Sub-population Identification of Multimorbidity in Sub-saharan African Populations

Abstract

This study develops an automated stratification approach to detect sub-populations with anomalously high or low multimorbidity rates in sub-Saharan African datasets, using survey data from Nairobi (Kenya) and Agincourt (South Africa). The method complements traditional confirmatory analyses, automatically scanning across all possible sub-groups. Results show consistency in high-risk populations across both areas, demonstrating the method's potential for scalable exploratory data analysis.