(PS3-C63) Multi-session Interpretation Bias Modification for Youth in the School Setting: A Pilot Randomized Controlled Trial

Cognitive bias modification (CBM) has received attention and criticism for its ability (or lack thereof) to modify cognitive bias and internalizing symptoms. Of the various cognitive biases (attention, interpretation, memory bias), interpretation bias – or the threatening or negative appraisal of ambiguous information (Beard, 2011) – appears to be a ubiquitous phenomenon in youth internalizing, or anxiety and depression, problems. There is some support that CBM for interpretation bias (CBM-I) reduces both bias and internalizing symptoms in youth and adults. However, most CBM-I studies have been conducted in laboratory settings, test one or a few training sessions, and do not examine whether effects on symptoms are maintained over time.

In this pilot randomized controlled trial, a four-week 12-session CBM-I intervention (n=22) was compared to an interpretation control condition (ICC, n=22) in youth (N=44) in the school setting.

Specifically, the current study aimed to test the effects of CBM-I on 1) performance-based and self-reported interpretation bias, and 2) youth- and caregiver-reported anxiety and depressive symptoms. Participants included 44 youth ages 10 to 17 (M = 12.33, SD = 1.69; 72.70% male, 38.6% racial/ethnic minority). Participants completed the Word-Sentence Association Paradigm (WSAP; Beard & Amir), a performance-based measure of interpretation bias; and the Children’s Automatic Thoughts Scale (CATS; Schniering & Rapee, 2002), a self-reported measure of interpretation bias. Participants and their parents completed the Screen for Child Anxiety Related Disorders (SCARED; Birmaher et al., 1999).

There was a significant group (CBM-I, ICC) x time interaction (F(2, 120.49) = 10.87, p< .001). The CBM-I group evidenced a significantly greater slope change than the ICC group from pre-to-mid intervention (β = .32, SE = .08, p< .001) and from pre-to-post-intervention (β = .31, SE = .08, p< .001). For the CATS, there was also a significant group x time interaction (F(2, 121.37) = 3.46, p=.04). The CBM-I group demonstrated significantly greater slope change than the ICC group from pre-to-mid-intervention (β = -8.50, SE = 3.37, p = .01). For the SCARED-C, there was a significant group x time interaction (F(2, 118.42) = 4.71, p = .01). The CBM-I group demonstrated significantly greater slope change than the ICC group from pre-to-mid intervention (β = -4.36, SE = 2.02, p = .03) and pre-to-post intervention (β = -6.28, SE = 2.14, p = .004). For the SCARED-P, there was a significant group x time interaction (F(2, 120.92) = 3.35, p = .04). The CBM-I group demonstrated significantly greater slope change than the ICC group only from pre-to-post intervention (β = -5.28, SE = 2.05, p = .01). Positive outcomes on self-reported interpretation bias and anxiety were maintained at two-month follow-up. These findings provide preliminary support for continued tests of multi-session CBM-I in youth, particularly those with elevated symptoms, and in real-world settings like schools.