Improving Function Through Primary Care Treatment of PTSD: The IMPACT Study Protocol

Abstract

Despite the availability of effective psychological interventions for PTSD, access to and retention in these interventions remains problematic. Of note, the Veterans Health Administration developed and implemented post-deployment health surveys that screen for PTSD in primary care (PC), but effective PC-based, psychological intervention treatment options have yet to be established. To address the literal physical gap between where the patients first present for care (i.e., primary care) and where they must go to receive first-line treatment for PTSD treatment (i.e., specialty mental health), study investigators developed a 4–6 visit Prolonged Exposure for Primary Care (PE-PC) treatment that has shown efficacy in reduction of PTSD. To extend previous work to recovery-based mental health care, the Improving Function Through Primary Care Treatment of PTSD (IMPACT) study examined function as assessed by the World Health Organization Disability Assessment Schedule [WHODAS 2.0; (Axelsson, Lindsäter, Ljótsson, Andersson, & Hedman-Lagerlöf, 2017)]. Veterans presenting in VA primary care mental health integration (PCMHI) clinics with PTSD or significant subsyndromal PTSD who met minimal inclusion and exclusion criteria were randomly assigned to PE-PC or treatment as usual (TAU). If proven effective in improving function, PE-PC would provide a new access point for high quality PTSD care and allow greater numbers of veterans to access effective PTSD treatment.

Introduction

Posttraumatic stress disorder (PTSD) is a debilitating and costly mental health condition with great impact on the Veterans Health Administration (VHA) (Medicine, 2014). The estimated two-year cost for mental health sequelae considering only the conflicts in Iraq and Afghanistan is $4.0 to $6.2 billion U.S. dollars, and providing evidence-based treatments for PTSD and depression could save an estimated $86.2 million (Watkins et al., 2011). Even modest reductions in PTSD severity are related to increased probability of positive functional outcomes (Mueller et al., 2019). Prolonged Exposure (PE) therapy is an effective, first-line treatment for PTSD (VA/DOD, 2017). However, despite successful PE dissemination efforts in VHA, most veterans do not access first-line treatments for PTSD offered through specialty mental health clinics (Maguen et al., 2020). Indeed, referral, per se, to specialty clinics for PTSD treatment contributes to low rates of treatment access (Possemato, Wray, Johnson, Webster, & Beehler, 2018). Of note, the VHA has developed and implemented post-deployment health surveys that screen for PTSD in primary care (PC) but has not established effective PC-based PTSD psychotherapeutic interventions (VA/DOD, 2017) to immediately respond to positive screens.

To address the gap between where the patients first present for care (i.e., primary care) and where they must go to receive first-line treatment for PTSD (i.e., specialty mental health), study investigators developed a 4–6 visit PE for Primary Care (PE-PC) treatment protocol. A previously completed pilot randomized clinical trial in military facilities supported efficacy of PE-PC for PTSD in that half of those with PTSD at baseline who received PE-PC no longer meeting PTSD diagnostic criteria six months posttreatment (Cigrang et al., 2017).

In addition to the compelling clinical rationale for PE-PC, the current study aimed to move beyond standard clinical symptom measures to focus our primary outcome on function. A recent review strongly recommended the use of functional outcomes assessment in PTSD treatment and clinical research, in addition to symptom measure in order to fully describe response to interventions (Wisco, Marx, & Keane, 2012). Function measures broaden the focus of assessment from PTSD specific symptoms to more global social, work, and family role performance, often with physical and mental health function separated. Global function measures and PTSD specific measures that are brief and show excellent psychometrics allows for closer study of these constructs (Bovin et al., 2018; B. P. Marx et al., 2015; Saltychev, Katajapuu, Bärlund, & Laimi, 2021). To extend the previous work to recovery-based mental health care, we examined the primary outcome of function in this study as assessed by the World Health Organization Disability Assessment Schedule [WHODAS 2.0; (Axelsson et al., 2017)].

