Comparisons between Patients Diagnosed with PTSD in Primary Care versus Mental Health Care in Five Large Civilian Health Care Systems

Abstract

Posttraumatic stress disorder (PTSD) is a serious mental health disorder that may not be adequately detected or treated in primary care (PC). The purpose of this study was to compare the clinical characteristics and health care utilization of PTSD patients diagnosed in PC versus in specialty mental health care (MHC) across five large, civilian, not-for-profit healthcare systems. Electronic claims and medical record data on patients treated during 2014 were analyzed. Treatment was considered in terms of initiation and dose (i.e., psychotherapy sessions; pharmacotherapy – prescription psychotropics). Of 5,256 patients age 15-88 with a diagnosis of PTSD, 84.4% were diagnosed by a MHC provider. Patients diagnosed by MHC providers had 4 times the rate of and more enduring psychotherapy than those diagnosed by PC providers. Receipt of psychotropcis varied by provider type, with generally higher prescription fill levels for patients in MHC. Strategies to better align patient needs with access and treatment modality in PC settings are needed.

Introduction

Exposure to potentially traumatic events, such as physical or sexual abuse/assault, serious accidental injury, terrorism, and mass shootings, is a significant public health problem (Magruder et al., 2017). In population-based surveys in 24 countries across six continents, 70% of individuals reported experiencing at least one traumatic event across their lifespan, with 30% reporting exposure to four or more such experiences. Although the majority of individuals who experience trauma will maintain or regain mental health, a significant number will develop posttraumatic stress disorder (PTSD) or other trauma-related conditions that could benefit from care (Roberts et al., 2011).

Untreated PTSD likely places a considerable burden on the health care delivery system. Individuals with PTSD use health services more often than those in almost all other psychiatric groups (Wang et al., 2005). In fact, there are several small studies in civilian populations showing that, even after controlling for age, depression, and chronic illness, those who screened positive for or had a PTSD diagnosis had more physical health complaints and higher health care costs due to more hospitalizations and longer lengths of stay (e.g., Kartha et al., 2008; Klassen, Porcerelli, & Markova, 2013; Walker et al., 2003).

The majority of research on PTSD in primary care and efforts to address it has been conducted in U.S. Department of Veterans Affairs (VA) or military settings. The VA is the largest integrated health care organization in the U.S., and likely one of the largest providers of PTSD care in the world. Over the past decade, the VA has made tremendous efforts and progress in implementing evidence-based models of integrated mental health services nationally in community based outpatient clinics, home-based primary care, and outpatient primary clinics at medical facilities (Zeiss & Karlin, 2008) as well as an unprecedented effort to training and supporting their workforce in the use of evidence-based psychotherapies, including for PTSD (Karlin & Cross, 2014). Since the VA has specialized PTSD programs and expertise that civilian health care systems likely do not have, those findings may not be generalizable to other health care systems, but can serve to inform research hypotheses in the civilian sector. Finding from research conducted in VA may differ from other health care systems due to operational and population differences. For example, it is likely that other health care systems in the U.S. are less coordinated and resourced as well more fragmented than VA. In addition, VA patients likely differ from civilians on their demographic and clinical characteristics, with much of veteran trauma being deployment-related. Thus the prevalence, duration and spectrum of severity of PTSD might diverge as well.

Even given the tremendous resources, VA still has issues in providing veterans with PTSD quality care. For example, using VA administrative data from over 20,000, predominately male veterans, Spoont et al. (2010) found that outpatient mental health treatment initiation varied among veterans with a recent PTSD diagnosis according to the type of VA clinic in which they received the diagnosis (i.e., PTSD specialty program or general mental health clinic). The odds of receiving any psychotherapy, as well as those of receiving at least eight sessions, were highest for individuals given a diagnosis in dedicated/specialized PTSD program. Whereas the odds of receiving pharmacotherapy were highest among patients given the diagnosis in a general mental health clinic. Thus, receipt of a PTSD diagnosis in a trauma specialty or mental health clinic likely conferred some advantage in appropriate mental health treatment initiation and receipt of at least minimally adequate treatment for this condition.

