(PS10-58) Exploring Whether a Primary Care Brief Mindfulness Treatment Can Open Doors to Behavioral Healthcare

Background: Behavioral healthcare engagement for Post-Traumatic Stress Disorder (PTSD) is low even within VA where high quality evidence-based psychotherapies (EBPs) are available (Ranney et al., 2023). Avoidance symptoms may contribute to low engagement (Sayer et al., 2009); mindfulness may mitigate avoidance through self-regulation and non-judgmental self-awareness (Kraemer et al., 2020). Therefore, mindfulness may support behavioral healthcare engagement in populations with PTSD. A Primary Care Brief Mindfulness Treatment (PCBMT) RCT showed that most participants engaged in post-study behavioral healthcare with no differences between PCBMT and psychoeducation (EDU) classes; PCBMT had higher PTSD symptom improvement than EDU (Possemato et al., 2022). Given that symptoms are strongly correlated with PTSD treatment utilization (Johnson & Possemato, 2019), symptom changes may influence behavioral healthcare following PCBMT.

Objective: Our hypothesis for this secondary data analysis of a RCT of PCBMT is that PTSD symptoms and mindfulness, specifically self-regulation and non-judgmental self-awareness, will predict behavioral healthcare.

Methods: 55 primary care Veterans with PTSD were randomized to PCBMT or EDU. A series of bivariate correlations was used to identify model predictors which fit our theoretical conceptualization but had low correlations with other model predictors. Due to overdispersion and high prevalence of zero counts, a Negative Binomial (NB2)-Logit hurdle model was the best fit to evaluate how post-treatment PTSD symptoms (PTSD Checklist-5, PCL-5), non-judgmental self-awareness (Five Factor Mindfulness Questionnaire-15, FFMQ-15 Observing subscale which measures noticing internal and external experiences), self-regulation (Self-Compassion Scale-Short Form, SCS-SF Over-Identification subscale which measures exaggerating negative aspects of personal experience), and randomization by intent to treat (ITT) predict behavioral healthcare appointments following PCBMT and EDU. Hurdle models combine a binary model and a count model to address excess zeros. 

Results: In the logit model component, PTSD symptoms, B(SE)=0.05(0.03), p=.049, predicted behavioral health appointments (yes/no); other predictors were not significant. In the negative binomial model component, PTSD symptoms, B(SE)=0.04(0.02), p=.047, observing, B(SE)=-0.19(0.09), p=.030, and over-identification, B(SE)=-1.02(0.33), p=.002 predicted more behavioral health appointments; randomization by ITT was not significant.

Conclusions: These results confirm previous research that PTSD symptoms are a strong predictor of behavioral healthcare. Contrary to expectations and prior research, high levels of observing internal and external experiences and self-regulation of personal experience negatively predicted behavioral healthcare engagement. In this case, low levels of these mindfulness facets may contribute to negative affect driving healthcare utilization . Alternatively, increases in mindfulness after PCBMT may have resulted in sufficient symptom relief, alleviating the need for more care. Future research should explore mechanisms for how mindfulness facets influence engagement in EBPs.