Receipt of Psychotherapy Following PTSD Diagnosis in VA and Community Settings

Abstract

The likelihood of receiving psychotherapy following PTSD diagnosis has been linked to patient and diagnosing clinician characteristics. This study sought to assess how the probability of receiving psychotherapy following an incident diagnosis of PTSD is influenced by patient rurality, location of diagnosis (VA or community care), and clinician type. We examined a cohort of veterans (n=663,256) who received an incident diagnosis of PTSD at a VA or community health care facility between 2019 and 2023. Patients were stratified by rurality, location of diagnosis, and clinician type. Covariates included patient demographics, psychiatric and medical comorbidities, prior VA service use, and prior psychiatric treatment. Probabilities of receiving psychotherapy following PTSD diagnosis were calculated using logistic regression with a three-way interaction of patient rurality, diagnosing clinician, and location of diagnosis. Compared to those diagnosed in the VA, patients diagnosed in the community were more frequently rural, less frequently diagnosed in an outpatient setting, and less frequently diagnosed by social workers, psychologists, psychiatrists, and non-physician prescribing clinicians. Across all rurality strata patients diagnosed by mental health clinicians had a high probability of receiving psychotherapy regardless of location of diagnosis. In contrast, across all rurality strata patients diagnosed by non-mental health providers had significantly lower probability of receiving psychotherapy when diagnosed in the community instead of the VA. PTSD diagnoses inside the VA are given by different clinician types than those given in the community, and, in some cases, this is associated with lower initiation of psychotherapy following diagnosis.

Introduction

The Department of Veterans Affairs (VA) is the largest health care system in the United States, comprised of 172 medical centers and 1,100 community-based outpatient clinics (CBOC) (Albanese et al., 2020), and is charged with providing primary care, mental health care and specialty care to 13.8 million urban and 4.4 million rural veterans. The VA mission is to provide the highest quality of care to veterans that is timely and geographically accessible. This presents significant challenges to rural veterans, who often need to travel long distances to obtain care at a VA medical center or outpatient clinic. To increase access to timely care, Congress passed the Veterans Access, Choice, and Accountability (CHOICE) Act of 2014 (Stroupe et al., 2019), and the VA Maintaining Internal Systems and Strengthening Integrated Outside Networks (MISSION) Act of 2018. These legislations enabled veterans who have long travel times to VA facilities to have the VA pay for care at non-VA affiliated community healthcare facilities that are spatially and temporally more accessible (“community care”). This approach appears to have been adopted by rural veterans for at least some of their healthcare. Following implementation of the MISSION Act, veteran outpatient care increased at community clinics, with rural clinics seeing greater proportional increases in visits for primary care and emergency/urgent care than urban clinics (Yoon et al., 2024). Therefore, policies like the MISSION Act, which mandate access to community care, may have important implications for the delivery of mental health care to rural veterans (Aggarwal, 2020).

Care for posttraumatic stress disorder (PTSD) is a priority for the VA (PTSD: National Center for PTSD. How Common is PTSD in Veterans?, 2025). In response to clinical need, the VA and the Department of Defense have created Clinical Practice Guidelines (VA/DOD Clinical Practice Guidelines: Management of Posttraumatic Stress Disorder and Acute Stress Disorder 2023, 2023) that recommend medications and psychotherapies that are condition-specific, patient-centered, minimize complications and optimize quality of life. However, recommended psychotherapies are underutilized and initiated variably across geographic regions (Dufort et al., 2021). Initiation and retention of psychotherapy treatment for PTSD seems to, at least partially, vary on patient demographics and location where diagnosis is made. Spoont et al. found that veterans diagnosed with PTSD in a PTSD or general mental health clinic had greater odds of receiving psychotherapy compared to those diagnosed in a general medical or primary care clinic (Spoont et al., 2010). Several studies have also found associations between patient demographics, comorbidities, and treatment history with receipt of psychotherapy following PTSD diagnosis at the VA (Doran et al., 2017; Spoont et al., 2017). Compounding this, utilization of psychotherapy for PTSD has been found to be low amongst non-VA community clinicians (Finley et al., 2018). These prior studies have highlighted that factors limiting psychotherapy initiation following a PTSD diagnosis may be more pronounced in VA-paid community care settings.