Less is known about the usual course of treatment and predictors of treatment utilization for patients with PTSD in civilian health care settings. For example, more than half (56%) of patients in an academic medical center had at least one therapy visit within the six months of receiving a PTSD diagnosis, although only 10% had 12 or more therapy visits (Nobels, et al., 2017). Overall, just over one fourth of patients (28%) did not receive any treatment, approximately 20% had medication alone, while nearly 30% had therapy alone, and 25% received both therapy and medication. To gain a better understanding of where patients with PTSD are diagnosed in the civilian sector, and what mental health treatment they receive, we utilized data from the Mental Health Research Network (MHRN). This is a consortium of public-domain research centers affiliated with 13 large, not-for-profit health systems that each provide comprehensive care (HCSRN, 2017). In this retrospective cohort study, we provide a description and analysis of patients diagnosed with PTSD and their follow-up for one year.

Utilizing claims and medical record data on patients treated during 2014 in five, large, civilian, non-profit integrated health care systems, we: (1) report the prevalence of PTSD diagnosed in primary/other versus mental health care settings; (2) examine sociodemographic characteristics; (3) characterize psychiatric comorbidities associated with PTSD; and (4) identify correlates of the number of psychotherapy sessions attended and of psychiatric medication use.

Method

Data were derived from the Virtual Data Warehouse (VDW) of the Health Care Systems Research Network (HCSRN) for the year 2014 from five member sites across the U.S., providing an examination of patients being treated for PTSD. The HCSRN develops and applies uniform variable definitions from claims and electronic medical record systems to conduct public-domain research into health services utilization and outcomes in its federation of private, not-for-profit health care systems (Ross et al., 2014). Those who contribute data to the VDW are insured patients who are representative of non-institutionalized U.S. residents. The VDW is virtual in that it is distributed across sites, rather than existing as a whole in any location. The five member sites were: Baylor Scott & White Health, the largest non-profit health care system in Texas (180,000 patients contributing data from the 25,000 square mile Central Texas branch); Health Partners, an integrated health care organization with over 1,800 physicians serving 1.2 million patients across Minnesota, Wisconsin, Iowa, Illinois, North and South Dakota; Kaiser Permanente Georgia, the largest integrated health system in Georgia, comprising 26 medical facilities with nearly 365,000 members; Henry Ford Health System, one of the nation’s largest group practices with nearly 30 medical centers in southeast Michigan serving over 100,000 patients annually; and Kaiser Permanente Hawaii, an integrated managed care system serving 252,000 members across 19 medical clinics on four islands (Hawaii, Maui, Oahu, and Kauai). Study procedures were approved or exempted from review by each health care system’s Institutional Review Board.

Patients aged 15-88 were included if they had a PTSD diagnosis in 2014, were insured (including commercially insured, Medicare and Medicaid) in that year with no more than a 3-month coverage gap, and had records of care including diagnoses. The independent variable of primary interest was location of care where patients first received a diagnosis of PTSD, categorized as mental health care (MHC) versus primary care or other non-mental health care (e.g., family medicine, endocrinology, general internal medicine, and gynecology); diagnoses by paraprofessionals, community health workers, and peer specialists were excluded. In 2014, behavioral health consultants (Reiter, Dobmeyer, & Hunter, 2017), who work to support the primary care provider as a generalist and typically provide a limited (typically less than 6) number of 15-30 minute treatment sessions, were not fully embedded into all primary care departments across these systems. Any PTSD visit in that year with an MHC provider put the patient in the mental health group.

Type of care was identified by the specialty of the attending clinician or by the outpatient clinic where care occurred. Psychotherapy visits were defined using standardized procedural terminology codes for diagnostic interviews/assessments and individual psychotherapy (HMSA Provider Resource Center, 2017) that were at least 30 minutes long. Comorbid conditions of interest, identified in the 365 days prior to the patient’s first PTSD diagnosis, included depression, anxiety, substance use disorders (SUD), psychotic disorders and bipolar disorder, as well as physical conditions commonly treated in primary care (hypertension, dyslipidemia). Psychotropic medications were identified by National Drug Codes and drug names in the categories of antidepressants, anxiolytics/sedative-hypnotics excluding benzodiazepines, benzodiazepines, and antipsychotics. Pharmacotherapy was defined as the number of 30-day prescriptions filled for psychotropic medications specified above.

Analytic Plan

Descriptive statistics included frequencies (percentages) and means (standard deviations; SD). Bivariate comparisons were conducted with chi-square for categorical measures and Student t-test, analysis of variance, or non-parametric equivalent (Wilcoxon rank-sum; Kruskall-Wallis) for interval-level data.