While previous trial results with PE-PC in military active duty showed reductions in PTSD and depression, examination of whether similar reductions in PTSD and depression would be seen in a veteran sample is warranted. If PE for Primary Care (PE-PC) treatment is proven effective in improving function, this further supports conclusions that that PE-PC represents a new access point for effective, high quality PTSD care that can allow more veterans effective PTSD treatment, consistent with a primary goal of the National Research Action Plan (2013) translating clinical research findings into clinical veteran care. Toward this end, the Improving Function Through Primary Care Treatment of PTSD (IMPACT) study evaluated both functional and symptom outcomes associated with brief, primary care-based PE-PC in patients with PTSD and/or subsyndromal PTSD symptoms presenting for treatment in VA primary care mental health integration (PCMHI) clinics.

Methods

Study Design

We examined the effectiveness (improvement in function and symptoms) of PE-PC compared to treatment as usual (TAU) in VHA primary care mental health integration (PCMHI) clinics in a randomized clinical trial (See Figure 1 for study flow). We randomized 121 veterans who presented in VA PC with chronic PTSD symptoms [PTSD Symptom Checklist for DSM5 (PCL-5; Bovin et al., 2016) PCL-5 ≥ 28 for at least 3 months] and met minimal inclusion/exclusion criteria to receive PE-PC (four to six, 30-minute, weekly, in person or telehealth visits) or PCMHI treatment as usual (TAU). Inclusion criteria were: 1) Any era veteran with PCL5 ≥ 28 seeking care for PTSD in VA Primary Care, 2) Age 18–70, 3), 3) English speaking, 4) Significant functional impairment on WHODAS (≥3 in any of the six domains), and 5) Desires PTSD treatment. Exclusion criteria were: 1) Other primary clinical issue that would interfere with PTSD treatment, 2) Level of suicidal risk that requires intervention based on C-SSRS), 3) Severe cognitive impairment that makes it unlikely that the patient can adhere to the study regimen, 4) Psychosis or unmanaged bipolar disorder, 5) Moderate to severe substance use disorder in the past 8 weeks, 6) Currently receiving talk therapy for trauma- related symptoms. TAU included the current best-practices standard of PTSD care in PCMHI (medication, referral to specialty mental health and seen within two weeks, etc.) at the clinic sites. The type of TAU used reflected the current practice in PCMHI and was not constrained by study design. Inclusion/exclusion were minimized to ensure that the model of study entry reflected a feasible path for PCMHI providers to use following the end of the study. We monitored and recorded medical and mental health care received.

Figure 1.

Study Design

All veterans completed baseline (week 0) and follow-up assessments at weeks 6 (post), 12, and 24. Participants completed interview assessments via phone or face-to-face and self-reports via paper and pencil or online survey software. The total score of World Health Organization Disability Assessment Schedule [WHODAS 2.0 (Axelsson et al., 2017)] was the primary outcome. Secondary outcomes included the Brief Inventory of Psychosocial Function [B- IPF (Kleiman et al., 2020)], Clinician Administered PTSD Scale-5 [CAPS-5; (Weathers et al., 2018) and Patient Health Questionnaire-9 PHQ-9(Kroenke, Spitzer, & Williams, 2001)]. Based on design, if WHODAS showed a floor or ceiling effect, the overall impairment score of the B-IPF would become the primary outcome measure. No floor or ceiling effects were apparent and WHODAS remained the primary outcome.

We hypothesized that:

Hypothesis 1:

PE-PC will result in larger improvements in functioning [WHODAS 2.0 (Axelsson et al., 2017)] than PCMHI-TAU (Week 0 to Week 6), and these differences will remain at 12- and 24-weeks post-baseline.

Hypothesis 2.1:

PE-PC will result in larger reductions in PTSD severity [Clinician Administered PTSD Scale (CAPS-5; Weathers et al., 2018), PTSD Checklist-5 (PCL-5; Bovin et al., 2016)] than PCMHI-TAU (Week 0 to Week 6) and these differences will remain at 12- and 24-weeks post-baseline.

Hypothesis 2.2:

PE-PC will result in larger reductions in depression [Patient Health Questionnaire-9 (PHQ-9; ] than PCMHI-TAU (Week 0 to Week 6) and these differences will be maintained at 12- and 24-weeks post-baseline.

