Reviews of Empirical Studies on Attention Placebo for Anxiety or Phobia Related Problems

Abstract

The term "attention placebo" has been used in an imprecise and often seemingly thoughtless way for decades, and it is often confusing to encounter "attention placebo group" in outcome studies. The purpose of this study was to investigate attention placebo empirically, focusing on anxiety or phobia related problems. The two aims were (1) to describe and summarize how researchers define attention placebo in empirical studies and the intervention procedures involved in these studies; and (2) to explore the possibility of estimating the magnitude of the attention placebo effects on anxiety or phobia related problems.A thorough literature search was carried out with the keywords of attention placebo synonyms and the resulting list of 1304 articles was screened for unique empirical articles on anxiety or phobia related problems. A total of 545 articles were identified as attention placebo empirical studies, and 83 involved anxiety or phobia related problems. These 83 constituent articles were coded for the purpose of qualitative overview, and 63 articles with sufficient data were included in the meta-analysis of attention placebo effects.The results of the qualitative overview of attention placebo on anxiety or phobia related problems showed that most articles did not discuss explicitly any rationale for using an attention placebo group, but simply included one as a comparison group. The arrangements for attention placebo interventions were diverse and often reflected little concern for emulating the nonspecific features of the treatment(s) of interest. Two meta-analyses were performed based on the comparison of attention placebo and a no-treatment control, and attention placebo and the verum. The overall estimated effect of attention placebo is Hedges's g=0.32, a small effect, significantly different from the effect size of the other control groups. Compared to the verum, the effects of attention placebo are not consistent in magnitude but were nearly all smaller than the effects of the verum treatments. Meta-regression and subgroup analyses indicated that estimating attention placebo effects is a sensible and meaningful research activity.In conclusion, attention placebo is not so much about attention, and it is possible to estimate the effects of attention placebo.