HSR&D Home » Research » CRE 12-083 – HSR&D Study
Motivational Coaching to Enhance Mental Health Engagement in Rural Veterans
Karen H Seal, MD MPH
San Francisco VA Medical Center, San Francisco, CA
San Francisco, CA
Funding Period: August 2014 - September 2018
There is a substantial burden of mental health (MH) problems in rural OEF/OIF/OND veterans. After a decade of war, over 51% of OIF, OEF, and OND veterans in VA healthcare have received MH diagnoses; the majority (27%) have received diagnoses of posttraumatic stress disorder (PTSD).1 Studies show that veterans residing in rural areas experience significantly greater MH severity and poorer outcomes than their urban counterparts2, 3, 9 Surprisingly, there are no published studies on the differential MH burden among OEF/OIF/OND veterans in VA healthcare based on rurality. To begin to address this knowledge gap, using rural-urban commuting area (RUCA) zip code data to define rurality 10, 11 and national VA administrative data to obtain ICD-9 MH diagnoses codes, our group found that increasing rurality was associated with a higher prevalence of MH disorders in OEF/OIF/OND veterans nation-wide and in VISNs 16 and 21. (Seal, preliminary data) For instance, compared to the prevalence of MH diagnoses among urban OEF/OIF/OND veterans in VISN 21 (44.7%), the MH burden was higher in rural veterans(47.4%) and even greater in "isolated rural" veterans (54.6%),(Relative Risk=1.22, 95% CI=1.11-1.34 for MH diagnoses in isolated rural vs. urban veterans) (Seal, preliminary data).
The majority of OEF/OIF/OND veterans with MH problems do not receive an adequate course of MH treatment. The VA Uniform Mental Health Services Handbook mandates that all veterans, including those receiving care at CBOCs serving rural veterans, have access to evidence-based MH treatments. Minimally adequate MH treatment has been defined as 8 MH treatment sessions or receiving 2 months of psychiatric medication plus > 4 visits within 1 year.5 Unfortunately, the majority of OEF/OIF/OND veterans have not received an adequate course of MH treatment as found in a nationally representative sample of veterans,12 and veterans enrolled in VA healthcare.6, 13 Indeed, at the San Francisco VA Medical Center (SFVAMC), our group demonstrated significantly improved MH treatment initiation in OEF/OIF/OND veterans who presented to our new co-located primary care-mental health clinic compared to usual primary care, but sustained engagement in specialty MH services remained poor with drop-out after 1-2 sessions.14
Poor retention in MH services may be even more pronounced for rural OEF/OIF/OND veterans.6 13 Our own research showed that living >25 miles from a VA facility was a strong predictor of failing to receive adequate MH treatment.6 In a more recent analysis of MH utilization in OEF/OIF/OND veterans in the first year of receiving MH diagnoses nationwide, we found a significant association between increasing rurality and drop-out after 1-2 MH visits. Specifically, in VISN 21, only 35% of isolated rural veterans with MH diagnosis compared to 45% of urban veterans completed 8 MH visits (RR for isolated rural=0.77, 95% CI=0.59, 1.00)(Seal, preliminary data) Geographical distance is a significant logistical barrier, but rurality may also serve as a proxy for other access and engagement barriers such as cultural, financial, and digital barriers, as outlined in the State of the Art Conference (SOTA) Access Framework, developed by Dr. Fortney (see CREATE Overview).15 SOTA hypothesizes that the most salient barriers to care for rural veterans may be cultural barriers, including heightened levels of stigma, and lack of perceived need including negative beliefs about MH treatment, stoicism, and self-reliance.16-18
Motivational Interviewing is an evidence-based practice to promote MH treatment engagement in veterans. Telephone-based referral care management (Behavioral Health Lab) consisting of 1 to 2 telephone MI sessions plus pre-scheduled appointments for 113 older, depressed veterans resulted in significantly improved MH treatment initiation, but no improvements in clinical outcomes.19, 20 In another study of 114 older veterans (mean age 56 years), those randomized to 4 sessions of group MI demonstrated greater retention in PTSD therapy than veterans who received 4 sessions of psychoeducation.21 Neither of these MI-based trials was conducted in OEF/OIF/OND or rural veterans, but results underscored that MI, including telephone MI, improved MH treatment initiation and/or retention in care. In contrast, results from a Cognitive Behavioral Therapy (CBT)-based MH treatment engagement intervention for a small sample of OEF/OIF/OND veterans were mixed. In OEF/OIF veterans, using a pre-post quasi-experimental design, Stecker et al., showed that among 26 National Guard and Reserve veterans, one session of telephone-administered CBT significantly increased veterans' self-reported intention to engage in treatment, but did not result in actual increases in MH treatment initiation.22