Hypothesis 2.3:

PE-PC will result in a greater proportion of veterans remitting from PTSD (Marx et al., 2021) than PCMHI-TAU at week 6, 12, and 24.

Our exploratory aim was to assess potential moderators of improvements in function and PTSD symptoms, including combat experience severity and exposure to prior traumatic events.

Eligibility Criteria

Veterans who presented in PCMHI at a Southeastern VA Medical Center (VAMC) and its community satellite clinics (CBOCs) with chronic PTSD symptoms (PCL-5 ≥ 28) of at least 3 months duration and who met minimal inclusion/exclusion criteria were randomly assigned to PE-PC (four to six, 30-minute weekly visits) or PCMHI-TAU. Potential participants were from any era with PTSD symptoms desiring treatment for this problem. Ranging in age from 18 to 70. Inclusion/Exclusion criteria were initially determined by the PCMHI provider at the first contact and confirmed by the independent evaluator (IE) at the baseline assessment following consent. All treatment and study visits were conducted via telehealth or face to face based on patient and provider preference and current requirements for health and safety (based on study conduct occurring during the COVID-19 pandemic).

Medication Stabilization.

Participants meeting inclusion/exclusion criteria were asked to maintain medications at current dosages when medically appropriate. Participants who had not started new psychotropic prescription medications in the previous 2 weeks completed the baseline assessment battery. Potential participants who had recently begun trials of psychotropic prescription medication were required to wait 2 weeks prior to completing the baseline assessment to ensure medication stabilization, at which point we administered/readministered the assessment battery.

Recruitment and Screening

Flyers and posters were displayed at the PC clinics and other high-traffic areas and directed those interested in participating to contact research staff for information or to inquire about the study with their PCP. PC providers were informed of the study and how to refer veterans to the PCMHI provider for treatment of PTSD in the PC setting. The study PCMHI provider was available for referrals from the PC providers as per the standard PCMHI procedures. Study staff screened for PTSD symptoms and significant impairment in function from these symptoms, as well as a desire for treatment to address these symptoms in primary care. Potential participants then had the study explained to them in a safe and private location with the opportunity to ask any questions. These visits occurred via televideo or face to face based on COVID-19 pandemic restrictions and patient preference once the restrictions were lifted. Research staff and the on-site PI delivered periodic briefings to PC providers to encourage assessment of PTSD symptoms and increase referrals to PCMHI. Research staff and investigators sought opportunities to brief larger groups such as health fairs, staff meetings, etc. to inform the facility of the study and how to refer veterans.

Randomization and Masking

After determining eligibility, we randomly assigned participants (1:1) to one of the two study conditions using a computer pre-generated block randomization schedule with random block sizes, stratified by site. We recorded the randomization results on the master tracking table. After randomization, patients remained in the intent-to-treat analytic cohort. The evaluators were not aware of assignment and reminded participants prior to each study assessment to not mention details of their treatment that might unblind the evaluator.

Interventions

Experimental Condition: Prolonged Exposure for Primary Care (PE-PC).

PE-PC was provided by the PCMHI or research provider in 4 – 6, 30-minute appointments scheduled approximately once a week over 4–6 weeks (to accommodate scheduling and missed appointments). All veterans received 6 visits unless they met the early response criteria (i.e., obtained a PCL-5 score of 20 or lower in visit 4 or later). PCMHI providers of PE-PC for this study included VA psychologists and licensed professional counselors who had completed the PE-PC training program. Treatment followed the PE-PC manual and veteran workbook developed in the pilot study, with minor modifications that occurred in our first 3 months of the study for use in VHA. Treatment content was drawn from the PE model (Edna B Foa, Rothbaum, Hembree, & Rauch, 2019) and includes smaller doses of all three components (imaginal exposure and processing, in vivo exposure, & psychoeducation) to fit the PCMHI brief visit format. At the first 30-minute appointment, the PE-PC provider reviewed the “Confronting Uncomfortable Memories” activity workbook to be completed at home and brought back for use in subsequent appointments. The workbook prompted the veteran to write a detailed first-person narrative of the event associated with the greatest level of current distress, including recollection of personal thoughts, feelings, and physical reactions. Emotional processing questions that focused on the meaning of the trauma at the time of the trauma and now followed the narrative in the workbook. In addition, the veteran and PE-PC provider planned for in-vivo exposure activities between appointments. The participant wrote a new memory exposure once per week (for 30 minutes) and then read that narrative for 30 minutes at least 3 times in the week between visits. At the second 30-minute appointment, the PE-PC provider reviewed the veteran’s exposures, problem-solved any implementation difficulties, and processed the exercises. Each visit started with the veteran reading the narrative and answers to the emotional processing questions out loud followed by about 10–15 minutes of processing. This appointment format and content was repeated in the subsequent appointments. At the end of treatment, the provider and veteran reviewed treatment progress assisted by results of the PCL-5 administered at baseline and during the last appointment. Early response was defined as a PCL-5 total score of 20 or lower at visit 4 or later. PE-PC treatment concluded at that visit if the veteran was an early responder. If PCL criteria were not met, PE-PC treatment continued for 6 visits. Potential outcomes of this collaborative review included concluding treatment at PE-PC completion or referring the veteran to specialty MH care for more intensive treatment.

