(PS5-C58) Predictors of Provider Turnover Following Participation in a Trauma-informed Community-based Learning Collaborative

A significant barrier to the sustainability of evidence-based practice (EBP) implementation is the high level of provider turnover in community-based mental health organizations. Previous research suggests that provider burnout and low pay predicts increased turnover (Aarons et al., 2009), while greater supervision support and staff incentives decreases turnover (Lau & Brookman-Frazee, 2015). Furthering our understanding of the individual and agency level factors that influence turnover can help tailor EBP implementation efforts to reduce these effects. This project extends this literature by examining whether pre-training provider EBP attitudes, feelings of burnout, and agency implementation climate predict turnover five years post-training. Participants (N = 68) were previously trained in a SAMHSA-funded trauma EBP training using the Community Based Learning Collaborative (CBLC) model (Hanson et al., 2016). A CBLC trains direct care providers, supervisors, and senior leadership from multiple organizations that provide psychotherapy (i.e., clinical track) and/or case management (i.e., broker track) services. This study included 47 clinical track, 15 broker track, and 6 senior leader participants. Participants were on average 40.8 years old (SD =10.5), 85.3% female, 64.7% Hispanic, and 73.5% held a Master's degree or higher. Training consisted of two, two-day, in-person learning sessions and bi-weekly consultation calls with a trainer over one year.  Before the first learning session, all participants completed the Implementation Climate Scale (ICS; Ehrhart et al. 2014) and two burnout questions (i.e., frequency of feeling 1. emotional exhaustion and 2. depersonalization at work; West et al., 2012). Clinical track participants also completed the Evidence Based Practices Attitudes Scale (EBPAS; Aarons et al. 2012). Turnover data was collected five years post-training through contact with the participant or their agency. Generalized Estimating Equations (GEE; Hanely et al., 2003) were used to examine predictors of turnover while accounting for nesting within agency. Analyses were conducted one predictor at a time for each ICS and EBPAS subscales and the two burnout items; significant individual predictors were then examined together. Forty-eight (70.6%) participants turned over. Contrary to study hypotheses, greater emotional exhaustion at work was associated with decreased odds of turnover (B=-0.35, p< 0.01, OR=0.71) and higher agency educational support for EBPs was associated with increased odds of turnover (B=0.71, p< 0.001, OR=2.04). Therapists those who viewed EBPs as more appealing were more likely to turnover (B=0.51, p< 0.022, OR=1.67). All other subscales were not significant. Follow up analyses of significant findings from the full sample indicated that turnover was associated with emotional exhaustion at work (B=-0.28, p< 0.05, OR=0.76) and agency educational support (B=0.65, p< 0.01, OR=1.91). Odds ratios indicated that a one-point increase in emotional exhaustion and agency educational support for EBPs were respectively associated with a 32% and 91% increase in the odds of turnover. Further analyses will explore these unexpected findings and implications for EBP trainings will be discussed.