Assessing Behavioral Threat Avoidance in Pediatric Anxiety With Naturalistic Games

ABSTRACT Behavioral avoidance exacerbates anxiety symptoms. Capturing and identifying maladaptive avoidance patterns holds clinical utility. However, effective assessment tools which provide unique information on behavior beyond self‐report are needed. Here, we assess the replicability of an interactive game to distinguish between clinically anxious and healthy youth's behavioral avoidance. One hundred and fifteen youth (8–17 years old) completed the Yale Interactive Kinect Environment Software (YIKES) paradigm, whereby behavioral avoidance of aversive and non‐aversive faces is captured. Our primary, preregistered analyses used mixed ANOVA models to test differences in behavioral avoidance by diagnostic status, clinician‐reported avoidance, and self‐reported social anxiety symptoms. Secondary, non‐preregistered analyses pooled data across two published studies ( n = 189 youth). Primary analyses found no association between clinician‐rated anxiety symptoms and avoidance ( ps > 0.296). A secondary, pooled analysis with additional data indicated a positive association between self‐reported social anxiety symptom severity and avoidance of aversive faces ( r = 0.25, p < 0.001). A positive association also emerged between overall self‐reported anxiety symptoms and avoidance of aversive faces ( r = 0.18, p = 0.014). Here, we provide evidence that a naturalistic interactive game is effective for capturing avoidance patterns. While no associations with clinician‐rated measures emerged, behavioral avoidance was correlated with self‐reported anxiety. This appeared stronger for social anxiety compared to overall anxiety symptom severity. Our results suggest that self‐report may be more sensitive to capturing behavioral patterns of avoidance than clinician‐report, although replication and further work building on this paradigm is warranted.