Pilot of a screening, brief intervention and referral to treatment process for symptoms of trauma among primary care patients

Abstract

Background

For populations with high rates of trauma exposure yet low behavioural health service use, identifying and addressing trauma in the primary care setting could improve health outcomes, reduce disability and increase the efficiency of health system resources.

Objective

To assess the acceptability and feasibility of a screening, brief intervention and referral to treatment (SBIRT) process for trauma and symptoms of posttraumatic stress disorder (PTSD) among American Indian and Alaska Native people. We also examine the short-term effects on service utilization and the screening accuracy of the Primary Care Posttraumatic Stress Disorder Screen.

Methods

Cross-sectional pilot in two tribal primary care settings. Surveys and interviews measured acceptability among patients and providers. Health service utilization was used to examine impact. Structured clinical interview and a functional disability measure were used to assess screening accuracy.

Results

Over 90% of patient participants (N = 99) reported the screening time was acceptable, the questions were easily understood, the right staff were involved and the process satisfactory. Ninety-nine percent would recommend the process. Participants screening positive had higher behavioural health utilization in the 3 months after the process than those screening negative. The Primary Care Posttraumatic Stress Disorder Screen was 100% sensitive to detect current PTSD with 51% specificity. Providers and administrators reported satisfaction with the process.

Conclusions

The SBIRT process shows promise for identifying and addressing trauma in primary care settings. Future research should explore site specific factors, cost analyses and utility compared to other behavioural health screenings.

Key Messages.

• Trauma screening, brief intervention and treatment referral were acceptable.

• Despite some increased feelings of stress, 99% would recommend the process.

• Patients who screened positive had higher behavioural health use afterward.

• T-SBIRT has the potential to address psychosomatic symptoms in primary care.

Introduction

Exposure to a life-threatening event, injury or violence can lead to distressing mental health symptoms and conditions including posttraumatic stress disorder (PTSD). PTSD is prevalent in primary care settings in United States (1). Compared to people without PTSD, individuals with PTSD experience higher rates of major depression, generalized anxiety disorder, tobacco and alcohol misuse, and chronic health conditions (2,3). People with PTSD are more likely to have high health service utilization (2,3), difficulty completing daily activities (4), reduced quality of life and elevated suicide risk (5). Importantly, people that do not meet the full diagnostic criteria but experience subclinical symptoms of re-experiencing, avoidance and hyperarousal, have similar levels of distress and dysfunction (6,7).

Risk of PTSD is associated with sociodemographic factors, psychiatric history and exposure to traumatic events (8–10). Globally, PTSD rates are higher among women, racial/ethnic minorities, those with low income, those lacking social support and those with severe or multiple trauma exposures (8). In the United States, Latino/a American, African American, and American Indian and Alaska Native (AI/AN) heritage have been associated with an elevated lifetime prevalence of PTSD compared to European American and Asian American heritage (9–11). AI/AN people have consistently been identified as having the highest prevalence with past year (12.9%) and lifetime (16.2%) rates of PTSD between two and three times European American counterparts (4.8%, past year; 6.3%, lifetime) (3). AI/AN people experience high rates of interpersonal violence (12–16), unintentional injury (17,18), intergenerational trauma (19–21) and multiple traumas, compared to their non-AI/AN counterparts (11,15,22,23).

Early detection of PTSD symptoms accelerates the development of therapeutic relationships, strengthens patient engagement and improves patient outcomes (24,25). Thus, it is critical for health systems to identify people with PTSD and address PTSD and related symptomatology. Screening, brief intervention and referral to treatment (SBIRT) processes have been successfully used in primary care to identify and intervene with alcohol or drug misuse and depression. Some healthcare systems are adapting SBIRT processes for PTSD in primary care settings serving Latino/a Americans and African Americans (26). An effective trauma-focused SBIRT program has the potential to improve health outcomes, reduce disability and allow more appropriate use of valuable health system resources (27). Self-report instruments can assist providers to identify people with PTSD symptoms for referral for further assessment and potential intervention (28,29).

While several PTSD screening approaches for primary care have been described (30–33), use of a comprehensive SBIRT process including universal screening, early intervention and timely referral and treatment is uncommon (34,35). We were unable to identify any published work on PTSD-focused services for AI/AN people in primary care despite the high trauma exposure and PTSD prevalence. In a previous publication (36), we describe the development of a trauma-focused screening, brief intervention and referral to treatment (T-SBIRT) process.