Our recent pilot RCT demonstrated the efficacy of telephone MI to increase MH treatment engagement in OEF/OIF/OND veterans. This efficacy data supports our taking the next step in conducting a hybrid effectiveness-implementation study. Seventy-three OEF/OIF veterans who screened positive for one or more MH problem(s), but were not engaged in treatment, received an MH referral and were randomized to either 4 brief sessions of telephone MI or 4 brief neutral attention-control telephone sessions at baseline, 2, 4, and 8 weeks. Of note, MI was conducted by non-clinician Master's-level research staff who had been trained in MI for this study. Blinded assessment occurred at 8 and 16 weeks. In intent-to-treat analyses, 62% assigned to telephone MI initiated MH treatment compared to 26% of Controls [Relative Risk (RR) =2.41, 95% Confidence Interval (CI) =1.33- 4.37, p= 0.004], which represented a large effect size (Cohen's h=0.74). We also observed significant reductions in stigma about MH treatment and in marijuana use (both p-values<0.05). In addition, while this MI trial was not focused on MH treatment retention, the MI group also demonstrated significantly greater retention in MH treatment than Controls (Incidence Rate Ratio (IRR) =4.36, 95% CI=1.96-9.68), signaling that telephone MI could be used to enhance MH treatment retention in OEF/OIF/OND veterans.7 While the evidence from this pilot study supports progression toward implementation, because we are now targeting a rural, and not urban population, we propose a pragmatic effectiveness study using a hybrid design that will also allow us to critically evaluate the implementation strategy at the same time as the interventions' effectiveness.8
The conceptual underpinnings of Motivational Interviewing (MI) support an MI-based coaching intervention to enhance MH treatment engagement in rural veterans. MI is a patient-centered counseling style for enhancing intrinsic motivation for change by exploring and resolving ambivalence.23 MI is based on the principles of interpersonal, patient-centered psychotherapy 24 and the Transtheoretical Model of Change 25 which posits that patients move in graduated stages from pre-contemplation through contemplation to action. Rollnick places the construct of "readiness to change" at the center of his model and identifies two elements: "importance" and "self-confidence" regarding change that contribute to readiness.23 A specific MI strategy that we will use in this study, involves the use of the "Readiness Ruler" to gauge an individual's perception on a scale from 0 to 10 of the importance, their confidence, and readiness to make behavioral changes such as engaging in MH treatment.23 In multiple randomized controlled trials, MI has been shown to enhance both MH treatment initiation as well as retention in care, 26,27 and MI has been successfully adapted for administration by telephone.28,29 MI has been effectively used in culturally diverse populations, including rural populations; peer counselors who "speak the same language" may be highly effective in engaging rural veterans in care. 30, 31 Cultural meanings of MH problems may be reflected back to patients to explore ambivalence, negative beliefs and stigma regarding MH treatment. In addition, culturally-based strengths and coping mechanisms, e.g., spirituality, family and peer supports may be leveraged to promote treatment engagement.32 Indeed, ancillary healthcare staff and peer counselors have been successfully trained to conduct MI.33, 34 VA has a long-standing tradition of peer counseling as evidenced by 207 Vet Center Centers nationwide, many of which are staffed by veteran peers.35 Veteran peer counseling programs capitalize on shared experience and camaraderie to foster credibility and trust, decrease stigma and promote help-seeking in veterans.35 VA has also successfully used peer counseling in substance abuse and MH treatment programs.36, 37 In January 2011, the VA/DoD Defense Centers of Excellence produced a White Paper identifying best practices in peer support, drawing on the experience and outcomes of numerous veteran peer counseling programs, e.g., Vet2Vet, Vets 4 Vets, Vet Center etc..38 The White Paper serves a blueprint that will guide us in the hiring, training, and monitoring of experienced veteran peer counselors who will become part of an MI coaching team (see below). As an adjunct to MI coaching, this study will also refer study participants to "Considering Professional Help," https://www.myhealth.va.gov/course/ConsiderProffHelpPresentation/index.html an Office of Mental Health Services program available on MyHealtheVet, which, consistent with MI principles, features veteran peers working through and resolving their ambivalence about engaging in MH treatment.
In a randomized multi-site pragmatic effectiveness trial, compare the effectiveness of MH Referral alone with MH Referral plus MI-based coaching to improve MH services engagement in veterans receiving care at CBOCs. Compared to MH Referral alone, MI coaching will significantly:
H2a. Increase MH services initiation and retention (number of MH visits) (Primary Hypothesis).
H2b. Increase the use of e-health "self-help" MH treatment options, such as afterdeployment.org.
H2c. Increase perceived need and readiness for MH treatment, and decrease barriers to MH services.
Secondarily, we will evaluate change in mental health symptoms, high-risk behaviors (e.g., driving under the influence, etc.), functioning, quality of life, perceived access to MH care, and satisfaction with VA healthcare.
