The Long-term Effects of Flooding on Mortality in Rural Bangladesh

Abstract

O-31C9-2 Background/Aims: Little evidence on the long-term health impact of flooding. To determine the extent to which flooding is accompanied by change in mortality over the 3 years following flooding in rural Bangladesh. Methods: Flooded and non-flooded areas at the time of severe flood in the summer of 2004 were identified by interviewing the head of 9524 baris (on average 5 households in a bari) which is 90% of all baris in the Matlab Health and Demographic Surveillance area with more than 220,000 residents. Mortality in the flooded areas in the 3 years after the 2004 flood was compared with mortality in the 3 years before the flood in the same areas, controlling for temporal patterns observed in “control” areas that were never flooded in 2004. Comparisons were stratified by sex, age group (0–1, 1–14, 15–59, 60–74, 75+), hygiene and sanitation, and socioeconomic status. Results: A total of 13,957 of interviewed households (31%) were flooded in the summer 2004. Overall, there was a slight but nonsignificant excess compared to pre-flood mortality during the flood, June–September (RR = 1.16; 95% CI: 0.93–1.45) and no excess after the flood; RR = 1.02 (95% CI: 0.89–1.17), 0.93 (95% CI: 0.81–1.07) and 0.99 (95% CI: 0.86–1.18) in the year 1, 2, and 3, respectively, after the flood. Subgroup analyses showed suggestive excesses in the year after the flood in persons aged above 75 (RR = 1.20; 95% CI: 0.92–1.57) and in the low income group (RR = 1.16; 95% CI: 0.82–1.62), but precision was limited. Conclusion: We found no evidence of an adverse effect of flood on mortality persisting beyond the flood period itself. Although power was limited, upper confidence interval suggest that risks above 1.2 in the population overall, and above about 1.6 in important sub-groups are unlikely, which contrasts with some other reports. Results for diarrhea or ARI morbidity, for which power will be greater, are being computed and will be presented.