Objective

In this paper, we report on the acceptability and feasibility of a T-SBIRT process developed and piloted in two large AI/AN primary care systems (36). Our primary objectives were: (i) to assess the overall acceptability and feasibility of the T-SBIRT as measured by survey scores and responses to patient and provider or administrator follow-up interviews; and (ii) to examine the effects of the T-SBIRT on service utilization. Secondary objectives were to assess the accuracy of the Primary Care-PTSD Screen (PC-PTSD), used to screen for trauma-related symptoms, to detect PTSD and correlate with functional disability.

Methods

Settings and participants

Recruitment occurred in adult primary care clinics at the Cherokee Nation Health Services (CNHS) in Tahlequah, OK and Southcentral Foundation (SCF) in Anchorage, AK. Each site provides primary care and behavioural health services in primary care clinics. CNHS has doctoral level and master’s level clinicians embedded in their primary care clinic. SCF has master’s level Behavioral Health Consultants (BHCs) embedded in six primary care clinics. The T-SBIRT was piloted in one of the six clinics. The Centers for American Indian and Alaska Native Health at the University of Colorado—Anschutz Medical Campus provided coordination. Before data collection, study procedures received Tribal and Institutional Review Board approval. Study reporting conforms to Strengthening the reporting of observational studies in epidemiology (STROBE) statement (Supplementary Table S1).

Eligibility criteria included AI/AN heritage, eligibility for healthcare at CNHS or SCF, age 18 years or older and a scheduled appointment with a participating provider team. Exclusion criteria included inability to give consent or presence of minor. Participating providers and administrators signed an informed consent prior to engaging in the pilot.

Sample size

Through engagement with clinical leadership at CNHS and SCF, it was determined that 50 participants per site were needed to assess acceptability and feasibility of the T-SBIRT. Power calculations for comparison of the positive PC-PTSD result and interview diagnoses assumed 30% of participants would be diagnosed with PTSD or Acute Stress Disorder in the interview, with 90% of those scoring positive with the screening tool. Under these conditions with a 5% significance level, 41 participants were needed to achieve 80% power to detect satisfactory agreeance.

Measures

Posttraumatic Stress Disorder Screen

The PC-PTSD consists of four questions assessing lifetime PTSD symptoms following traumatic events (37). In the first question, ‘nightmares’ was replaced with ‘dreams’ as a more culturally appropriate term.

Net-Structured Clinical Interview for DSM-IV Disorders and World Health Organization Disability Assessment Schedule 2.0

To determine the performance of the PC-PTSD to accurately detect PTSD and functional disability among AI/AN outpatients, study psychologists assessed a subset of randomly selected participants using two instruments: the web-based version of the Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version, Non-Patient Edition (NetSCID) SCID-I/NP (38); and the 36-item World Health Organization Disability Assessment Schedule (WHODAS) 2.0 (39). The SCID, non-patient version, is a structured interview administered to diagnose disorders included in the DSM. Psychologists conducting interviews were blinded to the PC-PTSD screening result. The NetSCID was used to detect mood episodes, psychotic symptoms, substance use, anxiety, eating and adjustment disorders.

The WHODAS is a 36-item self-report assessment of functional disability across six domains: cognition, mobility, self-care, getting along [with others], life activities and participation. Item response scales range from zero (no difficulty) to four (extreme difficulty or cannot do). Item ratings are summed to generate domain and overall scores (39).

Demographic and acceptability survey

The survey collected participants’ demographic characteristics (AI/AN heritage, gender, income, education, household size and employment) and 14 true–false items assessing satisfaction with the screener and procedures, brief intervention and brochure, and the acceptability of the PC-PTSD and T-SBIRT components.

Telephone follow-up interview

A semi-structured interview consisting of seven open-ended questions about participants’ experiences with T-SBIRT was conducted at 2 weeks post-enrolment and completion of study procedures. Open-ended question constructs included acceptability of the PC-PTSD and T-SBIRT components. In addition, administrator and provider participants were asked for feedback regarding impact on clinic flow and their thoughts on implementation of T-SBIRT as a standard care process.

Electronic health record queries

A qualified staff member approved to access health records did so after the study follow-up period ended via manual extraction (CNHS) or constructed record queries (SCF). Service utilization data were extracted for number of visits to primary care, ER/urgent care, inpatient, other clinical/ambulatory and behavioural health clinics were collected for the periods 12 months prior to T-SBIRT completion and 3 months following. The screening date was excluded from each period.

