Models to predict the intensity of Plasmodium falciparum transmission: applications to the burden of disease in Kenya.
Snow RW., Gouws E., Omumbo J., Rapuoda B., Craig MH., Tanser FC., le Sueur D., Ouma J.
There is an increasing need to provide spatial distribution maps of the clinical burden of Plasmodium falciparum malaria in Africa. Recent evidence suggests that risk groups and the clinical spectrum of severe malaria are related to the intensity of P. falciparum transmission. Climate operates to affect the vectorial capacity of P. falciparum transmission and this is particularly important in the Horn of Africa and parts of East Africa. We have used a fuzzy logic climate suitability model to define areas of Kenya unsuitable for stable transmission. Kenya's unstable transmission areas can be divided into areas where transmission potential is limited by low rainfall or low temperature and, combined, encompass over 8 million people. Among areas of stable transmission we have used empirical data on P. falciparum infection rates among 124 childhood populations in Kenya to develop a climate-based statistical model of transmission intensity. This model correctly identified 75% (95% confidence interval CI 70-85) of 3 endemicity classes (low, < 20%; high, > or = 70%; and intermediate parasite prevalences). The model was applied to meteorological and remote sensed data using a geographical information system to provide estimates of endemicity for all of the 1080 populated fourth level administrative regions in Kenya. National census data for 1989 on the childhood populations within each administrative region were projected to provide 1997 estimates. Endemicity-specific estimates of morbidity and mortality were derived from published and unpublished sources and applied to their corresponding exposed-to-risk childhood populations. This combined transmission, population and disease-risk model suggested that every day in Kenya approximately 72 and 400 children below the age of 5 years either die or develop clinical malaria warranting in-patient care, respectively. Despite several limitations, such an approach goes beyond 'best guesses' to provide informed estimates of the geographical burden of malaria and its fatal consequences in Kenya.