Synthesis of quantitative and qualitative evidence in systematic reviews: an example from public health

Article type
Authors
Campbell F, Goyder L, Johnson M, Messina J, Guillaume L, Campbell F
Abstract
Background: The issue of combining data from different traditions is of growing importance. This is particularly apparent in systematic reviews of public health interventions where an understanding of attitudes and behaviours are of central importance. The need for informed policy from a wider evidence base is clear and systematic reviews of both quantitative and qualitative studies are required (NICE 2006). While there is a strong body of research on methods for reviewing quantitative and qualitative evidence, methods to enable the synthesis of the two remain in their infancy (Dixon-Woods et al 2004). Objectives: This review of interventions to reduce excessive weight gain in pregnancy comprised both a review of quantitative and qualitative primary research. We sought to explore methods to synthesise the results of both. This abstract is a presentation of one of those methods. Methods: The qualitative review explored factors that influence maternal weight gain in pregnancy. Interventions studies (randomised and non-randomised) were subject to narrative analysis and where appropriate meta-analysis was performed. The results of the quantitative and qualitative analysis were then synthesised using thematic analysis. Results: The synthesis of the quantitative and qualitative reviews was presented in a matrix and enabled a greater understanding of the mixed pattern of effect seen in the intervention studies. Important factors that influence maternal weight gain were not addressed by the interventions such as the contradictions between lay health beliefs and healthy eating messages. Conclusions: The synthesis of the quantitative and qualitative evidence allowed greater understanding of the issues that undermine the effectiveness of interventions to reduce excessive weight gain. Shortcomings of the approach included assumptions about external validity between populations in both the quantitative and qualitative studies and a lack of transparency in the creation of themes.