Nationwide variation in the effects of temperature on infectious gastroenteritis incidence in Japan

Although several studies have investigated the effects of temperature on the incidence of infectious gastrointestinal disease in a single city or region, few have investigated variations in this association using nationwide data. We obtained weekly data, gathered between 2000 and 2012, pertaining to infectious gastroenteritis cases and weather variability in all 47 Japanese prefectures. A two-stage analysis was used to assess the nonlinear and delayed relationship between temperature and morbidity. In the first stage, a Poisson regression allowing for overdispersion in a distributed lag nonlinear model was used to estimate the prefecture-specific effects of temperature on morbidity. In the second stage, a multivariate meta-analysis was applied to pool estimates at the national level. The pooled overall relative risk (RR) was highest in the 59.9 th percentile of temperature (RR, 1.08; 95% CI: 1.01, 1.15). Meta-analysis results also indicated that the estimated pooled RR at lower temperatures (25 th percentile) began immediately but did not persist, whereas an identical estimate at a higher temperature (75 th percentile) was delayed but persisted for several weeks. Our results suggest that public health strategies aimed at controlling temperature-related infectious gastroenteritis may be more effective when tailored according to region-specific weather conditions.