SAH training program overview
Training sites and participating clinicians
As part of larger implementation efforts following the VA DV/IPV task force recommendations, six hospitals were designated by the VA DV/IPV task force as “rapid implementation sites” for the program. An additional three sites were added, for a total of nine in-person SAH trainings attended by personnel from 11 different hospitals in the first year of the roll-out. All sites participated voluntarily in the training program. The six rapid implementation sites were chosen by the VA DV/IPV taskforce because these sites had evidenced early commitment to the development of programming to address DV/IPV. Three additional training sites were added; two of these three sites were considered to have evidenced success in implementing other DV/IPV programming (though they were not designated as rapid implementation sites for the national program). Another training site was chosen due to its prior research experience with SAH; however, this site subsequently dropped out of the training program and is not included in analyses. In addition, clinicians from two other hospitals in close proximity to one of the designated pilot sites also completed SAH training. In total, 10 VA hospitals and 79 clinicians received SAH training and were enrolled in the pilot training program. All sites were required to have an IPV Coordinator tasked with overseeing the program and at least 4 clinicians who would volunteer to receive the training and commit to facilitating groups. All clinicians were required to be licensed mental health providers at the VA (i.e. social worker, psychologist, marriage and family therapist, licensed professional counselor). The training expectation was that at least one group would be initiated at each site within six weeks of the in-person training.
Training process and content
Prior to the training, IPV coordinators received a brief pre-training orientation call to answer any questions they had regarding the training, program expectations, and implementation tasks. Training was delivered in person over two days by the three SAH master trainers who were involved in developing SAH as well as in the conduct of the original SAH clinical efficacy trial [8, 9]. Each master trainer was a licensed clinical psychologist with years of experience working with individuals who use IPV, experience treating veterans and military families, and detailed knowledge of the SAH curriculum. The training consisted of one day focused on the theoretical underpinnings, program evaluation, and general therapeutic approach of SAH, with approximately 3 h of training in the use of motivational interviewing skills with veterans who may be using IPV. Attention was paid to ethics and mandated reporting statutes as well as to basic information on risk for IPV, dynamics of IPV in veterans, and clinical work with individuals using IPV. This was followed by a second day of detailed training on how to facilitate each of the 12 SAH sessions, augmented with role-plays of critical skills. Clinicians were provided with a treatment manual that described in detail the theoretical aspects of SAH and included all treatment materials. Also included in the manual was an assessment protocol detailing how to conduct the SAH intake interview and partner calls, clinician fidelity checklists, IPV fact sheets, SAH research publications, advice and materials for administering the protocol with LGBTQ and women veterans, and sample materials for publicizing the program and communicating with the justice system. Shortly after the in-person training, all clinicians were invited to an assessment and program evaluation training call.
After facilitating each SAH session, clinicians used a fidelity checklist to track their completion of all required elements for that session (using a yes/no checklist). All clinicians were required to complete weekly clinical consultation with the lead program developer while they led groups, and IPV Coordinators were required to attend monthly implementation consultation calls. In addition, the master trainers were available by phone and email for supplemental consultation as needed. To successfully complete the training program, clinicians were required to: 1) attend the full 2-day SAH training, 2) complete at least 2 SAH groups, 3) attend and participate in at least 75% of consultation calls, and 4) return program evaluation materials. After completing these requirements, clinicians could be cleared to facilitate SAH groups without concurrent clinical consultation.
Program evaluation procedures
Demographic information, session attendance, and scores on the program evaluation measures completed by program participants at the intake assessment and again at session 12 were summarized by site clinicians, and de-identified data were sent to the coordinating site. Summary information on recent IPV gathered from calls with veterans’ partners was likewise sent to the coordinating site. As the purpose was program evaluation only, no identifying information was collected by the evaluation team. To facilitate de-identification of the data, each participant was assigned a unique code by clinicians, and the same code connected veteran and partner data. The institutional review board at VA Boston reviewed the procedures and determined that the program evaluation did not meet the criteria for human research and thus was exempt from further IRB review.