When the VA increased geographic access to mental health care through VA CBOC’s and community care, rural veterans increased utilization of those resources (Shiner, Gottlieb, et al., 2022). However, veterans overall use of community care for outpatient behavioral healthcare following the implementation of VA CHOICE remained low. The patients that did utilize that care displayed characteristic differences from patients that relied solely on the VA for outpatient behavioral healthcare and were more likely to receive care from non-doctoral clinicians (Vanneman et al., 2023). As an example, rural veterans received mental health care, including psychotherapy, from non-doctoral clinicians like social workers and non-physician prescribing clinicians (NPPCs); physician assistants, nurse practitioners, and pharmacists; at greater rates than their metropolitan counterparts (Rozema et al., 2024). Therefore, given 1) community care was implemented with an explicit focus on rural veterans; 2) receipt of psychotherapy following PTSD diagnosis has been associated with patient demographics and the source of diagnosis; 3) utilization of psychotherapies for PTSD appear to be low amongst non-VA community clinicians; and 4) behavioral healthcare delivered by non-VA community clinicians is more often delivered by non-doctoral clinicians; the interest of the present study was to investigate the influence patient rurality, location of diagnosis (VA or community), and clinician type had on the probability of receiving psychotherapy following an incident diagnosis of PTSD.

Methods

This study was approved by the Research and Development Committee at the White River Junction VA Medical Center. It was also reviewed by the Veteran’s Institutional Review Board of Northern New England and designated as exempt human subjects research under Exemption Category 4(iii) of the 2018 Common Rule.

Cohort Definition

Our study was performed in a cohort of VA patients receiving VA provided or VA-paid care between fiscal years 2008 and 2023. Within this cohort we identified patients with a PTSD diagnosis (ICD-9: 309.81; ICD-10 F43.1x), defining a patient’s earliest PTSD diagnosis as their “incident” diagnosis of PTSD. This cohort was then limited to patients whose incident PTSD diagnosis occurred between fiscal years 2019 and 2023, and whose diagnosis was given by a clinician with diagnosing capabilities (physician, NPPC, psychologist, social worker, counselor, etc.). This period was chosen to align with the availability of the Integrated Veteran Care Consolidated Data Sets (IVC CDS), which contains comprehensive data on VA-paid community care claims processed on or after October 1, 2018. The incident diagnosis was further allowed to be given in either an outpatient or inpatient setting. Additionally, the patients were also required to have a designated address in the year of diagnosis for rurality assignment. Patients were categorized into three rurality bins based on Rural-Urban Commuting Area (RUCA) codes linked to their home address and corresponding to the University of Washington’s Categorization B classification: rural (7.0, 7.2, 7.3, 7.4, 8.0, 8.2, 8.3, 8.4, 9.0, 9.1, 9.2, 10.0, 10.2, 10.3, 10.4, 10.5, 10.6), micropolitan (4.0, 4.2, 5.0, 5.2, 6.0, 6.1), or metropolitan (1.0, 1.1, 2.0, 2.1, 3.0, 4.1, 5.1, 7.1, 8.1, 10.1).

Covariates were selected based on prior work in this cohort (Rozema et al., 2024). They included patient demographics at the time of diagnosis, any service-connected disability percent over 70%, diagnosing clinician specialty, location of diagnosis (VA or VA-paid community care), psychiatric and medical comorbidities in the two years preceding diagnosis, VA service use in the year preceding diagnosis, and psychiatric treatment in the year preceding diagnosis (Table 1). Era of service and combat status were determined through their corresponding flags in the patient demographics and encounters data within the VA corporate data warehouse. These fields were not mutually exclusive. Diagnosing clinician type was determined through National Uniform Claim Committee (NUCC) taxonomy codes. Frequencies of clinician types included in user-created bins (i.e. non-psychiatrist physicians) are included in the supplementary materials (Supplemental Table 1 & 2). Standardized mean differences (SMDs) were used to compare covariate profiles between rurality and source of diagnosis strata.

Table 1.

