Challenges faced and lessons learned from doing overviews of systematic reviews of complex interventions

Article type
Authors
Schoonees A1, Durao S2
1Centre for Evidence-based Health Care, Faculty of Medicine and Health Sciences, Department of Global Health, Division of Epidemiology and Biostatistics, Stellenbosch University
2Cochrane South Africa, South African Medical Research Council
Abstract
Background: We are conducting an overview of systematic reviews of interventions to improve food security in low- and middle income countries (LMICs). For overviews covering broad topics and complex interventions, selecting eligible studies, extracting useful data, and synthesising and presenting results in a meaningful way can be challenging.

Objectives: To highlight methodological challenges faced when conducting overviews of systematic reviews of complex interventions; and to suggest approaches to address these challenges.

Methods: We describe the challenges we encountered when selecting eligible systematic reviews, extracting data and presenting the findings for our overview. Furthermore, we asked colleagues who had done overviews before for guidance on ways to overcome the identified challenges. We also searched for published examples of overviews of systematic reviews of complex interventions. We drew on these sources to identify potential approaches to address the challenges we faced.

Results: Table 1 details the challenges encountered regarding study selection, data extraction, quality assessment, and summarising and presenting results of our overview. Potential approaches to address these challenges were drawn from input from three colleagues; one paper on the risk of bias in overviews; one systematic review, and five overviews of systematic reviews. We also drew on tools such as AMSTAR, ROBIS, SUPPORT summary checklist, and critical appraisal worksheets for intervention questions.

Conclusions: Although some of the challenges of doing overviews of systematic reviews may be unique to the specific overview question, many may have been encountered by other researchers. It is valuable to share experiences and learning to inform future approaches.