This community innovation has been accepted at the 2025 DHIS2 Annual Conference
Deriving population estimates for DHIS2 data
In resource limited settings, population estimates are commonly derived from inaccurate or outdated national registries or census data, and may not represent catchment populations for the administrative unit relevant to the monitoring or planning activity being evaluated. Geo referenced Infrastructure and Demographic Data for Development (GRID3) is a novel modelling application that works with countries to generate and validate geospatial data by combining current neighborhood scale microcensus surveys with national scale data from satellite images and digital maps to derive population estimates. Using GRID3, we estimated health facility catchment populations to construct model denominators (number of reproductive aged females in each health facility catchment area) in an evaluation of an intervention on contraception use in Nigeria using DHIS2 data. We derived catchment population estimates for 1,172 facilities in the states of Kano, Jigawa, and Katsina, based on an accessibility model (Euclidean distance) to the nearest facility using GRID3 health facilities point locations. Using ArcGIS Pro 3.3.2 (ESRI) software, we created 2 kilometer buffers and Thiessen polygons for each health facility, and clipped the polygons and buffers to create non overlapping catchment areas. Population estimates were summed for each facility catchment area with each 100m2 grid cell providing an estimated population range (with uncertainty levels). The mean value was the assumed population for that grid. We successfully matched and determined catchment populations for 1,172 facilities (97%); 35 facilities were unavailable in the GRID3 system. Catchment population estimates ranged from 64 people to 32,717; these estimates were used to determine rates for our evaluation outcomes. GRID3 provides a feasible and accessible method to calculate catchment populations for use with DHIS2 and other health system data.
Primary Author: Andrea Stucchi
Keywords:
routine health information system data, population estimates, geospatial data, administrative catchments, health monitoring and planning, evaluation outcomes
3 Likes