Socio-demographic, not environmental, risk factors explain fine-scale spatial patterns of diarrhoeal disease in Ifanadiana, rural Madagascar

Proceedings of the Royal Society B

A team of PIVOT community health workers traverse the mosaic landscape of Ifanadiana

Corresponding Author: Michelle Evans

Summary by Ethan Hackmeyer

Despite being both preventable and treatable, diarrhoeal disease (DD) causes over 700,000 child deaths annually due to inadequate  water and sanitation infrastructure. These losses are  primarily concentrated in 15 low-income countries. Distribution of DD is highly influenced by environmental factors, such as climate, in addition to socio-economic and cultural factors. Disease dynamics can be difficult to predict for DD, as successful transmission can vary widely within fine spatial scales, and most data is available at coarse scales (i.e. national or subnational level). Precision health mapping is a method of tracking disease spread in finer spatial scales that could possibly be used to better understand how certain diseases such as DD are transmitted. CEID Member Michelle Evans is lead author on a new study that seeks to determine whether precision health mapping is an effective tool for predicting DD occurrence within a health district in order to inform local health programs. This study is a collaboration of researchers at the University of Georgia; the Ministry of Health, National Institute of Statistics and PIVOT in Madagascar, Harvard Medical School, Institut de Recherche pour le Developpement, and Cornell University.

Dr. Evans and collaborators used multiple spatio-temporal datasets to identify socio-ecological variables associated with DD in the rural health district of Ifanadiana in southeastern Madagascar. They then assessed their ability to use precision health mapping to predict disease risk at a scale relevant to public health managers. From the climatic and socio-economic data collected in Ifanadiana, generalized linear mixed models were constructed to predict spatial distribution of DD.

Precision health mapping based on climatic factors alone was found to be suboptimal at predicting DD occurrence. Though the models found that environmental variables were not strong predictors of DD incidence and prevalence, climatic variables predicted strong seasonality. Socio-economic factors were found to be strong predictors of disease incidence and risk at fine spatial scales, and events such as national holidays were found to be strongly correlated with a rise in cases of DD.

More research will be necessary to determine the proper situations in which the use of precision health mapping could provide accurate information. With the addition of socio-demographic covariates to precision health mapping models, it is possible that the method could be useful for a larger range of situations at the fine scale.

Evans Michelle V., Bonds Matthew H., Cordier Laura F., Drake John M., Ihantamalala Felana, Haruna Justin, Miller Ann C., Murdock Courtney C., Randriamanambtsoa Marius, Raza-Fanomezanjanahary Estelle M., Razafinjato Bénédicte R. and Garchitorena Andres C. 2021. Socio-demographic, not environmental, risk factors explain fine-scale spatial patterns of diarrhoeal disease in Ifanadiana, rural Madagascar. Proc. R. Soc. B. 288:20202501. http://doi.org/10.1098/rspb.2020.2501