Abstract
•Pediatric enteric illness hospitalizations 1–2 days after heavy rainfall increased most in wet and dry regions.•Detected a “harvesting” phenomenon after 2 day lag in enteric disease hospitalizations.•Case-crossover study nested within a prospective child cohort study design.•The lowest-observable-adverse-effect-level is well below the threshold of weather warning for public health protection.
Heavy rainfall is associated with increased risk of waterborne disease. However, it is not known whether the risk increment differs between wet and dry regions. We examined this question in New Zealand, which has a wide geographical variation of annual rainfall totals (10th–90th percentile difference ≥3000 mm).
We conducted a nested case-crossover study within a prospective child cohort (born in 2009–2010) for assessing transient health effects when modified by longitudinal exposures to rainfall. Short-term heavy rainfall effects on hospitalizations due to enteric bacterial and viral infectious causes at lag of 0–14 days were assessed using a Cox regression model adjusted for daily temperature, relative humidity and evapotranspiration. We derived quantiles of time-weighted long-term rainfall levels at the children’s homes and these were added as an interaction term to the short-term effect model.
Hospitalization risks were higher two days after heavy rainfall days (hazard ratio [95% confidence interval]: 1.73 [1.10–2.70]). The lowest-observable-adverse-effect-level was detected at the 94th percentile of daily rainfall total. Hospital admissions 1–2 days after heavy rainfall increased most in locations with the lowest and highest long-term rainfall. An interaction of this kind between short-term weather and long-term climate has not been reported previously. It is relevant to climate change risk assessments given global projections of increasing intensity of precipitation, against a background of more severe, and possibly more frequent, droughts and flooding.