Provider Training and Fidelity.

IMPACT followed the established training protocol for PE-PC developed by the authors (SR, JC). For this study all providers were required to have previous training at the level of PE provider status, a 4-hour PE-PC Training Webinar led by the PI (SR), and then 6 months of 30-minute weekly PE-PC case consultation calls. In addition, training cases were recorded and rated using the PE-PC fidelity rating forms to establish competence prior to providers seeing randomized study patients. All PE-PC providers were required to receive a minimal 85% competency on the PE-PC fidelity manual elements based on the PE-PC fidelity rating forms completed by the fidelity rater prior to seeing randomized cases since the brief intervention requires a high level of competence. A trained fidelity evaluator reviewed and rated a randomly selected 10% of visit recordings. Finally, all providers attended weekly ongoing supervision for PE-PC providers for the duration of the trial to further ensure provider competence and fidelity.

Control Condition: Treatment as Usual Condition.

Veterans assigned to PCMHI-TAU received standard PCMHI care for PTSD. This included services in PCMHI as well as possible referral for specialty care (including specialty MH), medication management, and/or general supportive contact while awaiting referral. All MH care documented in the medical record was recorded, including tracking encounters, current ICD-10 codes, duration of contacts, no shows, and cancelled appointments. Psychoactive medications were tracked, including medications prescribed, changed, and filled.

Data Collection.

Patients completed clinical interviews and self-report measures of function and symptoms at intake (Baseline/Week 0), week 6, week 12 and week 24 (see Table 1). We repeated week 0 measures if time since baseline exceeded 4 weeks prior to visit one and used these new data for baseline values. We also collected administrative data to examine utilization of VA care the 24 weeks prior to randomization and 24 weeks following treatment. To accommodate participant schedules, pandemic public health requirements, and/or instances in which a participant may have left the local area at the time of a follow-up assessment, full or partial assessments occurred in person or via phone or electronic data capture using a secure link to the HIPAA-compliant database.

Table 1.

Research Materials.

Instrument	Wk 0	Wk 6	Wk 12	Wk 24
1. LEC	X
2. CAPS-5	X	X	X	X
3. WHODAS	X	X	X	X
4. C-SSRS	X	X	X	X
5. Health Interview		X	X	X
6. Demo & Military Service Characteristics	X
7. B-IPF	X	X	X	X
8. PCL-5*	X	X	X	X
9. PHQ-9*	X	X	X	X
10. PTCI	X	X	X	X
11. CES	X
12. CEQ	X	X	X	X
13. MINI (AUD/SUD modules only)	X
14. COVID-19 Research Therapy Impact Questionnaire		X	X	X

^*These measures will be administered for PE-PC treatment visits as well as all assessment time points.

Evaluator Training.

All evaluators were blind to treatment condition and received training on WHODAS 2.0 and CAPS-5 prior to the start of the study. Interviews were recorded for recalibration and inter-rater reliability assessment. All evaluators completed recalibration assessment reviews twice per year for the WHODAS and monthly for the CAPS-5.