Program evaluation measures
IPV
Recent use of IPV was measured with an adapted version of the Centers for Disease Control’s 2010 National Intimate Partner and Sexual Violence Survey [19]. This 30-item measure assessed four types of IPV behaviors: 1) Psychological Aggression (e.g. “I acted very angry towards my partner in a way that seemed dangerous”), 2) Coercive Control (e.g. “I tried to keep my partner from seeing or talking to their family or friends”), 3) Reproductive Control (e.g. “I tried to get my partner pregnant when they did not want to get pregnant”), and 4) Physical Aggression (e.g. “I slapped my partner”). At two time-points (pre and post-treatment), veterans and their partners reported whether or not the veteran had engaged in each IPV behavior in the past 3 months on a dichotomous scale: 0 (no), 1 (yes). Following the recommendations of Straus and colleagues, prevalence scores were then calculated by summing the number of positive responses for each subtype of IPV, and converting the sum into a dichotomous variable representing the presence or absence of each type of IPV in the past three months [20]. Prevalence scores indicate whether one or more acts on the subscale occurred in the past three months. Therefore, a score of one indicates one or more types of IPV behaviors measured on each subscale occurred, whereas a score of zero indicates none of these behaviors occurred. In accordance with previous methodology, veteran-reported and partner-reported items were compared and if either person reported that a behavior occurred it was coded as present [21]. Summary scores representing the number of subtypes of IPV (i.e., psychological, coercive, reproductive, physical) were then computed (possible range = 0–4).
PTSD symptoms
PTSD symptoms were measured with the PTSD checklist for DSM 5 (PCL-5) [22]. The PCL-5 is a 20 item self-report measure of PTSD symptoms in the past month. Items are rated on a 5 point Likert scale (0 = not at all, 4 = extremely) and participants endorse symptoms based on “a very stressful experience.” Items are summed with higher scores reflecting greater symptomatology in the past month. A cut score of 33 is suggested as indicative of probable PTSD [23]. The measure evidences good reliability (Internal consistency = .96; test-retest = .84), discriminant and convergent validity [23].
Alcohol misuse
The 10-item Alcohol Use Disorders Identification Test (AUDIT) was used to assess problem drinking behavior over the past year [24]. The AUDIT is a screening questionnaire with three questions on the amount and frequency of drinking, three questions on alcohol dependence, and four questions on problems caused by alcohol; responses (on a 0 to 4 point scale) are summed, and higher scores reflect greater levels of alcohol misuse. A cut score of 8 is suggested as indicative of problem or hazardous drinking [25]. The AUDIT has a reported median reliability coefficient of .83, and adequate construct and criterion related validity [26].
Treatment satisfaction
Veterans also completed a 10-question post-treatment satisfaction measure, which was based on the 8-question Client Satisfaction Questionnaire (CSQ) [27, 28]. The adapted measure inquired into patient satisfaction with SAH and services received. Patients rated their satisfaction on a 4-point scale (from 0 to 3), and also responded to an open-ended question regarding the most and least helpful aspects of the program. Analyses of the open-ended responses were outside the scope of the current project and are not reported here. The CSQ-8 has been shown to evidence excellent internal consistency and to correlate positively with self-reported symptom change, as well as client and therapist reported improvement [28].
Implementation outcomes
Hospital adoption was the length of time in days between receipt of the in-person training and the start of the first group session for each site. Clinician adoption was the length of time in days between receipt of the in-person training and the start of the first group session for each clinician. We also report the percentages of clinicians who fully or partially completed the training program, as well as those who dropped out of the training program and their reasons for drop-out. For those sites who were not successful in implementing the program in the first year, we report the reasons why this may have occurred.
Analyses
Paired samples t-tests were used to determine the significance of pre to post-treatment changes in measures of patients’ use of IPV, PTSD symptoms and alcohol misuse. Effect sizes for repeated measures designs (Hedge’s grm) were also calculated [29]; an effect size of 0.2 is considered small, 0.5 is considered medium, and 0.8 is considered large [30]. The McNemar Chi Square test (without continuity correction) was computed to examine changes in the proportion of veterans using each subtype of IPV from pre-to-post treatment. All participants with data from both the pre and post-treatment assessment were included in these analyses. The IPV data, as described in further detail below, represent the report of veteran-to-partner IPV based on either the veteran’s self-report or the partner’s report (where possible). However, in order to mitigate against the influence of differential sensitivity for some veterans (as partner reported data was not available for everyone), all IPV analyses were repeated using veteran report only. The pattern of significant findings was identical; therefore, we report analyses only for the combined veteran-partner IPV data.