Peer support during in vivo exposure homework to reverse attrition from prolonged exposure therapy for posttraumatic stress disorder (PTSD): description of a randomized controlled trial

Treatment “dose,” measured in terms of number of sessions completed post randomization, will be greater for participants in the PE + Peer Support during the in vivo exposure assignment condition compared to the PE + Standard Supportive Peer condition.

Treatment “outcome,” measured in terms of PTSD symptom reduction scores, will be significantly more improved for participants in the PE + Peer Support during the in vivo exposure assignment condition relative to the PE + Standard Supportive Peer condition at each assessment point (post treatment, 3-month and 6-month follow-ups).

Patients meeting the inclusion criteria will be asked to maintain medications at current dosages where medically possible. Participants who have not initiated new prescription medications in the previous 4 weeks will initiate treatment 1 week following the assessment battery. However, potential participants who have recently begun trials of prescription medication will be required to wait 4 weeks post assessment to ensure medication stabilization, at which point the assessment battery will be re-administered. This re-administered battery will be used for pre-treatment data in analyses.

Potential participants will be directly contacted by study staff when recruitment derives from the weekly PTSD clinic staff meetings. Potential participants who respond to brochures, or who are referred by other providers will be directed to the study coordinator for face-to-face explanation of the study in a private treatment room at the VAMC, where they will review informed consent and have all questions answered. All potential participants will first complete the assessment interview; immediately after the assessment interview, participants meeting inclusion and exclusion criteria (specified above) will be randomized to condition by the study coordinator using the randomization protocol designed by the study statistician. Participants will be told to expect a telephone call from their counselor to start treatment within 7 days.

All participants will receive US$30 for the baseline assessment, US$30 for the post-treatment assessment, US$35 for the 3-month follow-up assessment, and US$45 for the 6-month follow-up assessment for a combined possible total of US$140.

Both treatment conditions use peers to supplement PE treatment, and the same peers will be used in both conditions, thereby controlling for non-specific factors associated with individual peers. Contamination across conditions is rather easy to avoid given the nature of the peer activities in the two conditions: the experimental condition uses peers who provide support directly during scheduled in vivo exposure homework (see below) and the control condition uses peers in a manner consistent with current VA practices insofar as they offer frequent telephone support to “check in” to see how VA treatment is progressing and offer support to continue treatment, but never engage in any support during in vivo exposure therapy homework. VA PTSD clinic therapists will be asked if they have any patients who have completed treatment and did very well, both overall and in particular with in vivo exposure therapy components. Therapists will then be asked to contact these patients, inform them about the exposure therapy workout buddy program, and ask them if they would like a member of our VA/DoD research team to discuss the program further. Upon being contacted, the research team will outline the nature of expectations and research characteristics, and schedule a training session. Thus, initial nominations for the program are based on therapist impressions of successful candidates after they have completed treatment. As such, the therapists know the candidates’ strengths. Peers who are candidates will complete informed consent, and then be evaluated for the presence of a PTSD diagnosis via the Clinician Administered PTSD Scale-5 (CAPS-5), and only those who no longer have the PTSD diagnosis will be permitted to participate in this study. Peers will not be matched on gender or age, except in cases of Military Sexual Trauma, for which peers are matched on gender. Peers will be selected on the basis of geographic proximity to participants.

The experimental treatment is a modified version of PE insofar as social support in the form of peer workout buddies for in vivo exposure homework is added to the protocol. Specifically, PE is a manualized treatment [12] based on emotional processing theory, which suggests that traumatic events are incompletely and inaccurately encoded in memory as “fear networks.” Gradual exposure to corrective information via the confrontation of (i.e., exposure to) conditioned traumatic stimuli within a safe and therapeutic environment results in a competing and antithetical memory structure that inhibits the conditioned fear response. PE relies on two primary therapeutic tools: in vivo exposure and imaginal exposure. During in vivo exposure, the patient confronts feared, but safe, stimuli or cues that elicit trauma-related distress. During imaginal exposure, patients “re-live” the traumatic event, providing a detailed verbal account that includes sensory information, thoughts, feelings, and reactions experienced during the traumatic event. PE includes the following session by session components: (1) psychological education about the common reactions to traumatic events and presentation of the treatment rationale (sessions 1 and 2), (2) repeated in vivo exposure to traumatic stimuli such as people, places, things, or situations that resemble or trigger memories of the traumatic event but are realistically safe (in vivo exercises are assigned as homework during sessions 3 through 11), (3) repeated, prolonged, imaginal exposure to traumatic memories (imaginal exposure is implemented during sessions 3–11; patients listen to session audiotapes for homework between sessions), and (4) relapse prevention strategies and further treatment planning (session 12).

