gpaasi/scd-structural-vulnerability-hidden-burden-uganda

Reproducible spatial analysis workflow for mapping health-facility reported sickle cell disease admissions and deaths, epidemiological co-risk, SCD service access, structural vulnerability, and hidden burden priority catchments in Uganda. # When low burden is not low need: structural vulnerability and hidden burden of health-facility reported sickle cell disease admissions and deaths in Uganda ## Summary This repository implements the analytic workflow for the paper, **“When low burden is not low need: structural vulnerability and hidden burden of health-facility reported sickle cell disease admissions and deaths in Uganda.”** The workflow analyses health-facility reported sickle cell disease (SCD) admissions and deaths across health-facility catchments in Uganda from 2020 to 2024. It integrates routine facility-reported SCD burden data, facility service-inventory data, health-facility catchment polygons, raster-derived haemoglobin S burden, *Plasmodium falciparum* parasite prevalence among children aged 2 to 10 years, population data, and subregional boundaries. The analysis constructs catchment-level measures of epidemiological co-risk, SCD service access, service deficit, structural vulnerability, observed-to-expected mismatch, and hidden-burden priority. It then fits Bayesian spatial Poisson models to examine how co-risk and service deficit shape the spatial visibility of recorded SCD admissions and deaths. The workflow supports the central interpretation of the manuscript: health-facility reported SCD admissions and deaths are measures of **recorded burden**, not direct measures of population need. Low recorded burden in high-risk, weak-service catchments may reflect limited diagnostic, referral, inpatient, or reporting visibility rather than low underlying SCD burden. ## Study aims This repository supports three linked aims: 1. To describe catchment-level heterogeneity in recorded SCD admissions and deaths, epidemiological co-risk, SCD service access, and structural vulnerability. 2. To estimate associations between co-risk, service deficit, structural vulnerability, and recorded SCD admissions and deaths using Bayesian spatial models. 3. To identify hidden-burden priority ca …