Procedures

Participants completed data collection between October 2014 and January 2015.

Screening and detection

A participating nurse (CNHS) or BHC (SCF) administered the PC-PTSD screen with consented participants in a private clinic room. A ‘yes’ answer to three of four questions was considered a positive screen for PTSD based on standard cut-offs recommended by the screener developers (37).

Brief intervention

After screening, primary care providers (CNHS) or BHCs (SCF) discussed the screening results and conducted a brief intervention with each participant that included reviewing a brochure about trauma with general psychoeducational information and resources specific to each site (Supplementary Figures S1a, S1b, S2a, S2b) (36). All participants, regardless of screening status, received the brochure. Providers used scripted language about the brochure that varied based on participants’ screening status that explained why all participants received it, including those who screened negative.

Referral to treatment

At the end of the brief intervention, the primary care provider or BHC assessed the participant’s need for referral to behavioural health or other services and charted an order to initiate the referral when indicated and acceptable to the participant.

Diagnostic interview and disability assessment

Psychologists trained to conduct the NetSCID completed interviews and the WHODAS with randomly selected participants regardless of PC-PTSD screening outcome.

Telephone interview

Researchers telephoned ten patient participants 2 weeks after completing the T-SBIRT. All interviews were recorded.

Data analysis

Quantitative analysis

Analyses were conducted using SAS version 9.4. We explored differences in participant demographic characteristics, T-SBIRT acceptability ratings, rates of service use before and after the T-SBIRT, how the PC-PTSD screen predicted functional disability levels from the WHODAS, current and lifetime prevalence of PTSD, mood, anxiety and substance use disorders as measured by the NetSCID, and how the PC-PTSD Screen performed compared to the NetSCID in detecting PTSD. We utilized χ² or Fisher’s exact tests to compare participant demographic (data other than demographic characteristics and acceptability were not stratified by site to protect the confidentiality of individual study participants) characteristics by site and functional disability levels by screening outcome. We utilized Wilcoxon signed-rank test and Wilcoxon rank sum test to explore the effect of T-SBIRT on service utilization, comparing these data from the 12 months before the T-SBIRT (divided by 4 for comparison) and the 3 months after. Finally, diagnostic performance of the PC-PTSD was determined by calculating rates of false or true positive or negative screens as compared to NetSCID diagnosis rates. Missing data were handled via pairwise deletion for individual analyses.

Qualitative analysis

Two researchers (LM, BB) reviewed and analysed transcribed interviews broadly for common themes. Findings augmented and informed the quantitative acceptability survey data.

Results

Participant demographics

Two-hundred-sixty-three potential participants were asked if they had interest in participating in the study. Of those, 212 were assessed for eligibility. Those that declined to participate mostly cited an inability or unwillingness to commit to research procedures that, depending on group assignment, were described as needing 3 h to complete or were simply not interested. Of the 103 randomized to the study, 99 participants completed T-SBIRT (Figure 1). Their average age was 46 years, and 63% were women (Table 1). The two samples were similar in most regards except CNHS participants were more likely to be American Indian (P < 0.0001), have attained high school education or greater (P = 0.010), be employed full-time or part-time (P = 0.001), be uninsured (P = 0.001) and not be disabled or too ill to work (P = 0.010). SCF participants were more likely to be insured by Medicare or Medicaid (P = 0.033).

Figure 1.

Participant flow through research activities and trauma-SBIRT clinical procedures

Table 1.

Characteristics of 99 American Indian and Alaska Native participants in a pilot study to determine the acceptability and feasibility of a Trauma-SBIRT process, by study site (2014–2015)

