Reporting of health estimates prior to GATHER: a scoping review

Background: Generating estimates of health indicators at the global, regional, and country levels is increasingly in demand in order to meet reporting requirements for global and country targets, such as the sustainable development goals (SDGs). However, such estimates are sensitive to availability of input data, underlying analytic assumptions, variability in statistical techniques, and often have important limitations. From a user perspective, there is often a lack of transparency and replicability. In order to define best practices in reporting data and methods used to calculate health estimates, the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER) working group developed a minimum checklist of 18 items that must be reported within each study publishing health estimates, so that users may make an assessment of the quality of the estimate. Objective: We conducted a scoping review to assess the state of reporting amongst a cross-sectional sample of studies published prior to the publication of GATHER. Methods: We generated a sample of UN reports and journal articles through a combination of a Medline search and hand-searching published health estimates. From these studies we extracted the percentage of studies correctly reporting each item on the checklist, the proportion of items reported per study (the GATHER performance score), and how this score varied depending on study type. Results: The average proportion of items reported per study was 0.47, and the poorest-performing items related to documentation and availability of input data, availability of the statistical code used and the subsequent output data, and a complete detailed description of all the steps of the data analysis. Conclusions: Methods for health estimates are not currently fully reported, and the implementation of the GATHER guidelines will improve the availability of information required to make an assessment of study quality.