Improving health providers smoking cessation care in pregnancy: A systematic review and meta-analysis

Introduction: Health providers are lacking in their provision of smoking cessation care during pregnancy. The aim of this study was to systematically review all available global studies on the effectiveness of interventions in improving health providers’ provision of smoking cessation care during pregnancy. Methods: Five databases were searched, Inclusion criteria included all intervention study types. Two reviewers screened abstracts and full texts independently. Interventions were characterized according to the Effective Practice Of Care taxonomy. Random-effects meta-analyses examined intervention effects on smoking cessation care components based on the 5As. Estimates were number of participants reporting each outcome, or mean score, transformed into Cohen's d. Crude meta-regressions, and meta-analysis subgrouping, were performed to examine whether intervention effects for ‘Ask’ ‘Advise’ and ‘Assist’ differed by intervention components. Results: Of 3165 manuscripts, 16 fulfilled inclusion criteria. Pooled analysis showed significant small to large intervention effects on the different care components (Cohen's d ranging from 0.47 for ‘Ask’ (95%CI 0.13–0.81) to 1.12 (95%CI 0.45–1.79) for ‘Setting a quit date’). Crude meta-regression suggested that for ‘Ask’ having a theoretical basis may improve effectiveness (Cohen's d difference 0.62, 95% CI 0.12–1.1). Subgrouping the meta-analysis suggested that audit and feedback possibly increases intervention effectiveness for ‘Advise’ and ‘Assist’. Conclusion: Interventions designed to improve provision of smoking cessation care during pregnancy show a small increase in care components. Studies vary substantially in design, intervention components, and outcome measurement, impacting ability to synthesize available data. Audit and feedback and enhancing intervention design by using behaviour change theories may improve effectiveness. Registration: PROSPERO CRD42016030143.