Increasing Accessibility to a Brief, Couple-Based Group Intervention for Veterans and their Intimate Partners: A Quality Improvement Project
+1Narayan B. Singh, PhD, ABPP & 4 others
May 19, 2025
Introduction
Veterans experience relationship difficulties including divorce (Cohan et al., 2005; Mouritsen & Rastogi, 2013) and intimate partner violence (McGinn et al., 2017) at higher rates compared to non-Veterans. Veteran intimate relationships also present risk and protective factors for Veteran mental health service utilization (Logan et al., 2012), morbidity (Bolkan et al., 2013; Martin et al., 2013), and mental health symptoms (Batten et al., 2009; Madsen et al., 2017; Monson et al., 2012). The Veterans Health Administration (VA) mandated the provision of marital and family counseling to all interested Veterans starting in 2019 (Veterans Health Administration, 2019). Implementing couple therapy demands adequate resources, which are limited in many VA clinics (Mouritsen & Rastogi, 2013). Instead, offering couple group therapy may allow clinics to meet this mandate, serve multiple couples simultaneously, and maximize resources
Communication skills training has been linked to improved relationships in non-Veteran couples (Perlick et al., 2011; Wittenborn et al., 2019). Several studies have shown advantageous effects of skills-based couples group interventions with Veterans and their intimate partners as well (Croake & Kelly, 2002; Fischer et al., 2013; Fredman et al., 2020; Sherman et al., 2011; Sherman et al., 2015). For instance, a 20-week psychoeducational couples group therapy contributed to improvements in self-reported intimate relationship satisfaction in a sample of Veterans with a diagnosis of schizophrenia or a bipolar disorder and their partners (Croake & Kelly, 2002). Given the small number of studies using Veteran couples, however, more research is warranted to evaluate the effectiveness of communication skills training on relationship satisfaction with Veteran couples in a brief group therapy format.
This paper presents a program evaluation of a 5-week couple communication skills group for Veterans and their intimate partners. We aimed to evaluate whether: 1) brief couple group therapy would significantly improve couples’ relationship satisfaction and, 2) the group would also reduce anxiety and depression symptoms due to existing preliminary evidence of such an effect (Denton et al., 2012; Durães et al., 2020; Knopp et al., 2022).
Method
Procedure
This program evaluation was given IRB exemption as a quality improvement project. Groups were conducted in outpatient mental health clinics at two VA hospitals in the Pacific Northwest between 2014-2019. Clinicians with graduate degrees in mental health facilitated the groups. The protocol consisted of five, 75- to 90-minute weekly sessions (see Table 1). Sessions 1-4 consisted of didactics, role-plays, group discussions, and homework. Sessions 2-4 included a review of the previous week’s homework. The fifth and final session included a review and discussion of all previously presented skills. Participants completed measures on the first and last sessions and were given the option to provide written qualitative feedback during the last session.
Table 1
5-Week Couple Communication Skills Group Protocol Overview
| Session # | Topics Covered | Homework Assigned |
|---|---|---|
| 1 | Introduction & active listening in communication | 1-hour practice of using active listening skills with prompts |
| 2 | Non-violence and assertiveness in communication | 1-hour practice of using non-violent communication skills |
| 3 | Fighting fair, repairs, and time outs in conflict resolution | Completion of time out plan and practicing one time out |
| 4 | Building intimacy | Reading article on building intimacy and using prompts for intimacy-building conversation |
| 5 | Review of all prior weeks’ material & group wrap up | Continued use of skills reviewed in previous weeks |
Recruitment
Veterans and their partners were recruited from outpatient mental health and post-traumatic stress disorder clinics at each VA. Clinic therapists and prescribing providers referred interested Veterans to complete brief telephone screenings with the group facilitators. Exclusionary criteria for participation included current intimate partner violence, active psychosis, acute suicidality, or substance use that would interfere with group engagement.
