Cochrane review and meta-analysis of trials of action and/or coping planning for health behaviour change

Abstract

Background: Action Planning and Coping Planning (Implementation Intentions) are well-tested techniques for changing health behaviour. Aims: conduct a Cochrane review of trials of interventions involving action and/or coping planning across health behaviours in adults, and identify effect modifiers (type of plan, outcome measure, follow-up length). Methods: We searched Cochrane, MEDLINE, Embase, CINAHL, PsycINFO, and trial registers up to January 2014. We included randomized trials of interventions involving action and/or coping planning reporting health behaviour outcomes. Random effects meta-analyses were conducted on trials with sufficient data and which delivered action planning only, coping planning only, or both; objectively-assessed or self-reported behaviour; and follow-up length (<1 month, 1-6 months, 7+months). Findings: We included 119 trials involving 34,252 participants. Meta-analysis results (96 trials); for objectively-measured continuous outcomes were: Action Planning: <1month (4 trials, Cohen’s d=0.37[95%CI=0.22-0.52]), 1-6months (8 trials, d=0.29[95%CI=0.17-0.42]), 7+ months (2 trials, d=0.14[95%CI=-0.04-0.32]). Coping planning: (no trials). Action+Coping planning: <1month (no trials), 1-6months (3 trials, d=0.28[95%CI=0.02-0.54)], 7+months (no trials). Meta-analysis results for self-reported continuous outcomes: Action Planning: <1month (15 trials, d=0.48[95%CI=0.25-0.71]), 1-6months (21 trials, d=0.19[95%CI=0.03-0.35]), 7+months (2 trials, d=0.19[95%CI=0.04-0.34]). Coping planning <1month (3 trials, d=0.67[95%CI=0.37-0.96]), 1-6months (8 trials, d=0.59[95%CI=0.18-0.99]), 7+months (no trials). Action+Coping planning: <1month (7 trials, d=0.17[95%CI=0.05-0.29]), 1-6months (18 trials, d: 0.42[95%CI=0.22-0.62)], 7+months (3 trials, d=0.77[95%CI=-0.40-1.94)). Discussion: In this largest review to date, spanning all health behaviours, Action and/or Coping Planning had a small-to-medium effect on objectively-assessed behaviour, and medium effect on self-reported behaviour. There remains considerable unexplained heterogeneity and a need for more trials using objective outcomes and longer-term follow-up.