Provision of Mental Health Services in the Veterans Health Administration: A Nationwide Comparison With Other Providers

One of the major concerns about expanding the role of the U.S. federal government in the funding and delivery of health services is that a large, centralized care system may reduce patients’ access to specialized services (1). For example, in the current decentralized U.S. health care system, fewer patients wait >2 months to see a specialist than comparable patients in other high-income countries but with single-payer health care (2). To evaluate concerns about the quality of such care within the current landscape of U.S. health care, the Veterans Health Administration (VHA) affords a unique opportunity as the only government-operated national health system. Serving >6 million Americans annually, the VHA provides comprehensive, lifetime services, albeit to a specific population (3).

Although previous studies have compared the availability of selected specialty health care services in the VHA with availability in non-VHA health settings and have reported mixed results, few investigations have focused specifically on mental health care (4, 5). A focus on mental health care may be particularly informative because the VHA has recently made highly visible efforts to boost the quality of its mental health care in response to developments in psychiatry and psychology, such as psychiatric rehabilitation models and psychopharmacologic advances, and national social challenges that also affect veterans, such as homelessness and mass incarceration (6, 7). However, most of the existing literature on VHA mental health care is neither national in scale nor comprehensive in scope, and no study has systematically compared VHA clinical entities with equivalent local public mental health entities (8–13). One study (14), for example, compared 13 Centers of Excellence of the U.S. Department of Health and Human Services with eight VHA women’s health centers, reporting that VHA facilities were more likely to offer mental health care and sexual trauma screening and counseling, but the study’s findings were limited by reliance on small local samples.

In this study, we used national survey data from the 2018 National Mental Health Services Survey (N-MHSS) and a common systematic sampling method to compare the provision of mental health treatment modalities, specialized services, and dedicated programs for specific patient groups or diagnoses between mental health treatment facilities self-identified as VHA or non-VHA. We further attempted to document differences in two operational characteristics for which data were available: patient volumes (an indicator of workload) and licensing and accreditation.

Methods

The N-MHSS is an annual, voluntary survey of all known U.S. mental health facilities, defined by their address and participation in past surveys. VHA and non-VHA facilities were identified through the same method and identified themselves as VHA or non-VHA without specifying particular VHA administrative structures (i.e., medical centers or clinics) responsible for their operation. This approach also did not specify administrative structures in which non-VHA facilities were situated. This procedure, although imperfect, reflects the decentralized nature of the U.S. mental health “nonsystem.” The survey gathered data on diverse characteristics, including facility licensing and accreditation, patient volumes, basic treatments provided, and specialized services offered (15). Of 13,554 eligible facilities, 86.2% (N=11,682) completed the survey. Of these 11,682 facilities, 9.9% (N=1,153) were excluded because they treated only patients ages <18 years. Of the remaining 10,529 facilities, 0.2% (N=21) were excluded because of missing data, resulting in a final sample of 10,508 facilities. Because the N-MHSS is a publicly available data set with no individual patient information, institutional review board approval for this study was not required.

Measures

The 2018 N-MHSS asked facilities to complete a checklist documenting provision of mental health approaches and modalities, specialized services and practices, and treatment programs and groups dedicated or designed exclusively for clients of a particular identity or diagnosis. Additionally, the survey asked key informants about facility characteristics such as the facility type (Veterans Affairs medical center [VAMC] or separately located component entity, psychiatric hospital, residential treatment center, outpatient clinic, or community mental health center), the number of patients served, and age groups served (children vs. adults). Secondary characteristics such as licensing and accreditation status and numbers of new admissions to inpatient, outpatient, and residential treatments in the past year were also recorded.

Analyses

Facilities that identified themselves as VAMCs or other VHA health care facilities (such as community-based outpatient clinics) were then compared with two types of non-VHA facilities: those that exclusively served adults (ages ≥18 years) and those that served patients of all ages, i.e., both adults and children. For those two comparisons, two-sample t tests and Cohen’s d (to assess effect sizes) were used to compare continuous variables (the mean numbers of treatment modalities, specialized services, and special group treatments, as well as the sum of all three service types). Chi-square tests and risk ratios (to assess effect sizes) were reported for these two comparisons across 12 dichotomously defined mental health approaches and modalities (e.g., psychotropic medication or individual psychotherapy), 19 specialized services or practices (e.g., legal advocacy and integrated primary care services), and 14 dedicated programs for specific patient groups (e.g., for lesbian, gay, bisexual, and transgender [LGBT] patients, or patients with eating disorders).

