How much hotter it is than normal for an area better predicts the deaths and pre-term births linked to heat waves than just how hot it is, according to a study published today in the journal Environmental Health Perspectives.
The new study identified which of 15 published heat wave definitions were most closely associated with non-accidental deaths and pre-term births found in Alabama vital records, along with what those definitions had in common. Non-accidental deaths were measured because they tend to result from a combination of disease, aging and heat, as opposed to those more closely linked to social factors, like car crashes and suicides. Preterm births have been linked to heat as well, with one theory holding that heat and dehydration trigger contractions.
“Heat alerts are issued based on weather forecasts and provide a warning to the public that it may soon get hot enough to be dangerous,” said Julia Gohlke, Ph.D., assistant professor in the Department of Environmental Health Sciences within the UAB School of Public Health at the University of Alabama at Birmingham, and corresponding study author. “It makes sense then that alerts based on a heat wave definition that better predicts adverse health outcomes would better protect the public.”
According to the National Weather Service, heat waves cause hundreds of deaths per year in the United States. The weather service and state health departments are looking for ways to more effectively warn local communities about extreme heat. Hot weather and humidity reduce the body’s ability to cool, with some research suggesting that it may exacerbate heart disease and breathing problems.
Meteorologists and public health researchers have studied heat waves for different reasons in recent decades, with the result that there is no standard definition.
“The lack of consensus makes it harder for scientists to track the impact of heat on public health and for authorities to communicate the risks to communities,” said Shia Kent, Ph.D., a postdoctoral scholar in Gohlke’s group and first author on the paper.
Statistical analyses revealed that heat waves defined by how much hotter it is than normal (relative) in a given locale had a statistically significant, positive association with the two health outcomes, and that those using absolute temperature did not.
Along the same lines, heat wave definitions based on average temperature across a 24-hour period (daytime and nighttime heat) better predicted non-accidental deaths and pre-term births than did any single temperature reading recorded during that period, perhaps because hot nights provide no relief. The study also found that heat waves defined by temperature alone better predicted heat-related health problems than did those that included humidity and wind speed.
In a glimpse at the potential impact of an evidence-based heat wave definition, the new study found that a heat wave defined as an average daily temperature hotter than 98 percent of all days since 1990 in a given ZIP code for two days in a row was associated with a 32.4 percent increase in the number of premature births found in public records when compared to non-heat wave days. A heat wave defined as an average daily temperature hotter than 90 percent of all days since 1990 in a given ZIP code for two days in a row was associated with a 3.7 percent increase in non-accidental deaths, again compared to non-heat wave days.
Determining heat-related health risk by ZIP code made for a much more accurate analysis of heat impact on health than did past studies that recorded temperature in one place — the weather station at the nearest airport — meant to represent several surrounding communities. The current study achieved a by-ZIP-code grid of heat wave definitions using a NASA temperature data set that combines satellite and ground data. Ben Zaitchik and Tiffany Smith in the Department of Earth and Planetary Sciences at Johns Hopkins University, experts in the use of the NASA tool, partnered with the UAB team for the study.
Better by region
The South has experienced more heat wave days from 1979 to 2011 than have other regions in the United States, according to another recent study by the JHU team and Gohlke.
“Other studies have found that heat waves have more dire consequences in the urban north, but our study suggests that the question should be further studied by region,” said Gohlke. “Our specific study results apply to Alabama, and we have met with the Birmingham office of the National Weather Service and the Alabama Department of Public Health about the potential for changing the definition used to trigger heat alerts here. Nationally, studies like ours can give local NWS offices information based on local health records to help them decide when to issue heat alerts.”
Lastly, the team found heightened risk of deaths and pre-term births in both urban and rural areas as temperatures rose, highlighting the need for better heat wave response systems in both places.
“The field has focused on cities because developed land heats up quicker, but people in rural areas may be more likely to be outdoors and farther from the nearest hospital if something goes wrong,” said Gohlke.
“The public often views severe weather as the only dangerous weather, but the research at UAB clearly indicates that heat-related deaths and illnesses are continuing in large numbers,” said Kevin Laws, science operation officer at the NWS Forecast Office in Birmingham. “As scientists, we often struggle with meaningful definitions for weather, especially when determining critical thresholds like heat waves. We will integrate the UAB study results into our definitions with the goal of saving more lives.”
“We agree that a better understanding of the heightened risk of deaths and preterm births from excessive heat will help to improve health outcomes,” said Jim McVay, DrPA, director of the Bureau of Health Promotion and Chronic Diseases within the ADPH.
Along with Gohlke and Kent, Leslie McClure helped to design the statistical analyses. The work was supported by a grant from the National Institute of Environmental Health Sciences (R21 ES020205)