Estimating the mortality effect of the July 2006 California heat wave

Abstract

Objective As a result of the California heat wave in July 2006, county coroners reported that the high temperatures during that period caused approximately 147 deaths. However, heat wave-related deaths are likely to be underreported due to a lack of a clear case definition and the multi-factorial nature of heat-related mortality. Public health policy suggests a need for a careful assessment of mortality following a heat wave. In addition, it is useful to provide a comparison of the mortality impact per degree change during heat waves versus high temperatures observed during non-heat wave periods. Design Daily data were collected for mortality, weather and ozone in seven California counties impacted by the July 2006 heat wave. The association between apparent temperature and daily mortality was assessed using a Poisson regression model and combined across counties in a meta-analysis. These results were then used to estimate the increases in the number of deaths during the heat wave. Results Our analysis indicated that during the July heat wave, there was a 9% (95% CI=1.6, 16.3) increase in daily mortality per 10° Fahrenheit (F) change in apparent temperature for all counties combined. This estimate is almost 3 times larger than the effect estimated for the full warm season of May–September, during the non-heat wave years. Our estimates also determined that actual mortality during the July 2006 heat wave was 2–3 times greater than the coroner estimates. Conclusion This multi-county analysis provides additional evidence that the attributable risk of mortality following a heat wave may be underestimated by examining only direct heat-related deaths. In addition, we have found that the mortality effect per °F is several times higher than that reported during non-heat wave periods.