Participants
Fifty-eight individuals (29 couples), 35 of whom were Veterans (60.3%) participated in one of five rounds of the group. Due to the program evaluation nature of this project, demographic variables aside from Veteran status were not collected. Anecdotally, facilitators observed that half (n = 29) of the sample were female-identifying in heterosexual relationships with male-identifying partners. Among the Veterans, six were female-identifying (17.1% of Veterans).
Measures
Primary Outcome Measure
The 4-item version of the Couples Satisfaction Inventory (CSI-4; Funk & Rogge, 2007) was used to assess relationship satisfaction over the last week, the degree of happiness in the relationship, warmth, and comfort of the relationship, how rewarding the relationship is, and overall relationship satisfaction. The first item is rated on a 6-point Likert scale (0 = Extremely unhappy to 6 = Perfect) and the subsequent three items are rated on a 5-point Likert scale (0 = Not at all to 5 = Completely). Items are summed for total scores from 0 to 21, with higher scores indicating greater relationship satisfaction. Scores below 13.5 suggest relationship distress (Funk & Rogge, 2007). Internal consistency for the present sample at baseline was excellent (α = .94).
Secondary Outcome Measures
The Patient Health Questionnaire-2 (PHQ-2; Kroenke et al., 2003) was used to assess self-reported symptoms of depression (i.e., anhedonia and depressed mood) over the preceding two weeks rated on a 4-point Likert scale (0 = Not at all to 3 = Nearly every day). Sum scores range from 0 to 6, with higher scores indicating greater depressive symptoms. Scores ≥ 3 suggest the likelihood of a depressive disorder (Löwe et al., 2005). Internal consistency for the present sample at baseline was excellent (α = .90).
Four items (#1, #2, #4, and #5) from the 7-item Generalized Anxiety Disorder scale (GAD-7; Spitzer et al., 2006) were used to assess self-reported anxiety based on their applicability to multiple anxiety disorders. The items in this modified GAD-4 assess anxiety over the preceding two weeks, rated on a 4-point Likert scale (0 = Not at all to 3 = Nearly every day). Sum scores range from 0 to 12, with higher scores indicating higher anxiety symptom severity. This modified GAD-4 showed good internal consistency for the present sample at baseline (α = .85).
Qualitative Feedback
Three open-ended prompts were generated to gather participants’ feedback about the group, including what participants found helpful and unhelpful about the group, whether they liked the group session length, and if they would be interested in attending a 12-week couples’ group.
Data Analysis
Multilevel models (MLM) with maximum likelihood estimation were conducted in R version 4.1.2 (The R Foundation, 2021). MLM permits fixed effects calculation while accounting for random effects and nested data (i.e., participants within dyads; Atkins, 2005). Three models were conducted predicting pre- to post-treatment changes in the level one variables: couple satisfaction, depression, and anxiety. The level two variables of individual participant gender and Veteran status were omitted to maximize power. Omitting these variables was permitted due to finding comparable results with and without inclusion, and there were no significant effects or model convergence failures when omitting them. The level three variable was each participants’ couple dyad. Intraclass correlation coefficients (ICCs) were calculated using tests of two-way mixed effects based on a single measurement and absolute agreement (see Koo & Li, 2016 for a review) using Package ‘irr’ (Gamer et al., 2019). ICCs for all models were moderate, ranging from .53 to .60, indicating sufficient variance both between- and within-participants to warrant tests of fixed and random effects. Regarding qualitative feedback, participants’ responses were grouped into common themes and theme frequencies were then tallied.
Results
The model predicting couple satisfaction showed a significant intercept (b = 11.84, SE = 0.63, p < .001) and fixed effect of time on participants’ couple satisfaction (b = 2.08, SE = 0.46, p < .001). On average, participants started treatment experiencing relationship distress (below the 13.5 cutoff mentioned above) and over the course of treatment, participants’ relationship satisfaction increased. The model predicting depression showed a significant intercept (b = 1.91, SE = 0.21, p < .001, t(112) = 9.09) and a fixed effect of time on depression (b = -.51, SE = 0.20, p = .011, t(112) = -2.58). On average, participants did not report high levels of depression at baseline, however, the symptoms that were reported decreased over the course of treatment. The model predicting anxiety showed a significant intercept (b = 5.00, SE = .40, p < .001, t(112) = 12.65) and slope (b = -.93, SE = .34, p = .007, t(112) = -2.73). On average, participants did not report high levels of anxiety at baseline and, similar to depression, the symptoms that were reported decreased over the course of treatment.