Explanation of Covariates

Treatment History in the Year Preceding Diagnosis
VA Antidepressant	Receipt of any antidepressant treatment from CN609 list
VA Psychotherapy	Receipt of any psychotherapy sessions
Service Use in the Year Preceding Diagnosis
VA Primary Care Visit	Visit with Primary Care Stop Code + primary provider that is a physician, NP, or PA
VA Mental Health Visit	500 series stop code
Patient Characteristics at Diagnosis
Age	Continuous
Sex	Categorical male or female
Race	Categorical Black Non-Hispanic, Hispanic, White Non-Hispanic, Other
Marital Status	Binary married vs. not
Military Exposures	Combat, Service-Connected Disability, Era of Service
Diagnoses in the Two Years Preceding Diagnosis
Psychiatric Comorbidities	Peltzman I-6 Index (Peltzman et al., 2023) – categorical low, medium high
Medical Comorbidities	Categorical low, medium, high
Diagnosis Characteristics
Diagnosing Provider Characteristic	Provider type based on NUCC taxonomy codes
Diagnosis Outpatient vs. Inpatients	Diagnosed in an outpatient or inpatient setting

Note: NP = Nurse Practitioner, PA = Physician Assistant, NUCC = National Uniform Claim Committee

Psychotherapy Receipt Following Diagnosis

Instances of outpatient psychotherapy provided at the VA or in the community were identified in the year following diagnosis using the CPT/HCPCS procedure codes provided in the supplementary material (Supplemental Table 3). Our definition for what was considered an outpatient claim within the IVC/CDS data is also included in the supplementary material (Supplemental Table 4). Patients were then assigned a flag indicating whether they received any or no psychotherapy in the year following diagnosis in either the VA or community care.

Our binary outcome was assessed using logistic regression controlling for all covariates included in Table 1 and included a three-way interaction term between patient rurality, location of diagnosis, and diagnosing clinician type. Missingness was present in several categorical covariates and was included as its own level of the variable. Predictive margins were calculated and reported with 95% confidence intervals. Marginal effects at the means were also calculated and reported with Bonferroni corrected confidence intervals. The predictive margins can be interpreted as the predicted probability of the event occurring (i.e., receiving psychotherapy) holding all covariates at their means, while the marginal effect at the mean is the difference in those probabilities at two levels of our predictor of interest (i.e. patient rurality).

Data set creation and statistical modeling was performed in SAS 9.4. Predictive margins and marginal effects at the means were estimated using the margins macro provided by the SAS Institute.

Sensitivity Analyses

We performed two sensitivity analyses to assess the impact of our cohort specifications on receipt of psychotherapy in the year following diagnosis. First, we limited our cohort to patients that did not receive psychotherapy at the VA in the year prior to diagnosis to eliminate the possibility that instances of psychotherapy occurring after diagnosis are merely a continuation of an existing treatment regimen. Second, we limited our cohort to include outpatient PTSD diagnoses only.

Results

Patients Diagnosed with PTSD in the VA and the Community

663,256 patients met our inclusion criteria of an incident diagnosis of PTSD between FY 2019 and 2023. Of these, 565,608 (85.28%) were metropolitan, 57,273 (8.64%) were micropolitan, and 40,375 (6.09%) were rural. 60,756 (9.16%) of the cohort was diagnosed within the community.

Patients residing in rural and micropolitan settings were largely similar while race, age, and Vietnam era of service had small differences between rural/micropolitan and metropolitan patients. Comparing rural and metropolitan veterans, rural veterans were on average older, less often female, more often White or American Indian/Alaskan Native versus other race groups, and more often Vietnam Veterans (Table 2). In addition to demographics, rurality influenced the characteristics of patient’s diagnostic care. Compared to metropolitan patients, rural patients were more often diagnosed by a community clinician and that diagnosis was provided less often by a psychologist or psychiatrist and more often by an NPPC.

Table 2.