Because of the large sample sizes, in which statistical significance can reflect very small and relatively unimportant differences, we used an alpha criterion for statistical significance of p<0.01 and assessed effect size differences by using Cohen’s d of greater than 0.20 or less than −0.20 and relative risks >1.5 or <0.67 to identify substantial differences (16).

To address the correlation between facilities within states and specific potentially confounding variables, we used an additional set of analyses comprising mixed-effects linear regression models with a random effect for state and fixed effects representing key facility characteristics (whether the facilities offered inpatient, outpatient, or residential services, as well as the total number of admissions per year) to compare the numbers of treatment modalities, specialized services, and special group treatments, as well as their total sum. We conducted all data analyses in R, version 4.0.0, and the lme4 package for mixed-effects modeling.

Results

Of the 10,508 facilities included in this study, 459 (4.4%) self-identified as VHA facilities, and 3,671 (34.9%) were classified as non-VHA facilities that exclusively served adults and 6,378 (60.7%) as non-VHA facilities that served all ages.

Overall, VHA facilities offered more total services (including different treatment modalities, specialized services, and dedicated programs) (mean±SD=24.2±8.9 services) than both non-VHA adult (15.4±6.8; Cohen’s d=1.11, p<0.001) and non-VHA all-ages (17.1±6.6; Cohen’s d=0.90, p<0.001) facilities.

VHA facilities offered more treatment modalities (8.5±2.3) than both non-VHA adult (6.6±2.6; Cohen’s d=0.77, p<0.001) and non-VHA all-ages (7.9±2.0; Cohen’s d=0.26, p<0.001) facilities. In particular, VHA facilities were more likely than non-VHA adult-only facilities to offer telemedicine, as well as couples and family therapy, trauma therapy, and electroconvulsive therapy (ECT), but they were substantially less likely to offer activity therapy. Compared with non-VHA all-ages facilities, VHA facilities were more likely to offer ECT and telemedicine (Table 1).

VHA facilities offered more types of specialized services (10.8±4.7) than both non-VHA adult (6.6±3.6; Cohen’s d=1.01, p<0.001) and non-VHA all-ages (6.4±3.6; Cohen’s d=1.06, p=0.007) facilities. In particular, VHA facilities were more likely than non-VHA adult-only facilities to offer therapeutic foster care, legal advocacy, integrated primary care, psychiatric emergency walk-in services, chronic disease and illness management, supported employment, vocational rehabilitation, housing services or supportive housing, consumer-run (peer support) services, diet and exercise counseling, suicide prevention, and education services (Table 2).

Compared with non-VHA all-ages facilities, VHA facilities were also substantially more likely to offer legal advocacy, integrated primary care, vocational rehabilitation, chronic disease and illness management, diet and exercise counseling, housing services or supportive housing programs, supported employment, consumer-run (peer support) services, psychiatric emergency walk-in services, education services, tobacco cessation treatment, and illness management and recovery; however, they were significantly less likely to offer therapeutic foster care and court-ordered outpatient treatment (Table 2).

For programs dedicated to specific identity groups and diagnoses, VHA facilities also offered more dedicated programs (4.9±3.6) than both non-VHA adult (2.2±2.8; Cohen’s d=0.82, p<0.001) and non-VHA all-ages (2.8±3.3; Cohen’s d=0.60, p<0.001) facilities. Compared with non-VHA adult-only facilities, VHA facilities were more likely to offer eight types of dedicated treatment programs: traumatic brain injury (TBI) treatment, LGBT support, posttraumatic stress disorder (PTSD) treatment, support for members of military families, treatment and support for Alzheimer’s disease or other dementia, eating disorder treatment, interventions for patients who have experienced trauma other than PTSD, and treatment for HIV/AIDS. However, the VHA facilities tended to be less likely to offer dedicated programs for transition-age young adults (Table 3).

Compared with non-VHA all-ages facilities, VHA facilities were substantially more likely to offer six dedicated treatment programs: TBI treatment, Alzheimer’s disease or other dementia treatments, PTSD treatment, LGBT support, senior or older adult services, and HIV/AIDS treatment; however, VHA facilities were significantly less likely to offer dedicated programs for transition-age young adults (Table 3).