Measures

Primary Outcome: World Health Organization Disability Assessment Schedule [WHODAS 2.0 (Axelsson et al., 2017)]:

The WHODAS is a brief assessment instrument (36-item version) used to assess health and disability across six domains of functioning: cognition, mobility, self-care, getting along, life activities and participation. The WHODAS has excellent psychometric properties, including high overall internal consistency (α = .96) and test- retest reliability (0.98).

Brief Inventory of Psychosocial Function [B-IPF (Kleiman et al., 2020)):

The B-IPF is a seven-item brief self-report of function. We examined the measure’s utility within this patient population and its sensitivity to show treatment change coincident with the well-established WHODAS. The B-IPF has good psychometric properties, including high overall internal consistency (α = .84) and adequate test- retest reliability (r = 0.65).

Clinician Administered PTSD Scale-5 [CAPS-5 (Weathers et al., 2018)]:

The CAPS-5 is a interview measure of PTSD severity and the primary outcome measure for the study. Current PTSD was assessed by blinded evaluators in relation to the trauma that was currently most upsetting. The CAPS-5 has excellent psychometrics including high overall internal consistency (α = .88) and adequate test- retest reliability (r = 0.83).and required about 45 minutes to complete.

PTSD Checklist – Stressor-Specific Version [PCL-5(Bovin et al., 2016; Brian P. Marx et al., 2021)]:

The PCL-5 is a 20-item self-report measure of PTSD severity in the past month with 4 subscales: Intrusion symptoms (items 1– 5); Avoidance symptoms (items 6–7); negative cognitions and mood symptoms (items 8–14) and hyperarousal symptoms (items 15–20). Each item ranges from 0 (not at all) to 4 (extremely). The PCL-5 has good psychometric properties, including high overall internal consistency (α = .96) and adequate test- retest reliability (r = 0.84).

Patient Health Questionnaire – 9 [PHQ-9(Kroenke et al., 2001)]:

The PHQ-9 is a 9-item, well-validated measure of depression. The PHQ-9 assessed symptoms of major depression in the past two weeks from 0 (not at all) to 3 (nearly every day). The PHQ-9 detects treatment changes in depression in PC settings and has excellent internal and test-retest reliability and construct and criterion validity. The PHQ-9 has good psychometric properties, including high overall internal consistency (α = .89) and adequate test- retest reliability (r = 0.84).

Demographic & Military Service Characteristics:

This form assesses standard demographics (ethnicity, sex, age) and military service information (e.g., rank).

Columbia-Suicide Severity Rating Scale [C-SSRS(Madan et al., 2016)]:

The C-SSRS is a 5-item that assessed past suicidal ideation, behavior, and risk. The C-SSRS has good psychometric properties, including high overall internal consistency (α = .95) and the total score and summary score from the ideation/behavior factor balances sensitivity (0.694) and specificity (0.652–0.674).

Posttraumatic Cognitions Inventory [PTCI(Edna B. Foa, Ehlers, Clark, Tolin, & Orsillo, 1999)]:

The Posttraumatic Cognitions Inventory (PTCI) is a 36-item self-report of negative thoughts about the self, negative thoughts about the world, and self-blame. The scale has good psychometrics and changes in these thoughts relate to change in PTSD symptoms with treatment. The PTCI has good psychometric properties, including high overall internal consistency (α = .97) and adequate test- retest reliability (r = 0.85).

Health Interview:

The Health Interview was created for a previous PTSD clinical trial (Rauch et al, 2017) and included items regarding general health, hospitalizations, current and past psychiatric medications, utilization of MH services, utilization of outpatient medical services, and caffeine and tobacco use. We minimized items to prevent redundancy with information pulled from medical record and other sources. After treatment, we ask about changes in military status and important life events since the last interview.

Combat Experiences Scale [CES(Keane et al., 1989)]:

CES is a seven-item measure of combat exposure severity and enquired about the frequency of various combat experiences. Total scores range from 0 to 41. The CES has good reliability and will characterize the level of combat exposure in the sample.

Life Events Checklist [LEC(Gray, Litz, Hsu, & Lombardo, 2004)]:

LEC is a 17-item measure of self-report of potentially traumatic life events where the veteran endorsed items as happened to me, saw it happen to someone else, and heard about it.

