(PS12-A13) The Role of Positive Affect in CBT Skills Training

According to the broaden and build theory of positive emotions (Fredrickson, 2004), while negative emotions narrow one’s thought-action repertoires to navigate dangerous or uncomfortable situations, positive emotions broaden these thought-action repertoires, allowing for increased approach behaviors such as creativity, exploration, and building of skills. Positive emotion and the associated behavioral urges may be advantageous in psychotherapy, which asks clients to explore and develop new coping skills. As such, we would expect those with higher positive affect (PA) to have higher outcome expectancies and be more willing to engage with the intervention. Additionally, given the positive associations of client expectancy (Beasley et al., 2017; Constantino et al., 2012) and engagement (LeBeau et al., 2013) with treatment outcomes, we might anticipate those with higher PA to also achieve greater improvement on depressive and anxiety symptoms. In the current study, we examined how a client’s baseline PA related to several process and outcome variables of a CBT skills-based intervention, including treatment expectancy at baseline, homework completion, and outcomes of depression and anxiety.

Participants (n = 291) were randomized to either a cognitive (n = 75), behavioral (n = 75), interpersonal (n = 68), or mindfulness (n = 73) module, which took place over the course of 3 weeks. Each module consisted of three 5-minute educational videos (one per week) on a specific coping skill. In addition, participants were expected to practice by completing 9 module-specific worksheets over the course of the 3 weeks. Assessments of client expectancy (CEQ; Devilly & Borkovec, 2000) and PA (PANAS; Watson et al., 1988) were taken as baseline. Homework completion was defined as worksheets completed by the end of the intervention (0 = no worksheets, 1 = at least 1 worksheet). Depression (QIDS-SR; Rush et al., 2003) and anxiety (GAD-7; Spitzer et al., 2006) symptoms were assessed weekly.

Multiple regression analyses were conducted for tests of the association of PA with expectancy and homework completion. For each of these, baseline depression scores were included as a moderator to examine the role of baseline differences in depression on the relationship. Repeated measures linear regression models were conducted for depression and anxiety symptoms regressed on time (in weeks), PA, and their interaction.

The results suggested that baseline PA is associated with greater expectations of treatment success, t(283) = 3.76, p < .001. However, this is qualified by its interaction with baseline depression scores, t = 2.30, p = .022. Analysis of simple slopes suggests that PA relates to expectations of treatment success only for those with moderate depressive symptoms or greater. However, the relationship between PA and homework completion was not statistically significant, z = 1.94, p = .052. In a planned analysis of simple slopes, PA tended to be related to homework completion only for those with higher depressive symptoms. Finally, despite significant or nearly significant effects of PA on treatment expectations and engagement, PA was not predictive of change in symptoms of depression, t(618) = 1.03, p = .303, nor anxiety, t(612) = 0.53, p = .596.