Article type
Abstract
Background: The Healthy Life Trajectories Initiative (HeLTI) will establish intervention research cohorts in Canada, China, India and South Africa to inform global policy and practice regarding interventions to reduce childhood obesity rates. We developed a method to prioritise the choice of interventions to be delivered specifically during pregnancy.
Objectives: To describe a systematic priority-setting method to identify pregnancy-related interventions for research cohorts.
Methods: Following identification of relevant systematic reviews (SR), two reviewers, independently and in duplicate, extracted data on publication year and included studies. A matrix of all included studies from the SRs identified the extent of overlap between SRs. Where more than two-thirds of studies overlapped, the most recent, high-quality SR, as evaluated using ROBIS, was selected for inclusion. For each pre-specified outcome, we developed a GRADE-based effectiveness matrix incorporating effect size and study quality to rank interventions as: 1) beneficial or harmful; 2) possibly beneficial or possibly harmful; 3) no effect; 4) possibly no effect; or, 5) uncertain effect. For interventions ranked as 1) or 2), additional data regarding study setting, participants, intervention composition, dosing, frequency, duration, feasibility, implementation co-factors and cost were extracted. We engaged with HeLTI stakeholders to examine common findings and potential reasons for differential effects by setting.
Results: The included studies' matrix identified 12 SRs in which there was more than two-thirds overlap, leaving 13 recent, comprehensive, high-quality SRs, for disaggregation. The effectiveness matrix identified four beneficial interventions, one harmful intervention and one possibly beneficial intervention.
Conclusions: Aggregated information from SRs, and specifically meta-analyses, often collapse or do not report details that are important for understanding variations in the effectiveness of interventions by settings. Our methods provide a systematic, practical approach to disaggregating SR information for selection of interventions for primary research or implementation.
Objectives: To describe a systematic priority-setting method to identify pregnancy-related interventions for research cohorts.
Methods: Following identification of relevant systematic reviews (SR), two reviewers, independently and in duplicate, extracted data on publication year and included studies. A matrix of all included studies from the SRs identified the extent of overlap between SRs. Where more than two-thirds of studies overlapped, the most recent, high-quality SR, as evaluated using ROBIS, was selected for inclusion. For each pre-specified outcome, we developed a GRADE-based effectiveness matrix incorporating effect size and study quality to rank interventions as: 1) beneficial or harmful; 2) possibly beneficial or possibly harmful; 3) no effect; 4) possibly no effect; or, 5) uncertain effect. For interventions ranked as 1) or 2), additional data regarding study setting, participants, intervention composition, dosing, frequency, duration, feasibility, implementation co-factors and cost were extracted. We engaged with HeLTI stakeholders to examine common findings and potential reasons for differential effects by setting.
Results: The included studies' matrix identified 12 SRs in which there was more than two-thirds overlap, leaving 13 recent, comprehensive, high-quality SRs, for disaggregation. The effectiveness matrix identified four beneficial interventions, one harmful intervention and one possibly beneficial intervention.
Conclusions: Aggregated information from SRs, and specifically meta-analyses, often collapse or do not report details that are important for understanding variations in the effectiveness of interventions by settings. Our methods provide a systematic, practical approach to disaggregating SR information for selection of interventions for primary research or implementation.