We will conduct a pragmatic effectiveness RCT of MH Referral plus the MI-based MH treatment engagement intervention vs. MH Referral only (Control) in veterans who receive care in VA CBOCs serving rural veterans (months 10-44, 34-month RCT). All participants will be enrolled and followed for 8 months. Enrollment will begin at study month 15 and will conclude at study month 46. The last wave of enrollment will begin at month 39 to allow a full 8-month follow-up period until month 46 (25 month enrollment period). This leaves 4 months for data analysis and manuscript preparation.
a. Study Setting for the RCT: The RCT will occur at the same VA CBOC sites (at least 2 CBOCs each in VISNs 16 and 21) identified during the Phase 1 developmental formative evaluation (DFE) (see Section B.1).
b. Sample Size for the RCT: Based on power calculations which already account for a 20% drop-out rate (Section B.6.n), we will need to enroll 140 participants per study arm (total N=280) over 25 months. Different from our efficacy study in which we recruited exclusively from administrative databases and had no prior knowledge of potential study participants, because this is an effectiveness trial, CBOC staff will recruit roughly 80% of the total sample of 280 (N=224) directly from CBOC sites (see below), referring veterans to the study who they believe to be eligible and interested in study participation. If we project that 60% of these veterans will enroll, we would need to recruit and conduct baseline assessments in roughly 375 veterans across the CBOC sites in 26 months or 188 from each VISN in 26 months or about 8 veterans/ month from CBOCs in each of the VISNs. To ensure adequate enrollment, we will plan for an additional 20% (N=56) recruitment using VA administrative data to identify veterans who use CBOCs in VISNs 16 and 21 (see below). Based on our prior study, we project an enrollment rate of 25%, and thus would need to recruit and conduct baseline assessments in about 7 veterans/month from administrative data in each of the VISNs. Thus, in sum, we project that we would need to recruit a total of 15 veterans a month from CBOCs across both VISNs.
c. Study Recruitment/Enrollment Strategy for RCT: Information derived from the Phase 1 DFE will inform the research team about which CBOC providers (e.g. MH social workers, PCMH-I staff, or RN nurse care managers etc...) at each participating CBOC facility will be designated to implement study recruitment. During Phase 2 formative evaluation (Aim 3), the research team and the Health Behavior Coordinator (HBC) will monitor the adoption and fidelity of the recruitment strategy and will provide feedback to individual CBOC staff conducting recruitment as well as discuss and adapt the recruitment strategy as needed during the Phase 2 EBQI meetings. This will contribute to the evaluation of adoption and implementation fidelity as outlined in the RE-AIM framework. Optimizing the implementation strategy for veteran recruitment from the CBOCs is fundamental to successful implementation and maintenance of the MI coaching intervention. Finally, as in our pilot study which enrolled 36% female veterans and 55% ethnic minorities, we will plan to over-sample women and ethnic minorities (see Human Subjects for details).7
Recruitment streams are as follows:
1. CBOC providers will refer veterans who they believe are moderately or highly ambivalent about engagement in mental health treatment. In addition, they will refer potential participants who have frequent NO SHOWs or who have prematurely dropped out of MH treatment during the past year. The CBOC Providers will refer eligible patients to the research staff to be assessed for study enrollment. Prior to referring veterans, CBOC staff will have the option to briefly explain the study, provide veterans an information sheet, and ask for their permission to be mailed recruitment materials. The research staff will then send recruitment materials outlined in (3) to the potential participant.
2. In addition, CBOC Providers will also identify their patients which they believe would meet study inclusion criteria and refer these names and last 4 of the SSN to the research staff (approved for a HIPAA Waiver for Recruitment). The research staff will then find the patient's contact information in CPRS and send recruitment materials outlined in (3) to the potential participant.
3. VA administrative data will be used to identify additional veterans in VISNs 16 and 21 who have received care at participating CBOCs within 1 year of the study start date. We will identify veterans who screened positive on VA MH screens or received MH diagnoses, but have never attended a MH visit (treatment naive), or attended 2 MH visits without follow-up (treatment drop-out) 90 days prior to the study start date. Potentially eligible veterans will be mailed Patient Letters, Study Information Sheets, Opt-out Letters (including a VA return envelope addressed to the local study site) using previously described IRB-approved "opt-out" methods. We will wait 2 weeks for Opt-out Letters to be returned. If subjects do not opt-out we will contact them via phone call. No cold calls will take place.
4. Self-referral will be used to identify Veterans interested in participating in the study through the use of various advertisement methods and word of mouth. Advertisement methods will include: posting flyers at places such as the VA clinics, Veteran Service Organizations, and public areas, using e-bulletins, direct outreach by the study team at events where large numbers of veterans congregate. Since the veteran will be contacting us directly to inquire, we will provide information over the phone and go into the eligibility screen if the veteran is still interested.