In the spirit of the new VA mandate to use peer specialists in specialty clinics such as PTSD clinics, we will offer those individuals who indicate that they have decided to drop out or have dropped treatment (e.g., stopped attending sessions) the opportunity to have a peer (see description of peer and training below) who has been through treatment successfully and no longer meets PTSD diagnosis help them to complete in vivo exposure trials as a peer, by offering social support and encouragement during exposure trials. Specifically, participants who have dropped out from treatment will be asked if they would like to try PE treatment again, this time with a peer for support.

Our pilot study experience allowed us to refine the methods of initiating the peer support to the following steps: once patients have agreed to re-initiate treatment, peers will be instructed by the project coordinator to call in via televideo/meet in person to the next therapy session to make introductions and listen to the patient and therapists review the next item of the in vivo exposure hierarchy in depth. The location, timing, outline, description, and parameters of the in vivo exposure therapy assignment will be reviewed and will be clear to patient, peer, and therapist. Once this clarity has been achieved, peer and patient finalize arrangements to meet at a set time and place to engage in this exposure trial and will be directed to arrange three such meetings per week for 3–4 weeks. The analogy of a workout buddy will be used, and the purposely time-limited aspect (one quarter of the normal treatment length) of peer support for in vivo trials will be restated.

For patients who dropped out of treatment prior to developing and engaging in hierarchy item exposure, the peer support “session call-in” will be delayed until this hierarchy is established. Once established and reviewed, the peer will call into the next session and discuss the logistics of meeting the patient at the item location with patient and therapist. Upon re-entry to treatment, participants, peers, and therapists will hold a conference call to review in vivo exposure trials with which exposure homework peer support will assist and a brief discussion of logistics will be held. Then, therapists will speak with only the participant to assure that they continue to want to participate with a peer. The therapists will ask the patients to comment on any problems or benefits associated with in vivo exposures accompanied by the peer during treatment sessions each subsequent week to obtain a progress report of how the exposure trials are going, and to determine whether there were any issues that should be discussed in person with the participant. After the first peer exposure event, therapists will also contact the peer to determine whether they would like to continue with the participant and if there are any issues or problems. If a peer indicates that they would no longer prefer to work with a particular participant, a different peer will be assigned. If two peers indicate for any one participant that they prefer not to work with that participant, the peer support will be terminated for that participant, and therapists will include these interpersonal events as topics of therapeutic focus.

All peers engage only in hierarchy items jointly determined by the therapist and participant, with assent of the peer. No new, or unscheduled exposure activities, initiated on the part of the peer or the participant are permitted. Peers are directed to communicate with participants if either party feels or seems to feel uncomfortable with the activity in question. The difference between discomfort arising from conditioned anxiety that is part of PTSD and discomfort arising from feelings that the situation is inappropriate are discussed with the therapists when needed.

Only situational activities, in clearly safe places, will be included in the in vivo exposure participation events by peers. This will be determined by review with the therapist and the peer. If either party feels that there is more than minimal risk, the activity will not be included. Note, the risk of anxiety on the part of the participant is not a reason to exclude an activity (indeed, producing and dealing with such anxiety in a supportive environment is the point of in vivo exercises). It is impossible to list every possible activity that should be avoided vs included. However, several common activities are precluded. Specifically, we will exclude:

Peers and participants are told that the peer support is only in place for a given patient for 3–4 weeks, three to four times per week, and is a means by which to allow participants to progress to their own, independently conducted in vivo exposure trials. Thus, “phase out” of the program is built in from the beginning. Peers will focus on helping participants to engage in those exposure exercises that typically are social in nature. Again, any exposure trials, such as driving, that require the participant to engage alone, or from which peers are specifically prohibited will not be the target of this program. If participants and workout buddies agree, the length of time for exposure trial assistance can be extended to 6 weeks, three to four times per week, but this will be based on therapist judgment with respect to therapeutic gains vs risk of becoming dependent on workout buddy.

Participants in this condition will receive PE (see Foa, Hembree, & Rothbaum, 2007 [12]) and will be assigned a peer volunteer who will not engage in any support during the in vivo exposure homework. The primary purpose of the peer in this condition is to emulate standard VA PTSD peer support procedures. Therefore, these peers provide general support, such as reminding and encouraging participants to attend appointments, expressing concern when therapy appointments are missed, checking up on them once per week via telephone and asking about current treatment progress, life stresses, opportunities, problems and successes, and offering general support with VA programs. The PE component of this treatment will be matched to the experimental condition in terms of session number and duration (12 sessions); treatment completion will be defined identically across both conditions (minimum 10 sessions or pre-specified reductions in PCL-5 score at two consecutive sessions). Counselors will deliver treatment according to the same schedule as the experimental condition, once weekly for 75–90 min. Thus, the level of peer and patient contact in the standard peer support condition is designed to be the same as that of the experimental condition. Thus, peers call several times per week for 3–4 weeks at about sessions 3–8, offering support and encouragement.