To summarize the quantitative results, participants’ average relationship satisfaction was in the distressed range at treatment outset, rising above this threshold by treatment end. Although participants’ average depression and anxiety symptoms at baseline were not high, they nonetheless decreased over the course of treatment. Thus, the treatment appeared to impact its primary target of relationship satisfaction, while also providing additional benefits for other mental health symptoms. Means and standard deviations for each outcome variable are presented in Table 2.
Table 2
Sample Means and Standard Deviations at Pre- and Post-Treatment
| Measure | Mean (Standard Deviation) | |
|---|---|---|
| Pre | Post | |
| CSI-4 | 11.84 (4.27) | 13.91 (4.04) |
| PHQ-2 | 1.91 (1.81) | 1.40 (1.39) |
| GAD-4 | 5 (3.20) | 4.07 (2.86) |
Note. N = 58; CSI-4 = Couples Satisfaction Inventory 4-item; PHQ-2 = Patient Health Questionnaire 2-item; GAD-4 = Generalized Anxiety Disorder 4-item modified.
Forty-three (74.1%) participants provided written feedback, which was summarized into common themes. Ten participants indicated the group was helpful and did not provide additional details. The following are the frequencies of specific aspects of treatment that participants found helpful: Learning skills (n = 12), hearing about other couples’ experiences (n = 12), homework (n = 7), handouts and psychoeducation (n = 5), and discussions (n = 4). Three participants indicated wanting more of a challenge and to have learned additional skills. Twenty-three participants liked the group’s length, and 14 participants indicated a preference for the group to be longer by one or several sessions. Finally, 15 participants indicated interest and 12 indicated disinterest in attending a 12-week couples’ group. Scheduling difficulties and limited availability were the reasons listed for disinterest in attending a longer group.
Discussion
To meet the growing need for brief, accessible interventions to treat relationship distress in Veterans and their intimate partners, we conducted a program evaluation of a 5-week couple communication skills group at two VA hospitals. Results showed that participants’ average relationship satisfaction increased pre- to post-group and no longer fell into the range indicating relationship distress. Given that most couples present with complaints about communication difficulties (Bradbury & Bodenmann, 2020), specifically targeting effective communication skills within a group format appears to be an efficient treatment target.
Anxiety and depression symptoms were not a focus of group materials or homework, yet this couple skills group was associated with decreased symptoms for both, consistent with prior research on couples-based interventions (Denton et al., 2012; Durães et al., 2020; Knopp et al., 2022). Qualitative feedback was also primarily positive, suggesting that the group was helpful. These results suggest that a group-based couple format may be a reasonable option for assisting Veteran couples with relationship concerns, with the ancillary benefit of improving anxiety and depression symptoms.
There are several limitations to this project including a small sample size, no comparison group, no follow-up assessment, and the sample appearing to have relatively minor relationship distress at baseline. Changes in couple satisfaction may represent regression to the mean or may not have been maintained over time. Changes in anxiety and depression symptoms may be due to confounding factors that were either not measured or unrelated to the group. Data regarding participant drop-out, treatment tolerance, and participant demographic variables were not collected. Therefore, the current results have limited generalizability. Future research should utilize larger sample sizes with diverse sample demographics, follow-up assessments, and couples with a wider range of distress.
Conclusion
Preliminary data suggests that a brief, skills-focused couple-based group may be a modestly successful offering in a larger spectrum of couple and family services within the VA. Existing resources within VA clinics are likely sufficient for implementing a couple communication skills training group. This may provide immediate relief to Veteran couples with relationship distress. Thus, this couple communication skills group provides an avenue for VA mental health clinics to align with the 2019 VA initiative by providing Veterans with relationship-specific interventions.