Compared with both non-VHA adult-only and all-ages facilities, VHA facilities were more likely to be accredited by both the Commission on Accreditation of Rehabilitation Facilities and the Joint Commission but were less likely to be accredited by the Centers for Medicare and Medicaid Services (Table 4). VHA facilities were also more likely than non-VHA adult-only facilities to have high patient volumes (defined as having >250 hospital admissions per month or outpatient initiations per year) and more likely than non-VHA all-ages facilities to offer inpatient psychiatric services. However, of those with inpatient services, VHA facilities were less likely than non-VHA all-ages facilities to serve >30 inpatients as of April 30, 2018 (Table 4). VHA facilities were more likely than non-VHA all-ages facilities to offer 24-hour residential hospital inpatient treatment services. In contrast, VHA facilities were less likely than non-VHA adult-only facilities to offer residential treatment services; however, of those that provided these services, VHA facilities were significantly more likely than non-VHA adult-only facilities to serve >20 residential patients on April 30, 2018.

In a final set of analyses (applied to a smaller subset of facilities [N=8,127] because of missing data), we used mixed-effects linear regressions to account for clustering within states and adjusted for inpatient, outpatient, and residential service provision, as well as numbers of new admissions per year. In these analyses, VHA facilities (N=271) were still found to offer more total services, more treatment modalities, more specialized services, and more dedicated group treatments than both non-VHA adult-only facilities (N=3,010) and non-VHA all-ages facilities (N=4,846) (see the online supplement to this article).

Discussion

The findings from this analysis of N-MHSS data suggest that in 2018, VHA facilities offered more mental health treatment modalities, specialized mental health services, and dedicated treatment programs for patients with specific diagnoses and personal identities than did non-VHA facilities, even when controlling for patient volumes and treatment setting. Although not based on a comprehensive method of sampling all mental health providers nationally, which is challenging to do because of the fragmented nature of the U.S. mental health system, our results suggest that VHA offers no fewer types of service than non-VHA providers and possibly more.

Although our analysis could not provide a longitudinal perspective on the development of different services within the VHA, their extent and diversity likely reflect, at least in part, the VHA’s recent efforts to improve and expand its provision of mental health services and provide evidence-based practices by offering training in evidence-based psychotherapies, monitoring pharmacotherapeutic prescribing practices, and expanding services to homeless veterans with mental disorders (17, 18). Furthermore, the greater availability of tele–mental health services in VHA facilities reflects vigorous efforts to expand VHA’s telehealth capabilities during the past decade (19, 20). This telehealth expansion within VHA is a notable positive finding in the era of the COVID-19 pandemic because it addressed the need for social distancing and limited transportation resources, even before the pandemic began.

That integrated primary care, chronic disease management, and diet and exercise counseling were more likely to be reported at VHA facilities than at non-VHA facilities may specifically have reflected a recent primary care mental health integration initiative. Starting in 2007, all VAMCs and large community-based outpatient clinics of the VHA were required to provide specialized mental health care onsite. Since then, studies have shown not only improved detection and follow-up of diagnosed mental illness but also reductions in mortality rates and emergency department visits among veterans with mental illness (21, 22).

VHA facilities also reported being more likely to offer psychosocial rehabilitation, case management, supportive housing, and supported employment programs—services that focus on social determinants of health and include nationwide dissemination of a modified version of assertive community treatment, as well as a broad shift of the entire VHA mental health care system toward community-based care. The VHA may have been more successful in this shift than other health care systems because of its centralized commitment of resources and infrastructure through its intensive community mental health recovery program (formerly called mental health intensive case management), a supportive housing program offered by Housing and Urban Development–Veterans Affairs Supportive Housing, and its compensated work therapy and individual placement and support programs (13, 18, 23, 24). Furthermore, more VHA facilities than non-VHA facilities offered specialized psychiatric emergency services and multicomponent suicide prevention programs, driven by national efforts to reduce higher suicide rates among veterans (25, 26).

Dedicated treatment programs for patients with specific diagnoses or identities were also reported more frequently at VHA facilities than non-VHA facilities. For patients with PTSD and comorbid TBI and dementia diagnoses, this increased frequency of these programs at the VHA was likely due to the VHA’s legislative responsibility for providing treatment for the sequelae of national combat service and the older age of the veteran population, reflecting the greater personnel demands of earlier compared with more recent military conflicts (27–29).