Credibility and Expectancy Questionnaire [CEQ(Devilly & Borkovec, 2000)]:

The CEQ is a 6-item measure designed to assess treatment expectancy and rationale credibility of interventions. The scale has high internal consistency and test-retest reliability.

Mini-International Neuropsychiatric Interview [MINI 2.0(Sheehan et al., 1998)]:

The MINI is a brief structured interview that assesses the criteria for DSM-V Axis I diagnoses. The MINI exhibits similar sensitivity and specificity to more time-intensive structured psychiatric interviews. Only the alcohol and SUD modules were used to screen for severe alcohol and/or substance use disorder.

Mental health and medical services utilization:

In order to capture whether receiving PE-PC intervention changes health care utilization we collected utilization data for all VA health services received by the study patients 24 weeks prior to study enrollment and the 24-week period of post randomization. Outpatient visits included a clinical encounter with a VA provider, including MH clinicians (e.g., psychologist, psychiatrist, social worker) and PC providers. We examined the number of inpatient and outpatient MH visits, the number of PC outpatient visits, ED visits, specialty clinic visits, and inpatient medical admissions.

COVID-19 Research Therapy Impact Questionnaire:

When the COVID-19 pandemic occurred, we wanted to capture exposure and impact on our study. As such, in collaboration with several VA investigators, we created and added a ten-item questionnaire that assessed the impact of the COVID-19 pandemic on research participation and daily life (McLean et al., 2022).

Treatment Drop and Missing Data.

We followed patients regardless of treatment completion to model treatment response using an intent-to-treat approach. We made every effort, including follow-up phone calls, to collect follow-up data to reduce missingness to a minimum. We will check for patterns of missingness and compare rates of missingness and dropouts at each follow-up time between groups. We will check if dropout depends on baseline covariates using logistic regression, and if we find baseline covariates to predict dropouts, we will include them in the analytic models. For missing items within scales, we will use recommended imputation procedures.

Data Analytic Plan.

We will examine potentially confounding baseline variables to compare PE-PC and PCMHI-TAU groups using t-tests and chi-square tests as appropriate. Although we randomized patients to intervention groups, we will include baseline variables considered as potential confounders in the relationship between treatment and outcomes as covariates as needed in adjusted outcome analyses. Potential covariates were military era (Vietnam, Persian Gulf, Operation Enduring Freedom (OEF)/Operation Iraqi Freedom (OIF)/Operation New Dawn (OND)), baseline function level, PTSD or depression severity, and demographic characteristics.

Hypothesis 1.1:

PE-PC will result in larger increases in function than PCMHI-TAU (Week 0 to Week 6), and these differences will remain at 12- and 24-weeks post-baseline.

To evaluate, we defined effectiveness as a group difference at 6 weeks and maintenance as a group difference at 12 and 24 weeks. To examine the effectiveness and maintenance of a brief PE-PC protocol compared to PCMHI-TAU, we used a linear mixed model with WHODAS total scores at 0, 6, 12 and 24 weeks as the dependent variable. Mixed models do not require the assumption of homogeneity of variance in the treatment groups, permit flexible specification of the covariance matrix of the repeated measures and give unbiased estimate with incomplete data assuming missingness at random. The model included each participant as random intercepts and adjusted for study site and potential confounding patient characteristics identified in baseline analysis. Primary predictors included PE-PC group indicator, time in weeks, and interactions of treatment group by time variable.

If time showed a non-linear effect (e.g., lack of maintenance at 12 and 24 weeks), we can model time as categorical indicators for each assessment time and will include interaction terms of PE-PC group by time indicators. Based on the model coefficients and contrasts of the coefficients, we will report gains at 6 weeks relative to baseline for each group, compare the outcome at 6 weeks (for effectiveness), 12 and 24 weeks (for maintenance) separately between groups. We will also assess for a significant drop in function from 6 to 12 and from 6 to 24 weeks within PE-PC group for loss in initial gain (another type of maintenance) if week 6 gain is significant. If no significant difference across the three time indicators by group interaction effects is seen, we will drop time indicators and the interaction terms and include a post randomization time indicator and its interaction with PE-PC group indicator where the coefficient of the interaction term will represent the time-averaged PE-PC effect during weeks 6 to 24.

