Estimating spatial disease rates using health statistics without geographic identifiers

Abstract

AbstractMorbidity statistics can be reported as grouped data for health services rather than for individual residence area, especially in low-middle income countries. Although such reports can support some evidence-based decisions, these are of limited use if the geographical distribution of morbidity cannot be identified. This study estimates the spatial rate of Acute respiratory infections (ARI) in census districts in Cúcuta -Colombia, using an analysis of the spatial distribution of health services providers. The spatial scope (geographical area of influence) of each health service was established from their spatial distribution and the population covered. Three levels of spatial aggregation were established considering the spatial scope of primary, intermediate and tertiary health services providers. The ARI cases per census district were then calculated and mapped using the distribution of cases per health services provider and the proportion of population per district in each level respectively. Hotspots of risk were identified using the Local Moran’s I statistic. There were 98 health services providers that attended 8994, 18,450 and 91,025 ARI cases in spatial levels 1, 2 and 3, respectively. Higher spatial rates of ARI were found in districts in central south; northwest and northeast; and southwest Cúcuta with hotspots of risk found in central and central south and west and northwest Cucuta. The method used allowed overcoming the limitations of health data lacking area of residence information to implementing epidemiological analyses to identify at risk communities. This methodology can be used in socioeconomic contexts where geographic identifiers are not attached to health statistics.