regulations tethered to the PHE also expired on this date. 6Others, particularly those associated with Centers for Medicare and Medicaid Services (CMS), are set to expire at the end of 2024. 7In this study, we conducted a national secret shopper analysis, comparing mental health treatment facilities' (MHTF) responses about availability of telehealth before vs after the end of the PHE.

Methods
We contacted MHTFs in a 3-month period between December 2022 and March 2023 and again 6 months later between September and November 2023.Callers posed as adult clients inquiring about availability and features of telehealth services.Data were combined with aggregate facility-and county-level characteristics for analysis.This cohort study was approved by the RAND Human Subjects Protection Committee.Informed consent was waived because the study was deemed not to constitute human participants research.This report follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for observational studies.

Sampling
The sampling frame comprised outpatient MHTFs throughout the US, recorded in the Substance Abuse and Mental Health Services Administration's (SAMHSA) Behavioral Health Treatment Service Locator (BHTSL) on August 22, 2022, 8 which represents a national inventory of psychiatric facilities.
The BHTSL does not include private practices.The BHTSL is updated on a monthly basis and includes facility characteristics such as types of services offered, insurances accepted, and public vs private ownership. 9We abstracted facility address and phone number for contact purposes.
Of 9568 outpatient MHTFs within the BHTSL on August 22, 2022, we randomly selected 25% (1938) for contact in wave 1, of which we successfully contacted 1404 (72.5%) between December 2022 and March 2023.In wave 2 (September to November 2023), we recontacted 1163 that we successfully contacted in wave 1, achieving a wave 2 sample of 1001 facilities (86.1%).In both waves, facility addresses were linked to county-level information using the Health Resources and Services Administration's Area Health Resource Files, 10 including county metropolitan status (metropolitan vs nonmetropolitan), percentage of residents who are Black, percentage of residents who are Hispanic, and median household income.

Procedures
Trained callers read from a standardized script. 11,12Callers posed as prospective clients with 1 of 3 clinical conditions-randomly assigned to callers at the facility-level-for which they were seeking services: major depressive disorder, generalized anxiety disorder, or schizophrenia.Callers inquired about specific aspects of telehealth availability, as described in the next section.Callers documented facility responses in Qualtrics XM (Qualtrics).

Outcomes
Our primary outcome was a binary measure of whether the facility was currently offering telehealth (yes vs no).Among facilities reporting they offered telehealth, we also inquired about the telehealth

JAMA Network Open | Health Policy
Telehealth for Mental Health Care During and After COVID-19 modalities offered (audio-only vs video requirement) and whether facilities offered telehealth services for individuals with a mental health condition and comorbid alcohol use disorder (AUD; yes vs no).Lastly, we assessed whether facilities offering telehealth provided 3 types of services: telehealth-based psychotherapy, telehealth-based medication management, and telehealth-based diagnostic services (yes vs it depends vs no).The full protocol can be found in the eAppendix in Supplement 1.

Statistical Analysis
To examine nonresponse bias, we compared responders in waves 1 and 2 with the full sampling frame, as well as responders in waves 1 and 2 vs 1 only.We determined whether the differences between these groups was statistically significant using a χ 2 test.Among those in the analytic sample (1001 MHTFs), we reported descriptive statistics comparing survey responses for wave 1 vs wave 2.
For each survey response item, we tested for significant change across waves using univariate logistic regression models, with SEs clustered at the facility level.
For the primary outcome (any telehealth offered), we categorized facilities into 4 groups: sustainers who responded yes in both waves; nonadopters who responded no in both waves; late adopters who responded no in wave 1 and yes in wave 2; and discontinuers who responded yes in wave 1 and no in wave 2. We then conducted fixed-effects multinomial regression analysis to examine the associations between group membership and both facility-level characteristics (public vs private and accepting Medicaid vs not) and county-level characteristics (metropolitan vs nonmetropolitan, percentage of residents who are Hispanic, percentage of residents who are Black, and median household income).SEs were clustered at the state level.
All analyses were conducted in Stata version 17.0 (StataCorp).Statistical tests were 2-sided, using an α threshold of .05.Analyses were conducted in January 2024.respect to telehealth for psychotherapy, there was a significant decline in the percentage reporting yes (OR, 0.39; 95% CI, 0.31-0.48)and significant growth in the portion reporting it depends (OR, 2.62; 95% CI, 2.10-3.26).The same was observed for medication management; there was a significant decline in the percentage reporting yes (OR, 0.59; 95% CI, 0.49-0.71)and significant growth in the portion reporting it depends (OR, 1.81; 95% CI, 1.48-2.21).

Trajectories of Telehealth Availability
Overall, 674 facilities (72.0%) were in the sustainer class, 106 (11.2%) were in the nonadopter class, 66 (7.0%) were in the late adopter class, and 94 were (9.9%) in the discontinuer class.The Figure depicts the geographic distribution.We note that discontinuers and nonadopters generally appear to be concentrated in the southeastern US.

JAMA Network Open | Health Policy
Telehealth for Mental Health Care During and After COVID-19

Figure . 8
Figure.Telehealth Availability Status at Mental Health Treatment Facilities During and After COVID-19 Public Health Emergency

Table 1 .
Characteristics of Telehealth Services at Mental Health Treatment Facilities During and After Public Health Emergency (PHE) a PHE expired onMay 11, 2023.bPvalues derived from univariate logistic regression models with SEs clustered at the facility level.JAMA Network Open | Health Policy

Table 2 ,
results suggested MHTF ownership was significantly associated with class membership.Relative to public MHTFs, private for-profit MHTFs had 72.0%lower odds of being a (aOR, 0.34; 95% CI, 0.16-0.69)relative to MHTFs that did not accept Medicaid.MHTFs in communities with the highest quartile of proportion of Hispanic residents were more likely to be nonadopters than MHTFs in communities in the lowest quartile of proportion of Hispanic residents (aOR, 1.97; 95% CI, 1.07-3.64).

Table 2 .
Facility-and County-Level Differences in Trajectories of Telehealth Availability by Class For all measures, reference group for outcome is sustainers category (684 facilities).Sustainers were those that offered telehealth during and after the end of the public health emergency (PHE).Discontinuers were those that offered telehealth during but not after the end of the PHE.Nonadopters were those that did not offer telehealth during or after the end of the PHE.Adopters were those that started offering telehealth after the end of the PHE.