Characteristic	Cherokee Nation Health Services N (%) or Mean ± SD	Southcentral Foundation N (%) or Mean ± SD	P
Participants, N (%)	50 (50.5)	49 (49.5)	0.999
Women, n (%)	30 (60.0)	32 (65.3)	0.493
Age, mean ± SD, years	47.4±10.2	44.5±14.6	0.259
Heritage, n (%)			<0.001
Alaska Native	0	44 (89.8)
American Indian	50 (100.0)	5 (10.2)
Education, n (%)
<High school	9 (18.0)	21 (42.9)	0.007
High school or GED	15 (30.0)	11 (22.5)	0.393
Trade school; some college	13 (26.0)	8 (16.3)	0.239
4-Year college or higher	13 (26.0)	9 (18.4)	0.361
Employment, n (%)
Employed, full or part-time	36 (72.0)	19 (38.9)	<0.001
Unemployed	3 (6.0)	14 (28.6)	0.003
Retired/Student/Homemaker/Subsistence	9 (18.0)	10 (20.4)	0.761
Disabled/too ill to work	3 (6.0)	12 (24.5)	0.010
Pre-tax income, n (%)
<$10 000	26 (52.0)	18 (36.7)	0.126
$10 000 to $29.9,000	7 (14.0)	11 (22.5)	0.276
$30 000 to $49.9,000	9 (18.0)	6 (12.2)	0.425
>$50 000	8 (16.0)	12 (24.5)	0.675
Insurance status, n (%)
Private Insurance	19 (38.0)	26 (53.1)	0.132
Medicare/Medicaid	5 (10.0)	13 (26.5)	0.033
No Third-Party Insurance	26 (52.0)	10 (20.4)	0.001
Number of people in household, n (%)
1	9 (10.0)	11 (22.9)	0.582
2–4	32 (64.0)	25 (52.1)	0.191
5 or more	9 (18.0)	12 (25.0)	0.430

Trauma SBIRT acceptability/feasibility

The median time to complete the SBIRT encounter was 5.0 min (interquartile range [IQR] = 7.0). Participants were highly satisfied with the entire process; almost all reported that the screening time was acceptable (100%), the questions were easy to understand (100%) and complete (100%) and were asked by the right staff person (99%) in a safe place (100%). All participants reported that the brochure was easy to understand and helpful. Of the 99 participants, 20 indicated that the questions increased their stress, and 6 were offended by the brochure. Finally, 99% of participants reported they would recommend the process to others. Post T-SBIRT interviews with 20 participants, 2 administrators and 6 providers in the intervention clinics were overwhelming positive, with high acceptability of PC-PTSD screening questions and low time burden to administer the screener.

Referral and use of clinical services

The total number of visits to all clinics combined and number of visits to either the behavioural health clinic or to see a Behavioral Health clinician (CNHS) or embedded consultant (SCF) differed significantly between the pre- and post-T-SBIRT periods. Among all participants, all clinical visits significantly decreased in the post-period compared to the pre-period, while visits to a behavioural health clinic or clinician increased significantly (Table 2). Those who screened positive visited behavioural health services of any kind significantly more frequently during the 3-month period after T-SBIRT versus the equivalent period 3 months prior to T-SBIRT compared to those who screened negative. This finding was echoed in the interview data in which several participants mentioned changing their behaviours based on the brief intervention.

Table 2.

Use of clinical services by 99 study participants at both study sites, before and after screening, overall and by primary care PTSD result (2014–2015)

Service type	All PC-PTSD screens (n = 99)				Positive PC-PTSD screen (n = 50)			Negative PC-PTSD screen (n = 48)
	Prea	Post	Δ b	All	Prea	Post	Δ b	Prea	Post	Δ b	PosΔ versus NegΔ
	M ± SD	M ± SD	M ± SD	P c	M ± SD	M ± SD	M ± SD	M ± SD	M ± SD	M ± SD	P d
Primary caree	1.6 ± 1.2	1.7 ± 2.1	0.1 ± 2.0	0.456	1.8 ± 1.5	2.2 ± 2.6	0.4 ± 2.4	1.3 ± 0.9	1.0 ± 1.3	−0.3 ± 1.4	0.177
ER/urgent caref	0.6 ± 1.0	0.6 ± 1.3	−0.1 ± 1.0	0.594	0.6 ± 0.8	0.7 ± 0.9	0.0 ± 0.9	0.5 ± 1.2	0.4 ± 1.6	−0.1 ± 1.1	0.267
Inpatient	0.0 ± 0.1	0.1 ± 0.4	0.1 ± 0.4	0.509	0.1 ± 0.1	0.1 ± 0.3	0.1 ± 0.3	0.0 ± 0.1	0.1 ± 0.5	0.1 ± 0.5	0.765
Other, ambulatory	1.9 ± 3.3	2.7 ± 4.4	0.8 ± 1.9	0.108	2.8 ± 4.2	3.9 ± 5.6	1.1 ± 4.4	1.0 ± 1.5	1.4 ± 2.1	0.4 ± 1.7	0.578
Any BH serviceg	0.3 ± 0.9	0.9 ± 2.2	0.6 ± 1.9	0.002	0.6 ± 1.2	1.6 ± 2.9	1.1 ± 2.5	0.1 ± 0.2	0.2 ± 0.9	0.1 ± 0.8	0.012
Any service	2.9 ± 0.9	2.0 ± 1.3	−1.0 ± 1.3	<0.001	3.3 ± 0.9	2.5 ± 1.3	−0.8 ± 1.4	2.5 ± 0.7	1.4 ± 1.0	−1.2 ± 1.3	0.111

^a12-Month pre-screening period counts were divided by 4 for comparison to 3-month post-screening period.