Patient Characteristics by Rurality

	Metropolitan		Micropolitan		Rural		SMD
	565,608		57,273		40,375		Metropolitan vs Micropolitan	Metropolitan vs Rural	Micropolitan vs Rural
	n	%	n	%	n	%
Diagnosis Characteristics
Diagnosed in the Community	48,808	8.6	6,503	11.4	5,445	13.5	0.09	0.16	0.07
Outpatient Diagnosis	531,802	94	53,505	93.4	37,345	92.5	0.03	0.06	0.04
Diagnosing Provider Type
Social Worker	87,525	15.5	11,088	19.4	7,533	18.7	0.10	0.09	0.02
Psychologist	97,835	17.3	7,307	12.8	5,068	12.6	0.13	0.13	0.01
Psychiatrist	94,353	16.7	7,374	12.9	4,740	11.7	0.11	0.14	0.04
Non-Psychiatrist Physician	169,812	30	15,590	27.2	11,628	28.8	0.06	0.03	0.04
Other Mental Health Provider	23,026	4.1	2,368	4.1	1,751	4.3	0.00	0.01	0.01
NPPC	93,057	16.5	13,546	23.7	9,655	23.9	0.18	0.19	0.01
Patient Demographics
Age, M±SD	46.97±15.80		50.01±16.46		52.03±16.55		0.19	0.31	0.12
Female	110,258	19.5	8,561	14.9	5,474	13.6	0.12	0.16	0.04
Race
Hispanic	66,749	11.8	3,103	5.4	1,555	3.9	0.23	0.30	0.08
White	280,941	49.7	39,903	69.7	29,893	74.0	0.42	0.52	0.10
Black	138,634	24.5	7,048	12.3	3,854	9.5	0.32	0.41	0.09
American Indian/Alaskan Native	6,746	1.2	979	1.7	999	2.5	0.04	0.10	0.05
Asian/Pacific Islander	20,021	3.5	961	1.7	475	1.2	0.12	0.16	0.04
Missing	52,517	9.3	5,279	9.2	3,599	8.9	0.00	0.01	0.01
Vietnam Veteran	77,638	13.7	11,477	20.0	9,968	24.7	0.17	0.28	0.11
OEF/OIF Veteran	132,719	23.5	12,710	22.2	8,580	21.3	0.03	0.05	0.02
Combat Veteran	244,371	43.2	22,453	39.2	14,757	36.5	0.08	0.14	0.06
Service-Connected Disability of +70%	264,217	46.7	24,188	42.2	16,103	39.9	0.09	0.14	0.05
	Metropolitan		Micropolitan		Rural		Metropolitan vs Micropolitan	Metropolitan vs Rural	Micropolitan vs Rural
Missing Service-Connected Disability	694	0.1	55	0.1	28	0.1	0.01	0.02	0.01
Married	277,674	49.1	30,827	53.8	22,309	55.3	0.10	0.12	0.03
Missing Patient Marital Status	24,258	4.3	2,289	4.0	1,488	3.7	0.02	0.03	0.02
Treatment History
Prior VA Primary Care Use	332,290	58.7	35,644	62.2	25,363	62.8	0.07	0.08	0.01
Prior VA Mental Health Care Use	294,791	52.1	27,045	47.2	18,170	45.0	0.10	0.14	0.05
Prior VA Antidepressant Use	151,785	26.8	16,173	28.2	11,633	28.8	0.03	0.04	0.01
Prior VA Psychotherapy Use	178,691	31.6	16,209	28.3	10,765	26.7	0.07	0.11	0.04
Medical Comorbidities, M±SD	1.12±1.84		1.33±1.96		1.40±1.99		0.11	0.15	0.03
Psychiatric Comborbidities, M±SD	0.74±0.93		0.69±0.89		0.66±0.86		0.06	0.09	0.04

Note: SMD = Standardized Mean Difference, NPPC = Non-Physician Prescribing Clinicians, OEF/OIF = Operation Enduring Freedom/Operation Iraqi Freedom

Compared to veterans diagnosed with PTSD in the VA, those diagnosed in the community were older, more frequently White, less frequently Black, more often a Vietnam veteran and less often an Operation Enduring/Iraqi Freedom (OEF/OIF) or combat veteran (Table 3). Those diagnosed in the community also had more physical and psychiatric comorbidities, with more frequent prior use of primary care, mental health care, antidepressants, and psychotherapy at the VA. Community care patients were diagnosed much less often in the outpatient setting, and these diagnoses were provided more often by non-psychiatrist physicians or other mental health clinicians. In addition, community care patients were diagnosed by social workers, psychologists, psychiatrists and NPPC less often than VA patients.