5. Finally, we will use a technique called "snowball sampling" to increase recruitment. After completing an interview, participants will receive a "thank you" letter which thanks them for their time and efforts as well as provides the names, addresses, and phone numbers of the referrals they received during the phone session. Enclosed in the letter will be small contact cards that the veterans can pass onto their peers who they think may be interested in participating in the study. Participants are under no obligation to distribute these cards or disclose that they are participating in the study. Choosing not to distribute the contact cards will in no way jeopardize their ability to continue participation in the study. The contact cards include study information. If a veteran who receives a contact card calls the study staff, the study staff with provide the veteran with information over the phone and go into the eligibility screen if the veteran is still interested.
d. Eligibility Criteria for RCT: As a pragmatic effectiveness trial, there are minimal exclusion criteria to maximize generalizability and facilitate future implementation. The Project Coordinator will call participants to conduct an initial telephone eligibility screen. After confirming initial study eligibility, the Project Coordinator will further describe the study, including risks and benefits, and obtain verbal consent to proceed to the telephone baseline assessment to confirm eligibility for the RCT (see below):
Inclusion criteria: from initial phone screen: (1) a Veteran of military service, over age 18 (2) a resident of VISN 16 or 21 catchment areas receiving with no plans to re-locate within 8 months of enrollment, and from baseline assessment: (3) positive for 1 of the following disorders PTSD, depression, generalized anxiety disorder, panic disorder, high-risk drinking, and/or illicit substance use (as ascertained at the baseline assessment).
Exclusion criteria: from initial phone screen: (1) hearing- impaired, (2) no working telephone, (3) Veterans with self-reported (and/or CPRS-confirmed) diagnoses of schizophrenia, psychosis or bipolar disorder(4) received mental health treatment within the last 60 days and/or future appointments for mental health treatment scheduled in the next 30 days (these questions will be repeated during the baseline assessment so that the information is documented in Qualtrics for data analysis) and from the baseline assessment: (5) active suicidality or homicidality. Prisoners will be excluded because they will not be able to engage in mental health treatment. Our staff only speaks the English language; therefore we must also exclude those subjects that are not proficient in the English language. We must also exclude those with impaired cognitive function because our psychometric measures are designed for self-response and cannot be validly answered by a subject's caretaker.
e. Baseline Telephone Assessment for RCT: In addition to determining RCT eligibility, the main objective of the telephone assessment is to obtain baseline and follow-up information on sociodemographics, military service, MH problems, psychosocial functioning, high-risk behaviors and MH treatment.
The Veteran Peer Coach will conduct the 60-min telephone-administered baseline assessment, entering responses directly in a web-based data management system. Studies have shown that telephone administration of mental health assessments is valid, reliable, and equivalent to face-to-face administration.94, 95
f. Real-Time Scoring of Baseline Assessments: A summary score from each of the 5 psychometric assessments will be automatically calculated using Qualtrics. Previously validated cut-scores and definitions will define a "positive" screen.82, 84-87
g. Randomization for the RCT: Otherwise eligible participants who screen positive for 1 MH problem(s) will be randomized to MH Referral + MI Coaching (Intervention) versus MH Referral alone (Control). Block randomization will be stratified by: (1) MH treatment history (treatment naive or treatment experienced), (2) binary MH disorder severity defined as at least one MH problem with a score indicating severe symptomatology, and (3) located in VISN 16 or VISN 21. Strata will optimize random distribution of veterans with differential likelihood of MH treatment engagement between the two study arms.
h. MH Treatment Referral (Both arms): At baseline, the Veteran Peer Coach will provide all enrolled participants in both arms MH assessment results followed by MH treatment referrals. During the Phase 1 DFE, with the assistance of VISN and CBOC staff, community providers, Drs. Prins (NCPTSD), Weingardt (OMHS), and Dundon (NCP), we will develop a comprehensive list of acceptable, feasible and effective MH treatment resources in close proximity to participating CBOCs in VISNs 16 and 21. This list will also include telephone resources (e.g., NCP's Telephone Lifestyle Coaching), smart phone resources (e.g., PTSD Coach), and online MH resources (e.g., www.afterdeployment.org, etc...)96 (see Appendix, E-Health Resources). Control participants will receive MH referral only.
i. Telephone MI Coaching Intervention: The MI coaching intervention proposed for this pragmatic effectiveness trial is the brief telephone MI intervention that we developed and pilot tested with OEF/OIF/OND veterans (see Appendix, MI Manuscript, In Press and MI Coaching Sessions).7 As an effectiveness trial, the implementation of this MI coaching intervention will be tailored to the individual CBOCs based on Phase 1 and 2 formative evaluations. Moreover, this trial will include 2 additional MI sessions focused on treatment retention for participants who initiate MI treatment because our pilot study signaled that MI could be used for retention and other studies, the VA and Congress are emphasizing the need to focus on treatment retention to complete a minimally adequate course of MH treatment.6, 7, 13, 58, 97 Within each VISN, the two separate MI coaching teams of experienced veteran peer counselors (as determined by Phase 1 DFE) will undergo the standard 4-hour VA NCP MI training conducted by the Health Behavior Coordinators based at the two VAMCs, with booster trainings as indicated by supervision (see below). Based on stakeholder preference (determined in the Phase 1 DFE), each VISN's MI coaching team will be centralized at one VA facility (i.e., Medical Center) serving the participating CBOCs. Considering that we will recruit 3-4 MI participants each month and coaches may carry participants for up to 6 months, if there are 2 part-time coaches, each coach may have a caseload of 9-12 veterans and if there are 3 coaches, each may carry 6-8 veterans at any given time, a caseload which was feasible in our pilot telephone MI study.
