Association between air raids and reported incidence of cholera in Yemen, 2016–19: an ecological modelling study

Background: Yemen continues to endure cholera outbreaks during ongoing conflict and destructive environmental events. Air raids have been used throughout the conflict to target military and civilian infrastructure. We aimed to assess the association between air raids and cholera incidence while taking into account geographical, environmental, economic, and demographic factors that drive outbreaks. Methods: In this ecological modelling study, we used data from Sept 12, 2016, to Dec 29, 2019, for the number of air raids, vegetation coverage, surface water, precipitation, temperature, economic variables, and cholera case and population data to model the association between conflict and the weekly incidence of cholera (per 100 000 people) in Yemen. Data were transformed into weekly intervals and governorates were categorised according to air raid severity (the number of raids in the previous 3 months). We used a negative binomial generalised additive model that accounted for geographical location and environmental, temporal, economic, and demographic variables to estimate incidence rate ratios for the association between air raid severity and cases of cholera. Findings: During the study period, 2 107 912 cases of cholera were reported in Yemen, and a minimum of 11 366 air raids were recorded. After controlling for relevant factors, compared with no air raids, all other levels of air raid severity were significantly associated with cholera incidence. The largest effect was noted in governorates with severe air raid levels (ie, ≥76 during the previous 3 months), which had an incidence rate ratio of 2·06 (95% CI 1·59–2·69; p<0·0001) for cholera compared with governorates with no air raids in the previous 3 months. Economic factors were also significantly associated with increased cholera incidence. Interpretation: Air raids were significantly associated with the burden of cholera in Yemen, even after controlling for other relevant factors. Quantification of this relationship further shows that the cholera outbreak is largely a result of human action rather than a natural occurrence, and demonstrates the conflict's devastating effects on health. Our findings highlight the need for ceasefire and peacebuilding efforts, as well as infrastructure and economic restoration, to reduce Yemen's cholera burden. Funding: None. Translation: For the Arabic translation of the abstract see Supplementary Materials section.