Patients with both mental health and primary care/other visits were included in the MHC group. Multivariate models assessed differences in number of psychotherapy sessions by type of care (mental health vs primary/other care) adjusting for patient clinical and demographic variables and for nesting of patients within site. A similar model was used to assess correlates of psychotropic medication fills. Covariates were chosen for their association with variation in the use of primary care or of psychotherapy: age, gender, hypertension, dyslipidemia, anxiety, depression, SUD, and bipolar disorder.

Negative binomial regression models handled the dependent count variables with variance exceeding the mean, producing regression coefficients for the log of the count of psychotherapy sessions or number of psychotropic medications, which are exponentiated to aid interpretation. The resulting incident rate ratios (IRR) and their 95% confidence intervals (CI) describe the increased or decreased rate of the outcome occurring for the group indicated by the predictor or per unit increase in a continuous predictor. The multi-level regression model used health care site as a random effect because the intraclass correlation coefficient was non-trivial. That is, patients within sites were more similar to each other than to patients in other sites. The random effect of site adjusted for this loss of independence among observations.

Results

There were 5,256 patients who received a diagnosis of PTSD in 2014 out of a total of 2.1 million patients, suggesting a prevalence of 251 cases per 100,000 persons (Table 1). Of these, 83.4% (4,383) were diagnosed by a mental health provider, while one in six (16.6%) was diagnosed by a primary care or other type of provider. The overall mean age of patients receiving a PTSD diagnosis was 40.3 years (SD = 14.6; range 15-88) including 7% under the age of 18, 76.2% female, and 45.1% non-White. In the year prior to PTSD diagnosis, most patients (59.1%) also had a diagnosis of depression, while 45.5% were also diagnosed with anxiety, 35.5% had both anxiety and depression, 18.3 % had SUD, 11.4% had bipolar disorder, and 4.5% had a psychotic disorder. Hypertension and dyslipidemia were diagnosed in 21.5% and 20.1% of patients, respectively.

Table 1.

Population Demographics and Clinical Characteristics by Location of Initial Diagnosis of PTSD: Mental Health or Primary Care

Characteristic	Mental Health Care (N = 4,383)		Primary or other non-MH Care (N = 873)		Total (N = 5,256)
	M	SD	M	SD	M	SD	p- value
Age in years (range 15-89)	40.3	14.6	48.6	16	41.6	15.2	<.0001
	n	%	n	%	n	%
Age Group
15-29	1,126	25.7	119	13.6	1,245	23.7	<.0001
30-64	3,040	69.4	593	67.9	3,633	69.1	0.40
65-90	217	4.9	161	18.4	378	7.2	<.0001
Gender
Female	3,339	76.2	501	57.4	3,840	73.1	<.0001
Male	1,044	23.8	372	42.6	1,416	26.9	<.0001
Race / ethnicity
White	2,423	55.3	461	52.8	2,884	54.9	0.20
Black	795	18.1	222	25.4	1,017	19.4	<.0001
Hispanic	176	4.0	27	3.1	203	3.9	0.20
PTSD*	1,933	44.1	273	31.3	2,206	42.0	<.0001
MDD	2,786	63.6	321	36.8	3,107	59.1	<.0001
Anxiety	2,124	48.5	268	30.7	2,392	45.5	<.0001
SUD	864	19.7	97	11.1	961	18.3	<.0001
Bipolar Disorder	551	12.6	48	5.5	599	11.4	<.0001
Psychosis	215	4.9	19	2.2	234	4.4	0.0004
Hypertension	841	19.2	289	33.1	1,130	21.5	<.0001
Dyslipidemia	806	18.4	248	28.4	1,054	20.1	<.0001

Note:

^*2,206 patients of 5,256 had a diagnosis of PTSD in 2013 (index date was the first diagnosis of PTSD in 2014).

Patients averaged 8.7 psychotherapy visits in the one year following their first PTSD diagnosis in 2014 (SD = 14.0; range 0-195); 44.6% (n = 1,647) had 9 or more sessions; 29.8% (n = 1,567) had no psychotherapy. Average psychotherapy visits were higher among those with a prior-year PTSD diagnosis in comparison to those with no prior-year PTSD diagnosis (7.1 [SD = 11.6] vs. 10.8 [SD = 16.6]). Those diagnosed by mental health providers averaged significantly more sessions (10.0 vs 2.0; p<.0001).