1. MI for MH Treatment Initiation: Veterans randomized to the MI coaching arm will be contacted by a coach for up to four 20-minute telephone MI sessions (0, 2, 4, and 8 weeks) initially focused on MH treatment initiation (Figure 1). In our pilot MI study, we demonstrated increased MI efficacy using non-scripted MI techniques,7 a finding corroborated by other studies.26 Thus, MI coaches will be guided by an MI coaching framework that emphasizes MI principles: (1) expressing empathy, (2) developing discrepancy between client's values and current behavior, (3) "rolling with resistance", and (4) supporting self-efficacy 23 (see Appendix, MI Coaching Sessions).
During the first MI phone session, coaches will provide participants personalized feedback about their baseline MH assessments. Coaches will use open-ended questions to better understand cultural meanings veterans ascribe to their MH problems as well as perceived access, need for and perceived effectiveness of MH treatment.32 Coaches will use reflective listening to elicit and reflect back veterans' culturally-based beliefs, concerns and priorities and will offer referrals to MH treatment services based on participants' preference and readiness for MH treatment. Subsequently, at the 2, 4, and 8 week telephone MI sessions, interviewers will begin with a Brief Assessment (BA) of perceived need, readiness for MH treatment (Readiness Ruler) and inventory whether or not veterans engaged in any VA or non-VA MH treatment, including telephone, smartphone, and e-health resources (using the MH Treatment Experiences questionnaire), and if so, the number of sessions. For participants who have not yet initiated treatment, coaches will work to build motivation to engage or make behavioral changes that may be hindering their engagement, such as drinking. Coaches will guide participants' in identifying culturally-based strengths and values (e.g., social support) that can be used to overcome ambivalence about MH treatment as they weigh the pros and cons of taking action. Coaches will use an MI technique of assisting participants to develop "discrepancy" by encouraging them to articulate goals for the future that might not match their current situation, thus tipping the decisional balance toward help-seeking.
2. MI for MH Treatment Retention: When a veteran in the MI arm indicates on Brief Assessment (BA) that they have initiated a qualified MH service (see below) they will enter the MI-for-treatment-retention track (Figure 1). The next telephone MI coaching sessions (2 and 6 weeks after treatment initiation) will employ the same MI principles described above, but will focus on overcoming barriers to and enhancing enablers of retention in MH treatment (see Appendix, MI Sessions). To avoid a duel therapeutic relationship, there will be only 2 MI sessions and coaches will remain focused on continued engagement in MH treatment.
j. Monitoring and supervision of MI Coaching fidelity: The MI coaching teams in each VISN site will be supervised primarily by the HBC and secondarily by the Project Coordinator. With participants' explicit permission, Drs. Manuel and Mesidor (HBC's) will supervise and rate at least 20% of each coach's MI sessions conducted over speakerphone. They will record the degree of fidelity (high, medium, or low) with which each coach performs MI coaching as rated by the HBC. This will contribute to evaluation of adoption and implementation fidelity in the RE-AIM framework. Following supervision sessions, HBCs at each site will provide feedback to coaches about the quality of the MI sessions as well as hold bi-monthly supervision calls for the coaching team to expand on feedback as well as review more challenging cases. When indicated, the HBCs will provide additional in-person MI booster trainings to individual coaches and/or the coaching team.
k. Outcome Ascertainment: An independent study RA, blinded to group assignment (to guard internal validity), will call participants to conduct assessments at the appointed time (see Outcomes Assessment schedule below). Participants will undergo Full Assessments (FA) at 8 and 16 weeks and a final Brief Assessment (BA) at 32 weeks (Table 2). In addition, for intervention arm participants, prior to each MI session, the MI coach will conduct brief assessments (BA) (as described in B.6.i.1) because information gleaned from the BA about participants' intervening MH treatment experiences, readiness and perceived need for MH treatment will be used to conduct the MI session. When there is both a FA and MI session (at 8 weeks), research staff will coordinate such that after the RA concludes the FA, the MI coach will be notified to contact the participant, which worked smoothly in our pilot study.
