A multidimensional space-time geospatial analysis for examining the spatial trends of vector-borne diseases: 20 years of malaria in Kenya

Malaria continues to be a global burden with a disproportionate share of cases reported on the African continent. Knowing where malaria is distributed can aid in the development of different intervention strategies. Since geographic information on malaria is now available for 20 years, we examined how the distribution of malaria and the use of malaria interventions such as bed nets and antimalarials have changed over time in Kenya.

Multidimensional space-time pattern mining methods were used to identify malaria hot spots and examine how these have changed over time. Twenty years of Plasmodium falciparum incidence, mortality and prevalence data and intervention use data for bednets and antimalarials were obtained from the Malaria Atlas Project. We conducted a local hotspot analysis (LHA) and an emerging hotspot analysis (EHA) on the full dataset (2000–2020) and at 10-year intervals (2000–2009 and 2010–2019). The EHA was used for identifying the directional shift in disease clusters and whether these hotspot clusters are intensifying or expanding over time. The LHA identified areas with disease hotspots, coldspots, outliers or a mixture of where these different types were distributed. For each LHA hotspot type, we further examined annual malaria prevalence and intervention use trends.

In this study, we found that in Kenya there has been a general decline in malaria prevalence with a slight increase in between 2016–2018. The spatial distribution of malaria is changing in Kenya. We identified four key malaria zones with an additional two areas where malaria has been increasing. Bed net and anti-malarial use has increased over time. Although malaria has been greatly reduced in Kenya, malaria continues to be a problem mainly in cross-border regions. Having long-term data are useful for evaluating changes in the distribution of malaria and exploring the factors contributing to these changes over time. The space-time methods used in this study are useful for identifying where high-burden areas are distributed and provide insights into the trends taking place within these geographic locations, and as such, are useful for refining malaria elimination strategies and helping reduce the burden of malaria. Since the data used in this study are openly available and available globally, the analyses conducted here can be used elsewhere.