Table 3.

Patient Characteristics by Source of Diagnosis

	VA		Community		SMD
	602,500		60,756
	n	%	n	%
Diagnosis Characteristics
Outpatient Diagnosis	585,738	97.2	36,914	60.8	1.00
Diagnosing Provider Type
Social Worker	101,226	16.8	4,920	8.1	0.27
Psychologist	106,337	17.6	3,873	6.4	0.35
Psychiatrist	100,038	16.6	6,429	10.6	0.18
Non-Psychiatrist Physician	171,646	28.5	25,384	41.8	0.28
Other Mental Health Provider	12,136	2.0	15,009	24.7	0.71
NPPC	111,117	18.4	5,141	8.5	0.30
Patient Demographics
Age, M±SD	47.22±15.87		50.78±16.51		0.22
Female	114,182	19.0	10,111	16.6	0.06
Race
Hispanic	65,376	10.9	6,031	9.9	0.03
White	314,397	52.2	36,340	59.8	0.15
Black	138,605	23.0	10,931	18.0	0.12
American Indian/Alaskan Native	7,769	1.3	955	1.6	0.02
Asian/Pacific Islander	19,591	3.3	1,866	3.1	0.01
Missing	56,762	9.4	4,633	7.6	0.06
Vietnam Veteran	86,763	14.4	12,320	20.3	0.16
OEF/OIF Veteran	143,802	23.9	10,207	16.8	0.18
Combat Veteran	261,068	43.3	20,513	33.8	0.20
Service-Connected Disability of +70%	274,678	45.6	29,830	49.1	0.07
Missing Service-Connected Disability	711	0.1	66	0.1	0.00
Married	302,827	50.3	27,983	46.1	0.08
	VA		Community		SMD
Patient Marital Status Missing	25,635	4.3	2,400	4.0	0.02
Medical History
Prior VA Primary Care Use	347,119	57.6	46,178	76.0	0.40
Prior VA Mental Health Care Use	300,913	49.9	39,093	64.3	0.29
Prior VA Antidepressant Use	153,330	25.4	26,261	43.2	0.38
Prior VA Psychotherapy Use	180,582	30.0	25,083	41.3	0.24
Physical Comorbidities, M±SD	1.09±1.80		1.84±2.32		0.37
Psychiatric Comorbidities, M±SD	0.70±0.91		1.03±0.98		0.35

Note: SMD = Standardized Mean Difference, NPPC = Non-Physician Prescribing Clinicians, OEF/OIF = Operation Enduring Freedom/Operation Iraqi Freedom

Receipt of Psychotherapy of PTSD Patients in the VA and Community

Several covariates were associated with receipt of psychotherapy (Figure 1). Of note, receipt of psychotherapy at the VA in the year preceding diagnosis was associated with increased odds (OR: 2.41; 95% CI: 2.37–2.46) of receiving psychotherapy following diagnosis. Additionally, prior VA mental health care (OR: 1.29; 95% CI: 1.26–1.31), being female (OR: 1.28; 95% CI: 1.26–1.31), being diagnosed in an outpatient setting (OR: 1.08; 95% CI: 1.05–1.11) and having higher mental health morbidity (OR: 1.10; 95% CI: 1.09–1.19) were all associated with increased odds of receiving psychotherapy. In contrast, prior VA primary care use (OR: 0.84; 95% CI: 0.83–0.85), OEF/OIF era service (OR: 0.84; 95% CI: 0.82–0.85), Vietnam era service (OR: 0.80; 95% CI: 0.78–0.82), and any service-connected disability percent of at least 70% (OR: 0.77; 95% CI: 0.76–0.78) were all associated with a decreased odds of receiving psychotherapy.

Figure 1. Covariate Odds Ratios.

Note: Error bars denote 95% confidence intervals.