In terms of treatment, receipt of any psychotherapy was three-fold among those seeing mental health providers (the MHC group) relative to those seeing other providers (unadjusted: 79.3% vs 24.4%; p<.0001). Receipt of antidepressants, antipsychotics, benzodiazepines, and anxiolytics/sedative-hypnotics other than benzodiazepines varied by provider type, with generally higher prescription fill levels for patients in MHC group (see Table 2).

Table 2.

Treatment Profiles for Patients with PTSD in Integrated Health Care Systems

Parameter	Mental Health Care		Primary/Other Care		Total		Chi- square
Type of Treatment	n	%	n	%	N	%	p-value
All Psychotherapy sessions	3,476	79.3	213	24.4	3,689	70.2	<.0001
All Psychotherapy Sessions with PTSD Diagnosis	2,946	67.2	71	8.1	3,017	57.4	<.0001
Psychotherapy Sessions in Mental Health Clinics	3,278	74.8	119	13.6	3,397	64.6	<.0001
Psychotherapy Sessions in Mental Health Clinics with PTSD Diagnosis	2,797	63.8	0	0	2,797	53.2	<.0001
Psychotherapy Sessions from Non-mental Health Providers1	770	17.6	119	13.6	889	16.9	0.005
Psychotherapy Sessions from Non-mental Health Providers with PTSD Diagnosis	441	10.1	71	8.1	512	9.7	0.080
SSRI/SNRI	2,435	55.6	310	35.5	2,745	52.2	<.0001
Other Antidepressant	1,632	37.2	191	21.9	1,823	34.7	<.0001
Sedatives Excluding Benzodiazepines	865	19.7	102	11.7	967	18.4	<.0001
Benzodiazepines	993	22.7	122	14.0	1,115	21.2	<.0001
Prazosin	342	7.8	26	3.0	368	7.0	<.0001
Antipsychotics
Typical	35	0.8	3	0.3	38	0.7	0.10
Atypical	957	21.8	71	8.1	1,028	19.6	<.0001

Note.

¹Non-mental health providers included licensed/credentialed providers in the health system, only. These do not include community health workers or peer specialists.

In the regression models that adjusted for site, demographics and comorbidity (Table 3, 4), patients in MHC were estimated to have four times the rate of psychotherapy sessions (IRR=4.25; 95% CI 3.75 – 4.81; p<.0001). Being female was associated with a 36% increase in rate of receiving psychotherapy sessions (IRR=1.36; 95% CI 1.24 – 1.49; p<.0001). Patients in MHC also had nearly one and a half times the rate of pharmacotherapy use (IRR=1.48; 95% CI 1.31-1.67; p<.0001) compared to the PC group. Among the comorbidities, depression, anxiety, and bipolar disorders were all associated with about a 30% increase in incident rate of treatment sessions or psychotropic medications.

Table 3.

Negative Binomial Regression using Multilevel Modeling on Number of Psychotherapy Sessions Attended for Patients Diagnosed with PTSD in Mental Health or Primary/Other Care Clinics (Patients nested within site as random effect)

Variable	Incident Rate Ratio	95% CI	p value
Mental Health Care	4.25	3.75 – 4.81	<.0001
Age	0.99	0.96 – 1.02	0.60
Female	1.36	1.24 – 1.49	<.0001
Dyslipidemia	1.04	0.93 – 1.17	0.50
Hypertension	1.00	0.89 – 1.12	0.90
Anxiety	1.31	1.20 – 1.42	<.0001
Depression	1.31	1.20 – 1.43	<.0001
Substance Use Disorder	0.98	0.88 – 1.09	0.70
Bipolar disorder	1.31	1.16 – 1.49	<.0001

Note: Reference group is patients diagnosed with PTSD in primary/other care. The intraclass correlation coefficient for the effect of site was ICC = 0.31.

Table 4.

Negative Binomial Regression using Multilevel Modeling on Medication Use for Patients Diagnosed with PTSD in Mental Health or Primary/Other Care Clinics (Patients nested within site as random effect)

Variable	Incident Rate Ratio	95% CI	p value
Mental Health Care	1.48	1.31 – 1.67	<.0001
Age	1.17	1.13 – 1.21	<.0001
Female	1.16	1.05 – 1.28	.0002
Dyslipidemia	1.09	0.97 – 1.22	0.20
Hypertension	1.01	0.9 – 1.14	0.90
Anxiety	1.51	1.38 – 1.65	<.0001
Depression	1.42	1.29 – 1.56	<.0001
Substance Use Disorder	1.20	1.08 – 1.35	0.002
Bipolar disorder	1.51	1.32 – 1.73	<.0001

Note: Reference group is patients diagnosed with PTSD in primary/other care. The intraclass correlation coefficient for the effect of site was ICC = 0.30.