Each assessment interview will be conducted by raters blind to treatment condition. Raters will be qualified to master’s level and specifically trained in diagnostic interviewing techniques using the CAPS-5. Raters will be required to complete five practice standardized interviews and achieve item-level inter-rater reliability of at least 90%.

The following interview and self-report measures have been widely used in the clinical evaluation of adults with PTSD, and will be used in the present study. They will be collected at baseline, post treatment, and 3- and 6-month follow-ups. The PCL-5 [22] and Beck Depression Inventory-II (BDI-II) [23] are also collected every 2 weeks during active treatment.

Major Depression, Panic, and Substance Dependence diagnoses will be evaluated using this structured clinical interview based on the DSM-5. The onset of each symptom set will be specified. With the previous version, Ventura and colleagues [25] found excellent inter-rater reliability across a variety of disorders (overall kappa = 0.85).

The PCL-5 is a new version of the PCL, among the most commonly used self-report measures of PTSD symptom presence and intensity. Like its predecessor, the PCL-5 is structured to correspond to the DSM-5 PTSD criteria. The 20 items are scored on a 0–4 Likert scale for each symptom corresponding to “Not at all” to “Extremely.” Total scores on the PCL-5 range from 0 to 80. Initial psychometric data are encouraging. With college student samples, Blevin et al. found that PCL-5 scores exhibited strong internal consistency (α = .94), test-retest reliability (r = .82), and convergent (rs = .74 to .85) and discriminant (rs = .31 to .60) validity. With veteran samples, Bovin et al. (2015) found that the PCL-5 had good internal consistency (α = .96) and test-retest reliability (r = .84). Moreover, signal detection analyses using the CAPS-5 indicated a diagnostic cutoff score of 31–33 on the PCL-5 optimally categorized PTSD diagnosis.

The DRRI is collection of self-report measures assessing 14 key deployment-related risk and resilience factors with demonstrated implications for veterans’ long-term health.

The BDI-II is a 21-item self-report scale, is among the most widely used instruments to measure depression. Beck et al. [28] demonstrated that the BDI-II has high internal consistency (α = .86–.91).

The SF-36 is a 36-item questionnaire that measures health status, social support, and functioning over the past 4 weeks. The SF-36 has good test-retest reliability as well as sensitivity to change in health status [29, 30]. Lin and Ward [31] found the SF-36 to have high internal consistency (Cronbach’s alpha > .87) and found that the subscale’s reliability coefficients ranged from .59 to .89.

Treatments in both groups will require patients to carry out “homework assignments.” These forms will be collected and monitored each week during treatment. The percentage of returned, completed forms will be computed as an indicator of adherence to homework assignments.

Patient satisfaction is an important treatment outcome variable, closely associated with clinical outcomes [33]. The CPOSS-VA is 16-item measure, with a Likert scale response format, based on a general measure of patient satisfaction [34]. In a sample of veterans, preliminary data showed excellent reliability (alpha = .96) and good convergent validity with anchors (“would you recommend this treatment?”) [32].

We will adopt a relatively direct analytic approach for this first objective. Specifically, if preliminary descriptive analyses reveal baseline differences in psychopathology or other potentially confounding variables, including number of sessions completed prior to dropout (i.e., prior to entry into this study) these will serve as covariates in analysis of covariance (ANCOVA), which will be used to compare the average dose received in terms of number of treatment sessions in each group. These ANCOVAs will be repeated with race and gender as between-group factors to determine whether any interaction is present between treatment condition and these demographic characteristics. A similarly direct approach will be taken for the dichotomous variable of treatment completion, for which we will initially use chi-square analyses for bivariate examination of treatment completion in terms of each demographic variable and each treatment condition, followed by logistic regression in which those variables producing significant chi-square relationships to this outcome will be included in regression models. We will use 95% confidence intervals (CIs) for single proportions to estimate the proportion of participants who complete treatment within each intervention group. We will also estimate the difference in completer proportions between groups using 95% CIs for difference in proportions. In addition, frequency distributions describing participants’ reasons for discontinuation of study participation will be developed. In this way, the effect of each treatment condition on dose measured in terms of overall number of sessions completed, and in terms of proportion of participants completing at least 10 sessions/achieving clinical cutoff will be described.

Our primary outcome is the PCL-5 continuous score for PTSD. Clinical (CAPS-5, PCL-5, BDI-II), functional (SF-36), and process (treatment credibility scales, patient satisfaction, session rating scale, service delivery perceptions, retention in treatment) outcomes will be compared using the generalized linear mixed models (GLMM) approach.