We present preliminary results for this Hybrid 2 Implementation-Effectiveness study. In accordance with this type of design, prior to the implementation of the trial we conducted a pre-implementation formative study at each of our study sites-eight rural VA community-based clinics in Northern California and Southeastern Louisiana. Our goal was to learn how best to implement the intervention (telephone motivational coaching by veteran peers) given the unique settings and contexts in which these rural veterans sought care. We first present our qualitative findings that informed the implementation of the randomized controlled trial (RCT) and then some preliminary findings from the trial that followed.
Specifically, for our first publication (Koenig et al., 2016), we used two pre-implementation strategies, Formative Evaluation (FE) research and Evidence-Based Quality Improvement (EBQI) meetings to adapt the intervention to stakeholders' needs and cultural contexts. FE data were qualitative, semi-structured interviews with rural veterans and VA clinic staff. Results were rapidly analyzed and presented to stakeholders during subsequent EBQI meetings. FE research results showed that VA clinic providers felt overwhelmed by veterans' mental health needs and they acknowledged limited mental health services at rural VA clinics. Rural veteran interviews indicated geographical, logistical, and cultural barriers to VA mental health treatment and a preference for self-management strategies in coping with mental health symptoms. EBQI meetings resulted in several intervention adaptations, including decreased reliance on clinic staff to assist with veteran recruitment, enhanced peer veteran coach training to include practice with actual veteran cases, the inclusion of high quality community resources, and an expanded definition of what the study team defined as mental health care engagement among rural veterans.
For a second publication (Abraham et al., 2017), we used data from semi-structured interviews with 37 veterans from Northern California and Southeastern Louisiana to illustrate how "health work" for mental health concerns are shaped by place, setting and context. Using health work as an orienting concept for analysis, we described differences between the two study sites regarding how veterans managed their mental health and well-being. In Northern California, there was a strong tendency to interact with nature (hiking, gardening etc.) to manage mental health symptoms, whereas in the South, veterans used religious practice to manage their mental health problems. In both settings, veterans spoke of cultivating "alone time" as key to managing mental health, although "alone time" was more pronounced in Northern California veterans than those from the South, were veterans spoke more about the importance of community.
For a third publication (Zamora et al., in preparation), we tackled the problem of uniformity in defining, operationalizing and measuring patient engagement in mental health care as well as the issue that patient-centered definitions of what constitutes mental health engagement have heretofore not been considered. In interviews with 37 rural-dwelling veterans, we assessed rural veterans' conceptualization of engagement in mental health care and how these definitions aligned with or departed from established definitions of mental health care engagement in the VA health care system. In most cases, we found that veterans' experiences of mental health engagement included community-based and self-care activities that would most likely not be captured using standard VA engagement metrics. Moreover, our findings illustrated how personal, social, geographical, temporal and financial factors influenced individual veterans' decision-making surrounding mental health care engagement in rural settings and helped us refine our conceptualization of mental health treatment engagement for our subsequent trial.
For the subsequent RCT, we screened 2,033 veterans for eligibility. We excluded 1,615 for not meeting eligibility criteria (n=488), declining to participate (n=804), unable to contact (n=164), and other reasons (n=159). We ultimately consented 418 participants for the trial. Between informed consent and the baseline assessment, three participants withdrew and one participant was excluded due to cognitive impairment. We thus conducted baseline assessment in 414 participants, of whom 127 were found to be ineligible after baseline assessment and seven withdrew before randomization. We thus randomized 280 participants, with 140 allocated to the intervention arm and 140 to the control arm. One participant from the intervention arm withdrew and requested that their data not be used. As a result, there were 139 in the intervention group and 140 participants allocated to the control group who were included in the baseline analyses (N=279).
Of the 413 participants who underwent baseline assessment, the 279 who were randomized did not differ significantly from the 134 who did meet eligibility criteria and were not randomized except that those randomized were significantly younger (50.7 vs. 56.3 years, P<0.001) and more were service-connected (71.3% vs. 60.5%, P=0.027). In addition, more randomized patients reported that they had received mental health treatment, counseling or support in the past five years than those not randomized (84.2% vs. 69.4%, P=0.002).
At baseline, there were no significant differences between the intervention group (N=139) and the controls (N=140) regarding sociodemographic or military service characteristics and there were few differences in mental health characteristics and mental health service utilization. Overall, women veterans made up 15.8% of the study population; the mean age was 50.7 years (SD +/- 13.5); 7.5% identified as Hispanic/Latino and of those, the majority (58.5%) identified as White, whereas 22.9 % identified as African American and 35.6% other races. Educational levels, marital status, income, service-connection, insurance coverage, quality of life and mental health screen results were also not significantly different between the two study arms. Across both arms, 87% screened positive for depression, 87% were positive for anxiety; 49.1% panic; 23.6% alcohol use disorder and 68.4 % screened positive for PTSD symptoms. Significantly more controls reported amphetamine use (p=0.05) and opioid use (p=0.01) than intervention participants, but in both cases less than 1% in either arm reported non-medical use of either substance.