Discussion

This retrospective cohort study provides an examination of patients diagnosed with PTSD and their follow-up across five large civilian health care systems in the U.S. The majority of patients with PTSD received their diagnosis from a mental health provider, while one in six was diagnosed by a health care professional in primary/other non-MHC. Course of treatment varied, as patients diagnosed by MHC providers had four times the rate of psychotherapy sessions and were nearly one and a half times more likely to receive psychotropic medications than non-MHC-diagnosed patients. It is possible that patients with PTSD in primary care may be choosing to remain solely in primary care because they want medication rather than psychotherapy. However, even receipt of appropriate pharmacotherapy was less likely in non-MHC clinics.

In the current study, among patients diagnosed by MHC providers, the frequency and intensity at which psychotherapy sessions occurred also appeared to be fairly low, although almost half had a diagnosis of PTSD in 2013, so a portion may have previously received adequate treatment. Adequate treatment for PTSD typically entails nine or more treatment sessions, ideally spaced at weekly intervals [e.g., (Foa et al., 2007; Resick et al., 2016)]. The current findings suggest that, among those patients who received psychotherapy, the majority likely did not receive sufficient PTSD treatment within these health care systems. On the other hand, compared to a sample of 200 veterans referred for PTSD psychotherapy in the VA, civilians in these health care systems are receiving, on average, more psychotherapy sessions (8.7 in HCSRN versus 7.0 sessions in VA), although a smaller proportion (29.8% HCSRN vs 37.5% VA) had zero sessions (DeViva, 2014).

Findings related to psychotropic medication use are somewhat congruent with previous research. For example, in a sample of 10,636 privately insured patients with a diagnosis of PTSD, 60% were prescribed psychotropic medications (Harpaz-Rotem, Rosenheck, Mohamed, & Desai, 2008). In that study, women were more than one and a half times more likely to receive prescriptions and, similar to our study, those with comorbid disorders were also more likely to receive pharmacological intervention with the largest effects observed by those with a comorbid diagnosis of bipolar disorder. In the current investigation, of those receiving psychopharmacologic intervention, 55.2% were receiving an antidepressant (selective serotonin reuptake inhibitors (SSRIs) or selective norepinephrine reuptake inhibitor), 18.4% received anxiolytics or sedative-hypnotics (excluding benzodiazepines), 21.2% received benzodiazepines 20.3% received antipsychotics. It is important to note that while some SSRIs are considered first-line treatments for PTSD, prescribing rates of benzodiazepines appeared high in this sample despite their contraindication for use (Department of Veterans Affairs & Department of Defense, 2017).

In addition to the question of quantity, relatively little is known about the quality of PTSD treatments being offered or received in civilian health care settings. Namely, it is unclear what evidence-based psychotherapies (EBPs) MHC providers are trained in, what patients are offered, and what gets delivered. The VDW administrative database does not capture the type of psychotherapy received. Future studies should examine this issue as providers may face systemic barriers to implementing EBPs for PTSD, particularly as they relate to time, training, and overall caseloads (Cook et al., 2014).

There are several EBPs that have demonstrated efficacy for PTSD. Multiple practice guidelines have been developed on PTSD care delivery to help improve the quality of care for these patients, notably those adopted by the Departments of Defense and Veterans Affairs. More research needs to be conducted within our non-federal health care systems to determine how they might improve their service delivery to align with such guidelines. It is worth emphasizing that we found that benzodiazepines were equally likely to be prescribed by mental health providers as by other providers. These findings suggest the need for additional efforts to foster adoption of evidence-based PTSD care guidelines across non-federal health care settings.

