No evidence for attention bias towards threat in clinical anxiety: a meta-analysis of baseline bias in attention bias modification RCTs

No evidence for attention bias towards threat in clinical anxiety: a meta-analysis of baseline bias in attention bias modification RCTs. Anne-Wil Kruijt, Sam Parsons, Elaine Fox. Pre-print 10.31234/osf.io/rfjup

Abstract

Background: Considerable effort and funding are spent on developing Attention Bias Modification (ABM) as a treatment for anxiety disorders. ABM is theorized to exert therapeutic effects through reduction of an increased attentional bias towards threat. Yet, the available meta-analytical evidence for the common assertion that clinical anxiety is characterised by this treat-related attentional bias is thin: the largest meta-analysis to date included n=337 clinically anxious individuals. We propose that baseline measurements in clinical ABM RCTs constitute a hitherto not assessed additional body of data on magnitude of biased attention in clinically anxious samples.

Method: We meta-analysed baseline dot-probe assessed bias for 1005 clinically anxious individuals enrolled in ABM RCTs.

Results: REML analysis indicated no evidence that mean bias index (BI) differs from zero (k= 13, n= 1005, mean BI = 1.8 ms, SE = 1.26 ms, p = .144, 95% CI [-0.6 - 4.3]. Additional Bayes factor analyses also supported the traditional point-nil hypothesis (BF10 = .23), whereas additional interval-based analysis indicated it unlikely that mean bias in clinical anxiety extends beyond the 0 to 5 ms interval. 

Discussion: We discuss our findings with respect to strengths (larger samples, possible bypassing of publication bias), limitations (lack of control comparison, repurposing data), and theoretical and practical context. We suggest that it may be prudent to no longer classify anxious samples as being characterized by bias.

Conclusion: Clinically anxious individuals enrolled in RCTs for Attention Bias Modification are not characterized by attentional bias towards threat at baseline.