Systematic reviews of qualitative studies can structure qualitative comparative analysis-based synthesis of intervention effectiveness

Tags: Oral
Melendez-Torres G1, Sutcliffe K2, Richardson M2, Burchett H3, Thomas J2
1University of Warwick, UK, 2University College London, Institute of Education, UK, 3London School of Hygiene and Tropical Medicine, UK

Background and objectives: Qualitative comparative analysis (QCA) is useful for the synthesis of complex interventions, particularly when the goal is not to render a pooled estimate but to identify configurations of conditions, or participant and intervention characteristics that form pathways to an outcome. QCA originally relied on theoretically guided condition selection, which may be less helpful when rich data on interventions, e.g. from qualitative studies, is available. We report how a systematic review of service user views structured condition selection, model construction and interpretation to identify pathways to effectiveness in weight management programmes (WMPs).

Methods and results: We updated a previous views synthesis and a systematic review of interventions. We identified 38 key themes in the views synthesis via thematic synthesis and translated themes into intervention conditions. We coded the ten most and ten least effective of 40 interventions as to the presence or absence of these conditions. Because of the number of conditions we coded, we relied on the views synthesis to identify overarching processes that users suggested were associated with WMP effectiveness, and selected conditions to develop three QCA models on the basis that they addressed these processes. We then checked models for contradictory configurations, i.e. where a combination of conditions included both most effective and least effective interventions. Whilst two of our models were consistent, one included contradictory configurations. We used the views synthesis to develop additional lines of enquiry and to bound our enquiry (prevent ‘data dredging’) by acknowledging when inferences were unsupported by the views synthesis. Finally, findings from the views synthesis contextualised minimised solutions for pathways to effectiveness.

Discussion: Because QCA is an abductive approach, it requires ‘theorising’ and understanding of intervention processes to yield a meaningful solution. We have demonstrated here how the findings of views syntheses can be used to structure, but also to discipline and bound, analyses of pathways to effectiveness in interventions.