Across all rurality strata the probability of receiving psychotherapy was high when diagnosed by a psychologist, social worker, or other mental health clinician in both the community and VA (0.831 to 0.915) while probabilities were moderately lower when diagnosed by a psychiatrist (0.665 to 0.751). Inversely, across all rurality strata patients diagnosed by a non-psychiatrist physician or NPPC had significantly lower probabilities of receiving psychotherapy in both the community and VA (0.384 to 0.597) (Figure 2). Comparing individual clinician types revealed significant differences between patients diagnosed in the community and those diagnosed in the VA across all rurality strata. First, patients diagnosed by non-psychiatrist physicians or NPPC in the community were significantly less likely to receive psychotherapy than their counterparts diagnosed in the VA. Specifically, patients diagnosed by a non-psychiatrist physician in the community had a 0.154 to 0.183 lower probability of receiving psychotherapy compared to those diagnosed by non-psychiatrist physicians in the VA depending on the rurality of the patient. Similar disparities were seen in patients diagnosed by NPPC (0.092 to 0.143) (Figure 3).

Figure 2. Probability of Receiving Psychotherapy by Location, Clinician Type, and Patient Rurality.

Note: Error bars denote 95% confidence intervals.

Figure 3. VA-Community Difference in Probability of Receiving Psychotherapy by Provider Type and Patient Rurality.

Note: Error bars denote Bonferroni corrected confidence intervals.

Sensitivity Analyses

Limiting to patients that had no prior psychotherapy resulted in several additional notable effects on top of those seen in the primary analyses. Specifically, patients diagnosed in the community by social workers had a 0.078 to 0.138 higher probability of receiving psychotherapy than those diagnosed by the same clinician type in the VA. Likewise, patients diagnosed in the community by other mental health clinicians had a 0.111 to 0.144 higher probability of receiving psychotherapy than those diagnosed in the VA. Limiting to outpatient diagnoses yielded similar new results, but to a lesser effect (Supplementary Tables 5–8).

Discussion

In this analysis of veteran PTSD care, we assessed the impacts of rurality, location, and clinician type on probability of receiving psychotherapy following an incident diagnosis of PTSD. In our cohort of veterans with an incident diagnosis of PTSD differences based on rurality were largely expected and in line with past descriptions of these strata (Rozema et al., 2024). Of note, we found that compared to metropolitan veterans, rural veterans were more frequently diagnosed with PTSD by community clinicians and that those clinicians were more often non-doctoral versus a psychiatrist or psychologist. This is in keeping with a recent report on primary and emergency care in the community and the results of previous work comparing differences in providers of mental health care based on patient rurality (Rozema et al., 2024; Yoon et al., 2024). Collectively, these studies demonstrate that rural veterans are taking advantage of legislation like the MISSION Act for their healthcare, and that the clinicians delivering that care tend to be clinicians with less specialized training.

When stratifying by location of diagnosis several interesting differences appeared with some contrasting with previous descriptions of veterans utilizing VA CHOICE for behavioral or mental healthcare (Carlo et al., 2024; Vanneman et al., 2023). Compared to veterans diagnosed inside the VA, those diagnosed in the community were more often rural and were diagnosed more often by non-psychiatrist physicians and other mental health clinicians rather than psychiatrists, psychologists, social workers and NPPC. Additionally, community care patients also tended to be older, have more comorbidities, were less often seen in an outpatient setting, and utilized VA services at higher rates prior to diagnosis. Our work demonstrates that across the spectrum of all VA-paid PTSD care, rural patients and those diagnosed in the community have demonstrably different care experiences than metropolitan patients and those receiving care inside the VA. Observed differences between VA and community patients present important considerations for care delivery in the two environments. Veterans receiving community care tended to be older, sicker, and were diagnosed less frequently in an outpatient setting. While not evaluated in this study, these characteristics predict higher costs of care for a vulnerable patient population treated in a more acute setting. Future work will be required to determine if these differences translate to inferior outcomes such as hospitalization mortality rates. The iterative improvements to mental health care delivery within the VA were driven by policy changes that improved access to care (Shiner, Gottlieb, et al., 2022), combined with measurement strategies that drove improvements in the quality of that care (Trafton et al., 2013). As the VA shifts to a model that increasingly relies on community partners to provide access, there may be a need for analogous measurement strategies to monitor and drive improvements to the quality of care for VA patients receiving treatments in community care settings. This will be challenging given that the VA quality measurement strategy has been highly reliant on the use and adaptation its internal electronic medical record (Shiner, Levis, et al., 2022), and it will be difficult to achieve this level of standardization across many community partners.