Only participants who had not received mental health treatment in the past 60 days were eligible to participate in the trial. Nevertheless, at baseline, most (84.9%) reported receiving some mental health treatment within the past 5 years; the majority reporting having received most mental health care from VA vs. non-VA sources. Of note, there were no significant differences between the two groups regarding readiness to receive mental health treatment, with the average level of readiness being 6 on a 0-10 scale. When asked about their perceived need for mental health treatment however (using the GUPI scale), slightly more in the intervention than control group did not perceive a need for or desire mental health treatment (p=0.05).
After randomization, 7 participants withdrew and provided no post-baseline follow-up data. Thus, the analysis of the primary outcome- VA and non-VA mental health engagement (clinician-directed) included 272 patients who had at least one observation of the primary outcome during follow-up, with the majority (n=235) based on self-report supplemented with those identified by the VA Corporate Data Warehouse (CDW, n=37) as having not reported a mental health encounter, but having had a qualifying mental health encounter in the VA EMR. This analysis is valid based on the missing-at-random (MAR) assumption.
For this analysis, the 135 patients in the control arm and 137 patients in the intervention arm did not differ in demographics and military characteristics (age, gender, race, ethnicity, education, marital status, living situation, employment status, income levels, VA service connection, health insurance). The two arms were balanced in terms of mental health symptoms (depression, anxiety, panic disorder, and PTSD) and most substance use (tobacco, alcohol, cannabis, cocaine, inhalants, hallucinogens, and other substances), but the control arm had significantly higher use of opioids (0.9 vs. 0.2, p=0.009) and amphetamines (0.5 vs. 0.2, p=0.042) than the intervention arm. In addition, using the standard WHOQOL instrument, we measured quality of life in 4 domains at baseline: physical health, psychological health, social relationships, and environment. We also asked patients to rate their quality of life and satisfaction with their health. At baseline, the two arms did not differ in any of these quality-of-life measures. Finally, the two arms did not differ in receipt of mental health treatment, counseling and support at baseline.
The primary outcome variable, mental health engagement, was defined as receiving any of the following types of mental health treatment on one or more occasions: (1) mental health clinic-individual or group counseling, (2) drug and alcohol abuse treatment-individual or group counseling, (3) inpatient psychiatric treatment, (4) mental health counseling by a social worker, (5) marital, relationship or family counseling, and (6) prescription of a new psychiatric medication. Mental health engagement in both groups was high and we found no significant difference between the two groups. Sixty-two patients (45.9%) in the control arm and 61 patients (44.5%) in the intervention arm engaged in clinician-directed mental health treatment as defined above in VA and the community after baseline (p=0.817). There was also no difference between arms in the types of clinician-directed mental health treatment. We also examined non-clinician-directed mental health engagement, defined as participating in self-help groups or using internet/mobile mental health support applications. Similarly, we could detect no difference between groups with 14 patients (10.4%) in the control arm and 13 patients (9.5%) in the intervention arm engaging in non-clinician directed mental health support (P=0.491). To compare mental health engagement between the two arms while accounting for baseline differences, we used a survival analysis adjusted for baseline opioid and amphetamine scores. We found that on average, the control arm had a slightly longer follow-up period than the intervention arm (207 days vs. 201 days). An adjusted cox proportional hazards regression showed that the intervention arm did not differ from the control arm in the risk/hazard of MH engagement at any time during follow-up [Hazard Ratio: 1.09 (95% CI: 0.76-1.57), p=0.633.].
Post-baseline repeated measures of depression, anxiety and panic disorder scores were available in 118 and 102 patients of the control and intervention arms, respectively. Compared with the control arm, during the follow-up period, the intervention arm achieved lower scores in all three mental health measures. The mean depression score was significantly lower in the intervention arm than in the control arm, 9.4 (std: 6.2) vs. 11.1 (std: 6.5), p=0.010. Similarly, among participants reporting a traumatic event, the mean PTSD score was significantly lower in the intervention arm than in the control arm, 25.1 (std: 18.4) vs. 29.7 (std: 16.7), p=0.034. During follow-up, most substance use scores were similar between the two arms, but notably participants in the intervention arm had significantly lower use of cannabis than those in the control arm: 3.1 (std: 4.8) vs. 4.6 (std: 6.7), p=0.011. In addition, compared with the control arm, during follow-up, the intervention arm achieved significantly higher domain scores in psychological health (13.4 vs. 12.7, p=0.004), social relationships (13.3 vs. 12.1, p=0.003), and environment (14.4 vs. 13.6, p=0.004).
Qualitative data collected during the trial indicated that participants in the intervention arm (veteran peer coaching) vs. those in the control arm (no coaching) may not have perceived a need for mental health treatment because they were receiving coaching from the veteran peers. The data indicated that participants found these coaching calls therapeutic. The following themes emerged from the qualitative data: (1) encouragement from the veteran peer coaches, (2) therapeutic value to calls; (3) a sense that someone cared; (4) psychoeducation; (5) help with problem solving; (6) accountability; (7) helpful resources; (8) assessment questions elicited more insight; (9) delivery of intervention by phone was more convenient than driving for care; and (10) peers were less judgmental and more relatable than mental health clinicians.
