Distributed lag effects in the relationship between temperature and mortality in three major cities in South Korea

In many cases, the effect of an environmental exposure event is not restricted to the period when it occurs. Understanding the extent to which high-temperature exposure hastens the occurrence of health outcomes is a key to interpreting public health risks correctly and developing appropriate intervention programs related to heat. We explored distributed lag effects in the relationship between high temperature and mortality in summer (June-August) in Seoul, Daegu, and Incheon, South Korea, from 1991 to 2008. A Poisson regression model adapted for time-series data was used to estimate short-term heat-related mortality impacts. To examine the distributed lag effects, we fitted a constrained distributed lag model that included lagged exposure variables as covariates, applying a function of days of lag according to B-spline bases. The effects on mortality caused by high-temperature exposure continued for about 5 days, and slight deficit effects after the initial mortality increases were observed, even if not initially apparent. Thirty days after high-temperature exposure, the cumulative effects were still high in Seoul and Incheon. This study shows a pattern of distributed lag effects of high-temperature exposure that the single-day model could not identify. Our results confirm that summer high temperature has an effect on mortality, not advancing the date of adverse events by a few days. Ultimately, it suggests that public health programs be amended to allow for differential high-temperature effects spread over multiple days.