Detecting the net effect of flooding on infectious diarrheal disease in Anhui Province, China: a quasi-experimental study

Abstract

Though a number of studies have shown positive relationships between flooding events and infectious diarrhea, there is a paucity of rigorous evidence regarding the net effect of flooding on diarrhea incidence, controlling for existing pre-trends and meteorological confounders. The study treats the 2016 catastrophic flood event in Anhui Province, China as a natural experiment using a difference-in-differences design with propensity score matching to exclude background variations of diarrhea occurrence and meteorological effects, thus isolating the net effect of flooding on diarrhea. A triple-differences analysis was further deployed to identify the potential effect modifiers, including gender, age, occupation and community health resources. By analyzing 359 580 cases of diarrhea that occurred before, during and after the flooding, we show that the 2016 flood event significantly increased the risk of dysentery (RR: 1.29, 95%CI: 1.15–1.46) in during-flood period, and also increased the risk of all-cause diarrhea (RR: 1.21, 95%CI: 1.17–1.26), typhoidal diarrhea, dysentery, and other infectious diarrhea in post-flood period. Children, males and non-farmers were particularly vulnerable to flooding impacts and the density of health professionals was found to be protective against diarrheal risk in both during-flood (RR: 0.81, 95% CI: 0.72–0.92) and post-flood (RR: 0.83, 95% CI: 0.77–0.88) periods. This study employs quasi-experimental design and provides a better understanding on both acute and sustained effects of flooding on diarrhea, which is important for accurate health impact assessments and developing targeted intervention strategies.