^b# of post-screening visits minus # of pre-screening visits.

^cWilcoxon signed-rank test for non-parametric paired data.

^dWilcoxon rank sum test for non-parametric unpaired data.

^eExcludes behavioural health consultant visits in primary care.

^fExcludes behavioural health services in emergency room/urgent care.

^gIncludes behavioural health consultant visits in primary care.

Primary care screener functional disability prediction

The screener differentiated between those who screened positive and negative in WHODAS measures of overall functional disability, and functional disability in domains of participation in society (P =0.002), self-care (P =0.03), and understanding and communicating (P =0.02) in the past 30 days. Greater disability in these domains was reported by participants who screened positive for symptoms of PTSD (P =0.04). The screener score did not differentiate between other disability domains of getting around, getting along with people, or life activities; nor did it predict days unable to work in the past 30 days (Table 3).

Table 3.

World Health Organization’s Disability Assessment Schedule 2.0 scores for a subset of 42 American Indian and Alaska Native participants at both sites, by primary care PTSD result (2014–2015)

Functional domain	Positive PC-PTSD screen (n = 23)a	Negative PC-PTSD screen (n = 18)a	P
	Mean ± SD	Mean ± SD
Overall score	33.9 ± 25.6	17.9 ± 20.0	0.043
Understanding and communicating	5.9 ± 4.4	3.0 ± 3.6	0.021
Getting around	4.9 ± 5.1	3.1 ± 3.3	0.257
Self-care	1.5 ± 2.4	0.5 ± 1.5	0.029
Getting along with people	4.4 ± 5.2	2.2 ± 3.2	0.141
Life activities	6.8 ± 7.1	5.9 ± 6.2	0.785
Participation in society	9.8 ± 6.9	3.3 ± 4.4	0.002
In the past 30 days, for how many days were you totally unable to carry out your usual activities or work because of any health condition?	3.3 ± 4.9	2.4 ± 7.1	0.188

^aFrequencies are lower than total sample because only participants completing the Structured Clinical Interview completed the WHO Disability Assessment Schedule 2.0.

Primary care screener diagnostic characteristics

Structured clinical interview diagnoses.

Forty-two participants completed the Structured Clinical Interview. Of the 16 people who experienced a DSM-IV traumatic event, four met criteria for PTSD at the time of the interview, eight in their lifetime (Table 4).

Table 4.

Prevalence of DSM-IV psychological and behavioural health disorders in 42 American Indian and Alaska Native participants at both sites, as determined by the NetSCID (2014–2015)

Diagnosis	N (%)
PTSD, lifetime	8 (19.0)
PTSD, current	4 (9.5)
Any anxiety disorder, lifetimea	17 (40.5)
Any anxiety disorder, current	11 (26.2)
Any mood disorder, lifetimeb	13 (30.9)
Any mood disorder, current	5 (11.9)
Substance use disorder, lifetimec	12 (28.6)
Substance use disorder, current	4 (9.5)
Any behavioural health condition, lifetime
0	15 (35.7)
1	13 (30.9)
2 or more	14 (33.3)
Any behavioural health condition, current
0	25 (59.5)
1	10(23.8)
2 or more	5 (11.9)

^aAny anxiety disorder includes: anxiety disorder not otherwise specified, panic disorder without agoraphobia, generalized anxiety disorder, panic disorder with agoraphobia, social phobia, specific phobia, posttraumatic stress disorder.

^bAny mood disorder includes: mood disorder due to general medical disorder, major depressive episode, bipolar I, bipolar II, dysthymic disorder.

^cSubstance use disorder includes any alcohol and/or drug abuse or dependence.