A recognized quality of care measure for PTSD is the ability of patients to receive psychotherapy following diagnosis (Forehand et al., 2022; Grubbs et al., 2017; Schnurr et al., 2024). It is not surprising that patients were more likely to receive psychotherapy if they were diagnosed by a psychologist, social worker, or other mental health clinician since this is a focus of care delivery for these clinicians. These findings are consistent with past work within the VA identifying lower odds of psychotherapy initiation when diagnosed within a general medical clinic as opposed to a mental health or PTSD specialty clinic (Spoont et al., 2010). Novel to our study is that this disparity appears to be greater when diagnosed in a community setting. It is concerning that patients diagnosed by non-psychiatrist physicians in the community had a significantly lower probability of receiving psychotherapy compared to their counterparts within the VA. These clinicians, who may not have specialized mental health training, also provided a higher proportion of diagnoses in the community compared to the VA. Thus, providing diagnostic care by less specialized mental health clinicians in the community translates to poorer follow-up care for these veterans. Compounding this, clinicians within the community that provide specialized PTSD treatment are relatively rare (Finley et al., 2018). In these cases, it might be helpful for community-based clinicians to have a mechanism to refer patients for more specialized mental health service available in the VA, either at a local VA facility or using telehealth when the required service is unavailable or inconvenient locally (Rice et al., 2026). Together, this highlights the need for continued investment in VA specialty care alongside expansions to VA-paid community care to ensure Veterans receive adequate care.

Several limitations exist in our study. First, the IVC-CDS data does not contain information on community claims prior to October 1, 2018. This results in a limited ability to control for mental health community care utilization within our cohort prior to this date. Offsetting this, prior studies indicate that utilization of community care for behavioral and mental healthcare during this period was low (Carlo et al., 2024; Vanneman et al., 2023). Additionally, in cases of routine care veterans are required to obtain a referral from their existing VA care team prior to receiving care in the community, making it likely that patients receiving an incident diagnosis in the community have some prior data available within the VA. Second, our study focused solely on receipt of psychotherapy following PTSD diagnosis. It is possible that other treatment options for PTSD, like antidepressants, are being utilized more often by non-psychiatric clinicians in the community. However, prior work found only mild differences in rates of antidepressant and psychotropic medication receipt in patients diagnosed in specialty vs. general outpatient clinics at the VA (Spoont et al., 2010). Furthermore, due to limitations with the claims-based IVC-CDS data, we were unable to identify if the psychotherapy being provided was evidence-based psychotherapy for PTSD (i.e. prolonged exposure), or some other form psychotherapy. Future work should investigate the utilization of evidence-based psychotherapy and non-psychotherapy treatments for PTSD in the community.

Providing comprehensive PTSD care requires leveraging a diverse mental health clinical workforce and providing access to care regardless of the veteran’s location. Recent initiatives like the MISSION Act aim to increase access and timeliness of care by expanding the VA workforce to include non-VA clinicians in the community. There is evidence that this approach has helped reduce geographic disparities in access to care, particularly in rural areas. However, it is just as important to assess whether the type of care is equitable regardless of the geographic location or healthcare environment. Here, we show that veteran PTSD care provided by the VA community care network is different than the care provided inside the VA; delivered by different clinician types and in some cases resulting in lower utilization of the psychotherapy. Given the established link between clinical mitigation of PTSD symptoms and risk of death by suicide (Forehand et al., 2022), the VA should develop strategies that address these limitations in care delivery when Veterans receive their PTSD care in the community through strengthening of referral mechanisms available to community providers.

Public Significance Statement:

VA and Department of Defense Clinical Practice Guidelines recommend psychotherapies as an effective treatment for veterans with posttraumatic stress disorder. Our study suggests that veterans newly diagnosed with PTSD are less likely to receive this psychotherapy when that diagnosis originates from a non-mental health clinician. This disparity appears to be more severe in VA-paid community care settings.