Hypothesis 2.1–2:

PE-PC will result in larger reductions in PTSD severity (H2.1: CAPS and PCL-5 (49, 50)) and depression severity (H2.2) than PCMHI-TAU (Week 0 to Week 6), and these differences will remain at 12 weeks and 24 weeks post-baseline. We will use a similar analytic approach as for Aim 1 using a mixed model for each outcome measure, with both baseline and all follow-up assessment data as the dependent variables. The model will include PE-PC group indicator, follow-up time indicators with week 0 as reference time, and interactions of treatment group by follow-up time indicators. As in Aim 1, we will also assess maintenance as drop at 12 or 24 weeks in improvement gained at 6 weeks for both PTSD and depression symptoms using appropriate contrasts based on the coefficients from the linear mixed model.

Hypothesis 2.3:

PE-PC will result in a greater proportion of veterans remitting from PTSD than PCMHI-TAU at week 6, 12, and 24. CAPS-5 below 20 will represent remission. We will report proportions remitted by the treatment group with confidence intervals. To compare remission between two groups at 6, 12 and 24 weeks, we will use a generalized linear mixed model with a logit link. We will assess the dependent variable of remission at 6, 12 and 24 weeks, and the primary predictor will be the PE-PC group indicator and indicators for 12-week and 24-week time point, and the interactions of follow-up time indicators by treatment group will test if the relative odds of remission between the two intervention groups differ across the follow-up times. We will adjust the model for baseline function score, symptom severity, and intervention site.

For the exploratory aim to assess potential moderators of improvements in function and PTSD symptoms, potential moderators include combat experience severity and exposure to life events. We will extend the primary mixed model to include these measures and their interaction terms with PE-PC arm to understand and assess if the treatment effect differs by the level of these potential moderators. We will also extend the mixed-effects model by including PTCI as a time-dependent covariate to explore if increased self-efficacy, as assessed by PTCI, decreases function or symptoms. Last, in the control (PCMHI-TAU) arm, we will include the use of specialty care during the prior six weeks as a time- dependent covariate to examine if it explains improvement in function or symptom in the control group. This exploration will assess whether contamination of participating study providers’ referring control patients to specialty care explains some of the improved function or symptom reduction in the control group, which might have reduced effects of PE-PC.

Discussion

The goal of the IMPACT study is to examine the comparative effectiveness of PE-PC and TAU in VA PCMHI on functional (primary) and symptom outcomes. Results will inform whether this brief intervention provided in PC can improve function and reduce symptoms of PTSD and related mental health problems in veterans desiring PTSD treatment in the PC setting. Whereas previous studies found initial efficacy of PE-PC in reducing PTSD compared to minimal attention controls (Cigrang et al., 2017), the current study will extend those findings to examine whether function, in addition to symptoms, changes more in PE-PC than TAU, and whether the changes seen in military treatment facilities are similar in a VA sample. Increasing veteran access to effective PTSD treatment is a priority for VHA, and PC is a setting that can accelerate access to care. In addition, PTSD includes significant avoidance. Providing quick access to effective treatment when PTSD is first identified or when the veteran first considers treatment can prevent some barriers to care and potentially increase retention and overall population outcomes.

In addition, providing simplified versions of effective interventions, like PE, can provide a springboard for implementation in new settings and modes of care. For instance, as PE-PC has been in development, additional work on written exposure formats and self-guided formats has built on the success of these new treatment models [Written Exposure Therapy (Sloan, Marx, Lee, & Resick, 2018); WEB PE (McLean et al., 2020)]. The expected mechanism/s of exposure therapy remain the same, though the intervention is provided in a manner that patients can access more easily and even work on independently. Bringing effective treatment out of specialty mental health and into new settings can transform how veterans and others with PTSD access care providing additional avenues to response and remission. While such modifications are exciting for the field, ensuring that these new models of care provide adequate effectiveness is critical and the IMPACT study design is intended to evaluate the effectiveness of PE-PC. The current protocol of comparative effectiveness in a new setting provides a model for additional future clinical trials in other setting outside specialty mental health.