PC-PTSD diagnostic accuracy

Of the 99 participants, 50 screened positive on the PC-PTSD, 48 screened negative and 1 screening was missed. Of the 42 screened participants who also completed a Structured Clinical Interview, 23 screened positive. Thus, for lifetime PTSD, there were 8 true positives, 15 false positives, 0 false negatives and 18 true negatives. For current PTSD, there were 4 true positives, 19 false positives, 0 false negatives and 18 true negatives. The PC-PTSD screen was 100% sensitive and 51% specific to detect current PTSD with an accuracy of 54%, a positive predictive value of 17% and a negative predictive value of 100%.

Conclusions

Summary of results

During development of the T-SBIRT process, described elsewhere, administrators and providers voiced concerns that T-SBIRT would take too much time, complicate visits, trigger adverse patient reactions and overload the health systems by identifying too many patients in need of behavioural health intervention (36). Our findings were contrary to most of these initial concerns. The T-SBIRT process was not overly time-consuming or complicating within primary care visits. Provider and administrative interview participants reported no adverse reactions to screening questions and no adverse impact on clinic flow at either site in which behavioural health services are routinely provided in primary care clinics. Nearly all patient participants were highly satisfied with T-SBIRT and would recommend it to others.

The PC-PTSD screen’s performance raises opportunities and concerns about its use in routine clinical operations. Overall, the T-SBIRT process demonstrated good clinical utility as it was 100% sensitive to identify PTSD, differentiated functional disability levels at time of screening and increased rate of behavioural health service use in the following 3 months for those screening positive. Although the screener had a false positive rate of 45% and 51% for lifetime and past year PTSD, respectively, it appears these false positives may consist of people with subclinical PTSD, and with similar levels of distress and dysfunction. Should these rates persist in daily operations, the concern of overwhelming an unprepared behavioural health system is not unfounded.

Considerations for implementation

How an organization chooses to use T-SBIRT depends on its institutional capacity. Organizations with adequate behavioural health resources may choose to fully implement T-SBIRT universally to address PTSD, subclinical PTSD, or other behavioural health needs that may be causing functional impairment. Such implementation has the promise to reduce overall costs to the health system. In other instances, a screening process in primary care should not be implemented unless resources are available (40). Organizations with limited resources could raise the PC-PTSD cut off score to 4 (out of 4) to reduce false-positives, potentially conserving resources. The T-SBIRT may be implemented alone or in conjunction with other screening efforts. For instance, a similar T-SBIRT has also demonstrated acceptability and feasibility in settings serving Latino/a American and African American populations when administered in conjunction with SBIRT screening for substance use (26). Organizations may be advised to examine gaps in existing screening practices and inventory existing service availability prior to implementation (40). Implementation of T-SBIRT, at any scale, represents a meaningful new avenue of patient engagement, an opportunity for providers to get to know patients, and the potential to more appropriately address psychosomatic symptoms in primary care. Consideration of implementation may be merited in healthcare organizations serving populations with high prevalence of PTSD, rates of trauma exposure, or in which trauma-informed services are an identified priority.

Limitations

The small sample precluded exploration of changes in patient diagnoses use after participation in T-SBIRT. The limited 3-month follow-up limited our ability to capture changes in service use. The diagnostic skip pattern of the NetSCID prohibited assessment of subclinical PTSD symptomatology and analysis of the T-SBIRT process based on types of trauma. Receiver operating characteristic curve analysis to identify an optimal screening cut-off score was precluded because documentation of screening outcome was an indicator variable rather than a score. Treatment referrals may have been affected by other variables, such as system capacity for effective T-SBIRT. Our randomized sampling strategy for patient follow-up interview questions did not allow targeted follow-up interviews to explore why six participants were offended by the brochure. Lastly, differences between samples at the two sites precluded meaningful cross-site comparisons.

Future research

The T-SBIRT process shows promise to improve the identification and treatment of PTSD in primary care settings serving AI/AN people. Follow-up studies are warranted with other populations with high prevalence of PTSD, rates of trauma exposure, or other risk factors (e.g. racial/ethnic minority or low-income populations). Follow-up studies should be at a larger scale and include receiver operating characteristic curve analysis to optimize a cutoff score. These studies should explore site specific and other factors impacting acceptability, feasibility, effectiveness, costs and cost-savings, and incremental benefit to identify a population in need of services not being identified by existing screening initiatives (e.g. depression and substance use).

Declaration

Funding: Funding received by the National Center on Minority Health and Health Disparities of the National Institutes of Health (grant number 2P60MD000507).

Conflict of interest: The authors have no conflicts of interests to declare.