Adoption of PTSD guidelines are important for important for several reasons. The prevalence of PTSD in primary care settings ranges widely from 2% to 39% (Greene et al., 2016), with higher rates in high-risk populations, such as veterans or urban, low socioeconomic individuals with high exposure to interpersonal violence. A recent review of 41 primary care studies reported that the median point prevalence of current PTSD was 12.5% (Spottswood et al., 2017). The current study observed low rates of PTSD diagnosing (251 cases per 100,000 persons or 0.25%) in these health care systems. Although we are unable to determine the reason for the low prevalence rates observed, we suspect this may be a result of underdiagnosing or patient populations whom are generally at low risk. We are aware that systematic screening of trauma and PTSD is not being conducted across these health care systems, whereas it is mandated for primary care settings within VA and the Department of Defense (Department of Veterans Affairs & Department of Defense, 2017).

Primary care settings remain the de facto MHC system in the U.S. (Regier et al., 1978). Indeed, primary care providers serve as important gatekeepers. In a national study in Israel of 2,975 patients from 26 primary care clinics, 39% met criteria for PTSD. Of those, only 2% were given a diagnosis of PTSD by physicians (Taubman-Ben-Ari et al., 2001), setting the scene for a missed treatment opportunity. Individuals with PTSD often have worse physical health, greater co-occurring mental health disorders, poorer functional status, and higher rates of medical service utilization compared to their non-PTSD counterparts, potentially leading them to primary and specialty providers (Greene, Neria, & Gross, 2016; Kartha et al., 2008; Leibschutz et al., 2007; Spottswood, Davydow, & Huang, 2017). These providers should be prepared to recognize and treat individuals with PTSD seeking care for physical disorders, but also in need of mental health care.

Recognizing that trauma and its sequelae are public health problems, multiple agencies have recommended specialized training for health care professionals (e.g., Institute of Medicine, 2014; SAMHSA, 2014). Indeed, improved provider identification is a crucial step in correcting the significant under- or untreated proportion of primary care patients with PTSD. In a qualitative investigation of barriers to care for trauma-related mental health problems among low-income minorities in primary care, providers reported that patient resistance, inadequate referral options, reimbursement issues, limited staff training, and lack of clarity about the term “trauma” were also deterrents (Chung et al., 2012). In addition, many reported feeling unprepared to deal with their patients’ trauma-related issues (Green et al., 2011). They expressed a strong desire for more training in trauma care and a hotline for providers who needed an immediate consultation.

However, Sonis (2013) explained that there is insufficient evidence at this time to recommend universal screening for PTSD. Rather, it may be more prudent for primary care providers to be alert to the possibility of PTSD for those with trauma histories, and consider initiating evaluation for PTSD among persons who are at higher risk of PTSD. Screening for trauma exposure may be an initial step in identifying those who are at risk for PTSD. To our knowledge, systematic screening for trauma and other adverse experiences is still not occurring in many health care systems. This limitation is relevant not only to the current study, but also has implications for practice. Understanding the nature, frequency, and duration of trauma exposure and its relationship to physical and psychiatric difficulties may help guide treatment decisions and shared decision-making between the provider and patient.

There are several limitations of this study. First, this data is cross-sectional, and thus causality between PTSD diagnosis location and health care utilization cannot be fully determined. Second, in addition, some patients had multiple mental health disorders, thus the mental health treatment they were receiving may have been for one particular disorder or their combination. The type of mental health professional seen and the type of psychotherapy received were not available in these large health care databases. In addition, it is unknown whether patients were offered, yet refused, mental health treatment. Prior-year treatment measures were also not available. Finally, it is not known if any sought mental health treatment outside of the study facilities.

Nonetheless, these findings from five large, civilian health care systems across the U.S. represent a strong picture of what PTSD patients are receiving in regards to psychotherapy and psychotropic treatments. As such, there are a number of implications. There appears to be a continued need for education and skills-based training of general health care providers, integration with their mental health counterparts, and a strong referral system to serve trauma-exposed patients (e.g. Possemato, Johnson, Wray, Webster, & Stecker, 2018). Future research should determine the best screening (including trauma exposure and multi-disorder screeners), diagnosis and treatment of PTSD approaches for adolescents and adults in primary care. It is important to explore how to improve rates of mental health treatment initiation for patients with PTSD. For example, more information is needed on what factors might prompt a patient who was offered psychotherapy to attend sessions. In addition, research is needed to determine the barriers and facilitators to delivery of EBPs for PTSD in PC as well as MHC care. Though there are likely times when treatment is best delivered by a mental health provider in specialty mental health settings, effective options need to be available for PTSD patients who may be unable or unwilling to engage in specialty MHC.