Intra-country climate variability and geographical distribution of diarrhoea among children under five in Bangladesh

Abstract

BackgroundAssociations between climate and childhood diarrhoea are likely, particularly in the face of climate change. We estimated the effects of climate conditions on hospitalizations due to childhood diarrhoea across three Bangladeshi regions to assist develop regional prevention strategies.MethodsFortnightly hospital admissions (July/2012-June/2017) for diarrhoea (≥3 loose stool in a day, and symptoms lasting ≤7 days) in children under five years of age and fortnightly recordings for average maximum temperature, relative humidity and rainfall for 3 regions (i.e., Mymensingh-R1, Rajshahi-R2 and Sylhet-R3) were modelled using a Poisson regression model with distributed lag linear terms. Flexible spline functions were used to adjust models for season and trend. Lagged (delayed) effects of maximum temperature were estimated over 0-4 fortnights.ResultsHigher temperature was associated with fewer diarrhoea hospitalizations across the 3 regions (IRR 0.90-0.92, 95% CI [0.87-0.89] - [0.93-0.95]). When seasonality and long-term trends were adjusted, estimates moved towards the null. Adding relative humidity to the model did not change the effect estimates. There is evidence of an association between diarrhoea hospitalizations and maximum temperature at different fortnightly lags. When lag effects were adjusted for each other using distributed lag models, a negative association at lag 2 (IRR 0.70, 95% CI 0.55-0.90) was reported in R1.ConclusionsOur results varied from previous studies in that higher temperature is associated with increased diarrhoea hospitalizations in Bangladesh. In addition, results underscore that the relationship between maximum temperature and diarrhoea hospitalizations could be confounded by season and long-term trends. Although the unadjusted lagged effects varied between regions in terms of directions, higher temperatures could cause fewer diarrhoea hospitalizations in Mymensingh for as long as a month. Our findings inform policymakers that climate change may have variable effects in different regions, hence childhood diarrhoea prevention strategies could merit region specific considerations over generic guidelines.