Downscaling African disease case data from hospitals to fine resolution maps

Health case data from Africa (e.g., from hospitals, health centres and dispensaries) are collected routinely on a monthly basis within Health Management Information Systems. Recent advances in the geo-locating of facilities and the recording of these case data electronically across Africa have opened up the opportunity for their analysis using space-time statistics. However, the data have some serious deficiencies that need to be addressed, and even when cleaned up, they are inadequate on their own to represent disease risk through space and time (they are just a set of time-series of counts represented at points across Africa). This talk describes the challenges associated with these data and how space-time statistical modelling can be used to address them. Specifically, a case study in Namibia is used to demonstrate how fine-resolution maps of malaria case incidence can be produced from these routinely collected facility case data using space-time statistics.