K-Means Clustering in Monitoring Facility Reporting of HIV Indicator Data: Case of Kenya

Abstract

Health management information systems (HMISs) in low- and middle-income countries have been used to collect large amounts of data after years of implementation, especially in support of HIV care services. National-level aggregate reporting data derived from HMISs are essential for informed decision-making. However, the optimal statistical approaches and algorithms for deriving key insights from these data are yet to be fully and adequately utilized. This paper demonstrates use of the k-means clustering algorithm as an approach in supporting monitoring of facility reporting and data-informed decision-making, using the case example of Kenya HIV national reporting data. Results reveal four homogeneous cluster categories that can be used in assessing overall facility performance and rating of that performance.