(PS6-97) Behavioral Activation-primary Care for Suicidal Ideation: A Secondary Analysis of a Randomized Controlled Trial Testing Behavioral Activation-primary Care for Depression in a Veteran Sample

Background. Given that nearly half of those who die by suicide will visit a primary care setting in the month before killing themselves (Luoma et al., 2002), embedded behavioral health providers who work in primary care (PC) should possess facility in behavioral interventions that are (1) effective at reducing suicidal ideation (SI), and (2) fast-acting. Behavioral activation (BA), which has already shown high efficacy for treating depression (Daughters et al., 2008; Ekers et al., 2011), is a promising candidate for meeting these criteria. Brief BA-Primary Care (BA-PC), which is delivered as two 30-minute appointments followed by two boosters, was designed to meet the needs of embedded behavioral health providers in fast-paced PC settings. In the clinical trial from which this analysis draws its data (a randomized controlled trial (RCT) which compared BA-PC to treatment as usual (TAU) for depression in a sample of 140 veterans), BA-PC was found to be as effective as TAU for reducing depressive symptoms, and was associated with stronger clinical retention and higher gains in quality of life than TAU (Funderburk et al., 2021).
Methods. This secondary data analysis focuses on a subset of participants (n = 42, 100% male) from the RCT described above who reported SI at baseline by endorsing item 9 of the Patient Health Questionnaire-9 (PHQ-9; Spitzer et al., 1999) before random assignment to either BA-PC (n = 25) or TAU (n = 17). SI was assessed again using the PHQ-9 at six, 12, and 24 weeks following treatment. SI remission was defined as a follow-up score of 0 on item 9 of the PHQ-9.
Results. Descriptively, we found that 57% of BA-PC participants reported SI remission compared to 50% of TAU participants at 12 weeks. At 24 weeks, 100% of BA-PC participants were in remission compared to 57% of TAU participants. Using logistic regression, we found no significant difference in SI remission between groups at 24 weeks (p = 0.14).
Limitations. Although this analysis did not yield statistically significant results, it was greatly limited by small sample size. Prior studies (Daughters et al., 2008; Ekers et al., 2011; Hemanny & de Oliveira, 2022) and our descriptive analysis together suggest that this line of research deserves further inquiry.