Key points are not available for this paper at this time.
The study of population health through network science is promising but suitable population health datasets covering low- and middle-income countries (LMICs) are not available. Covariate-based methods used to produce small-area estimates (SAEs) combine national health surveys with covariates from varied sources through various methods limiting their use for producing network representations of populations by injecting unquantifiable uncertainty into estimates of node attributes, affecting the comparability of representations across countries and time. Here, we present SEEDNet (Settlement-level Epidemiological Estimates Datasets for Network Analysis), a multi-country data library of population health indicators across human settlements. Our datasets are produced through a covariate-free method that uses georeferenced national surveys to produce SAEs of health indicators and include complete mapping of population settlements of all sizes. Our open-access library is intended to be used as the basis for network representations of population health in LMICs. Novel aspects include automated estimation process, harmonized data inputs, complete settlement mapping and the adoption of settlements as the functional units for network-based analysis of epidemiological data.
Darooneh et al. (Tue,) studied this question.