Can answering questions change health behaviours? A systematic review and meta-analysis

Abstract

Questioning behaviour and/or cognitions (question-behaviour effect; QBE) may change health behaviours. The present review aims to consolidate QBE studies regardless of behaviour domain and design. Three databases were searched for papers (1980-April 2014). Included papers measured cognitions and/or behaviour without another intervention vs an unassessed control. Effect sizes were calculated using meta-analysis. Meta-regressions assessed moderator and effect size associations. 96 studies were included from 381 screened papers. Results supported a small significant QBE (g=.14, p<.001). Bias was a significant moderator, low risk of bias studies had a smaller effect (g=.07) than unclear/high risk of bias (g=.22). QBE was found to increase healthy behaviours (g= .16, p<.001), and reduce some health risk behaviours (g= .08, p= .04). This review shows QBE as a somewhat effective intervention to increase healthy behaviours. Small effects in low risk of bias studies warrant caution in determining its effectiveness. The limited evidence in risk behaviours suggests it is generally ineffective in reducing these behaviours. Further high quality studies and focus on risk behaviours are needed.