This study evaluated the impact of rural culture on MH referral and engagement processes at rural CBOCs in VISN 16 & 21 and used this information to adapt and implement a Motivational Interviewing coaching intervention to improve access to mental healthcare for rural veterans.
This research also helped illuminate barriers to care and preferences for mental health services among rural veterans with mental health symptoms. Information from this project can be used to develop implementation toolkits for MH treatment engagement interventions for rural veterans. This project also filled a gap in the scientific literature about the effectiveness of peer motivational coaching for mental health treatment engagement among rural veterans with mental health symptoms who were not in treatment at baseline. We found that mental health treatment engagement was not superior among participants randomized to the peer motivational coaching intervention arm compared to controls. Nevertheless, we found that mental health treatment engagement after baseline assessment (by veteran peers) was higher than expected (~45%) in both groups. Importantly, we found greater mental health symptom improvement among participants in the peer motivational coaching intervention compared to controls and that in most cases these improvements were statistically significant. We also found that participants randomized to the peer motivational coaching intervention reported significant improvements in several quality of life domains compared to controls. Qualitative data collected from trial participants during the intervention reveled that veterans found the calls themselves to have therapeutic value, perhaps obviating the need for mental health treatment, thereby explaining these findings.
The study has already resulted in additional operational work in this area. The team has a one-year VA operations grant from the Office of Patient-Centered Care and Cultural Transformation (OPCC-CT) which extends the scope of our VA HSR&D CREATE-supported COACH study. Specifically, the OPCC-CT pilot funding allows us to test the VA Whole Health coaching model delivered by a mix of veteran peers and research staff for both urban and rural veterans to support self-management approaches to chronic disease. We have adapted the VA Whole Health coaching model to include the motivational coaching intervention currently being tested by veteran peer coaches for rural veterans to support engagement in mental health services through our CREATE study.
We also have a one-year grant from the VA Office of Mental Health services (OMHS) to pilot the use of the COACH motivational interviewing intervention by veteran peer specialists for veterans seen in primary care needing assistance with various behavioral interventions to support health and wellness.
Research Impact: This manuscript describes the process of implementing our veteran peer motivational coaching intervention to promote mental health care engagement among rural veterans who use VA CBOCs. It highlights the importance of tailoring the implementation of the evidence-based intervention to the needs and preferences of stakeholders in the rural communities served. Koenig, C.J., Abraham, T.H., Zamora, K.A., Hill, C., Kelly, P.A., Uddo, M., Hamilton, M., Pyne, J.M., and K.H. Seal. (2016). Pre-implementation strategies to adapt and implement a veteran peer coaching intervention to improve mental health treatment engagement among rural veterans. Journal of Rural Health 32(4):418-428. doi: 10.1111/jrh.12201
Research Impact: This manuscript qualitatively describes the diverse practices used by rural veteran to self-manage mental health symptoms when there are no services or barriers to engagement in those services. These self-management practices are largely informed by geographic and cultural contexts. Abraham, T., Koenig, C., Zamora, K., Hill, C., Uddo, M., Kelly, A., Hamilton, M., Curran, G., Pyne, J, and Seal KH. Situating Mental Health Work in Place: Qualitative data from interviews with veterans in Southeastern Louisiana and Northern California. Health & Place (Accepted).
Research Impact: These presentations leverage qualitative data demonstrating individualized definitions of what it means to be "engaged" in mental health care and how that differs from the VA institutional definition of mental health services engagement, a metric our group was instrumental in defining.
Zamora, K., Koenig, C., Abraham, T., Hill, C., Pyne, J., & Seal, KH. (March 31, 2016) The Diversity of Veteran Engagement in Mental Health Care. Society for Applied Anthropology (SfAA) Annual Meeting. Vancouver, B.C., Canada.
Zamora, K., Koenig, C., Abraham, T., Hill, C., Pyne, J., & Seal, KH. (September 12-13, 2016) A Qualitative Study of the Diversity of Veteran Engagement in Mental Health Care. State of The Art VA Rural Health Summit. Washington, D.C.
External Links for this Project
NIH ReporterGrant Number: I01HX001125-01
Dimensions for VADimensions for VA is a web-based tool available to VA staff that enables detailed searches of published research and research projects.
If you have VA-Intranet access, click here for more information vaww.hsrd.research.va.gov/dimensions/
VA staff not currently on the VA network can access Dimensions by registering for an account using their VA email address. Search Dimensions for this project
DRA: Mental, Cognitive and Behavioral Disorders, Health Systems
DRE: Treatment - Comparative Effectiveness
MeSH Terms: none