Imagery interventions in health behaviour: a meta-analysis

Abstract

Background: Imagery-based interventions represent an inexpensive, potentially effective technique for changing health behaviour yet have shown considerable variability in effects across studies. The present review provided a quantitative synthesis of the effectiveness of mental imagery interventions in health behaviour and tested key moderator effects. Methods: A systematic database search for studies adopting imagery interventions in health behaviour and related outcomes was conducted with additional manual searches and direct author contact for unpublished studies. Data were extracted for imagery intervention effects on behavioural, psychological, and physiological outcomes, and for candidate moderators. Findings: Twenty-six studies of mental imagery intervention effects comprising 33 independent data sets met eligibility criteria for review inclusion. Mental imagery interventions led to non-trivial, small averaged corrected effect sizes on post-intervention behaviour (d+ = 0.23), intention (d+ = 0.19), perceived control (d+ = 0.08), and attitude (d+ = 0.12), and a small-to-medium sized effect of imagery interventions on post-intervention physiological measures (d+ = 0.29). The substantive heterogeneity in the effects meant that a search for moderators was warranted. Moderator analyses indicated stronger effects for imagery interventions on health behaviours in studies on older samples, when detailed instructions were provided, in studies with higher methodological quality scores, and in studies of longer duration. Effect sizes for imagery on behavioural and physiological outcomes were larger than effects on psychological outcomes. Discussion: Results support effects of mental imagery interventions on health behaviours, identifies the conditions where they may be more effective, and points to how future imagery interventions might be optimized.