The effects of monthly temperature fluctuations on mortality in the United States from 1921 to 1985

Abstract

The impact of short-term temperature fluctuations on mortality has been studied mainly on historical populations, thus providing a limited ability to generalize to contemporary conditions, which would be more useful in determining public health policies aimed at reducing mortality. Therefore, this study examined the effects of monthly temperature fluctuations on mortality in the United States from 1921 to 1985. Monthly data about mortality from the Vital Statistics and temperature from the National Oceanic and Atmospheric Administration and the US Department of Agriculture Weather Bureau were used. Six states were selected to be studied (Massachusetts, Michigan, Washington, Utah, North Carolina, and Mississippi). The analysis was carried out using distributed lag models. The analysis showed that warmer than usual temperatures in July and August, and unusually cold temperatures from January to June are linked to higher mortality. From September to December unusually low temperatures are associated with higher mortality in most states, while temperature has no significant effect on mortality in June and September. In January and February mortality is especially affected by unusually cold weather in the southern states of Mississippi and North Carolina. For example, a one degree drop in the mean temperature in 1921 is associated with a more than 3.5% increase in the February crude death rate in Mississippi and North Carolina and a less than 1% increase in the four other states examined. Finally, in the months from January to March the relationship between monthly fluctuations in the crude death rate and temperature declined over time and became relatively weak by 1985.