VHA facilities were also reported to be more likely than non-VHA facilities to provide dedicated treatment programs for LGBT patients. This finding was somewhat surprising given the U.S. military’s history of LGBT discrimination (i.e., “Don’t Ask, Don’t Tell”) and studies from as recently as 2014 suggesting poor training among VHA psychotherapists for treating LGBT patients (30). However, our finding likely reflects recent efforts to expand specialized care for LGBT veterans, in particular, a central mandate to appoint at least one clinician in every facility to serve as an LGBT veteran clinical coordinator (31).

Given recent administrative initiatives to improve VHA mental health services, our findings inform public consideration of the potential effectiveness of nationally centralized health care leadership, even though these initiatives are often driven by political processes, professional standards, and national performance data (32, 33). Previous studies of VHA mental health care have similarly concluded that the VHA provides high-quality, accessible mental health care (17); however, these studies mostly relied on data from within the VHA, without non-VHA data for comparison. Our investigation was distinct, because it used common sampling and assessment procedures to compare the provision of multiple mental health services in VHA with those provided in non-VHA facilities in a national sample.

Our overarching objective was to contribute to the debate on the topic of national, centralized, and government-run versus local and privately run mental health organizations. The national survey data examined here suggest that, on average, the VHA may provide more, and certainly not fewer, comprehensive mental health services within single facilities than do non-VHA organizations. However, one cannot conclude from these results that VHA facilities outperform non-VHA facilities in terms of access to or quality of care. Although non-VHA facilities reported offering on average fewer services than VHA facilities, nonveteran patients often receive treatment from several different facilities (e.g., inpatient treatment from a hospital and outpatient management from a private provider). Thus, non-VHA mental health services may be more divided among different facilities, but they are not necessarily less available than VHA mental health services.

Several methodological limitations of this study deserve comment. First, participant facilities were identified and selected for the N-MHSS on the basis of a database maintained by the Substance Abuse and Mental Health Services Administration (SAMHSA) of known treatment facilities that is updated with assistance from state mental health agencies, professional mental health organizations, and individual requests from facilities themselves. However, discussions with SAMHSA staff clarified that although active efforts are made to update the list of facilities regularly, the fragmented nature of the U.S. mental health system precludes definitive validation of the comprehensiveness and inclusiveness of the facilities surveyed (15).

Second, VHA facilities were not asked to specify their place in the national VHA administrative structure, i.e., as medical centers, freestanding clinics, or community-based outpatient clinics. Although VHA facilities therefore cannot be identified as specific components in the administrative organization of the VHA, this limitation also applied to non-VHA facilities and was offset by the major advantages of the facilities having been identified through a common nationwide sampling effort and evaluated with common measures. Furthermore, in view of the robustness of the observed VHA–non-VHA differences, it is unlikely that our results reflect a systematic sampling bias in which VHA facilities with extensive services were oversampled and those with fewer services were excluded. We cannot rule out such a bias in the sampling, but it seems unlikely that such bias would have had any more effect on data from VHA facilities than on data from non-VHA facilities. At the very least, the strong associations between VHA facilities and extensive service delivery make it unlikely that VHA, as a national system of care, provides fewer specialized services than locally based systems.

Third, because the N-MHSS 2018 was a voluntary, self-reported survey by facility administrators, the estimates of patient volumes may not have reflected precise workload measurements. An additional limitation was that data on service provision were based on information provided by key informants and were not independently validated. However, we again consider it unlikely that potential biases in reporting would have affected reports on VHA service delivery to a greater extent than reports from non-VHA facilities. Furthermore, the survey data addressed only whether specific services were offered, without addressing the quality of such services, measured, for example, as patient satisfaction, outcomes, or accessibility. The N-MHSS does not include cost data, precluding comparison of the efficiency of service delivery.

Finally, it was not possible to detail the many centralized VHA quality improvement initiatives relevant to each indicator on the survey—both because these initiatives were too numerous for a comprehensive accounting in this study and because some initiatives have not been described in the literature. A plausible explanation of the apparent advantage of a national health care system is that once emerging clinical needs (e.g., homelessness, incarceration, and opioid addiction) are identified, evidence-based treatments and solutions can be effectively disseminated through systemwide funding and educational and performance-monitoring initiatives within a well-established administrative network (18).

Conclusions

Despite the aforementioned limitations, these findings are noteworthy for their national scale and suggest that VHA facilities provide not fewer and possibly more comprehensive specialized mental health services within single facilities than do non-VHA mental health facilities—perhaps reflecting a distinct advantage of efforts to maintain quality in a centralized, integrated health system, at least in the area of mental health service delivery.