Background: To describe spatiotemporal patterns of diarrhoea in Bhutan, and quantify the association between climatic factors and the distribution and dynamics of the disease.
Methods: Nationwide data on diarrhoea were obtained for 2003 to 2013 from the Health Information and Management System (HIMS), Ministry of Health, Bhutan. Climatic variables were obtained from the Department of Hydro Met Services, Ministry of Economic Affairs, Bhutan. Seasonal trend decomposition was used to examine secular trends and seasonal patterns of diarrhoea. A Bayesian conditional autoregressive (CAR) model was used to quantify the relationship between monthly diarrhoea, maximum temperature, rainfall, age and gender.
Results: The monthly average diarrhoea incidence was highly seasonal. Diarrhoea incidence increased by 0.6% (95% CrI: 0.5-0.6%) for every degree increase in maximum temperature; and 5% (95 Cr I: 4.9-5.1%) for a 1 mm increase in rainfall. Children aged <5 years were found to be 74.2% (95% CrI: 74.1-74.4) more likely to experience diarrhoea than children and adults aged ≥5 years and females were 4.9% (95% CrI: 4.4-5.3%) less likely to suffer from diarrhoea as compared to males. Significant residual spatial clustering was found after accounting for climate and demographic variables.
Conclusions: Diarrhoea incidence was highly seasonal, with positive associations with maximum temperature and rainfall and negative associations with age and being female. This calls for public health actions to reduce future risks of climate change with great consideration of local climatic conditions. In addition, protection of <5 years children should be prioritize through provision of rotavirus vaccination, safe and clean drinking water, and proper latrines.
Keywords: Bayesian analysis; Bhutan; Diarrhoea; Spatial analysis; Time series analysis.