Comparison of temperature-mortality associations estimated with different exposure metrics

Abstract

Background: Studies of the short-term association between ambient temperature and mortality often use temperature observations from a single monitoring station, frequently located at the nearest airport, to represent the exposure of individuals living across large areas. Population-weighted temperature estimates constructed from gridded meteorological data may offer an opportunity to improve exposure assessment in locations where station observations do not fully capture the average exposure of the population of interest. Methods: We compared the association between daily mean temperature and mortality in each of 113 United States counties using (1) temperature observations from a single weather station and (2) population-weighted temperature estimates constructed from a gridded meteorological dataset. We used distributed lag nonlinear models to estimate the 21-day cumulative association between temperature and mortality in each county, 1987–2006, adjusting for seasonal and long-term trends, day of week, and holidays. Results: In the majority (73.4%) of counties, the relative risk of death on extremely hot days (99th percentile of weather station temperature) versus the minimum mortality temperature was larger when generated from the population-weighted estimates. In contrast, relative risks on extremely cold days (first percentile of weather station temperature) were often larger when generated from the weather station observations. In most counties, the difference in associations estimated from the two temperature metrics was small. Conclusions: In a large, multi-site analysis, temperature-mortality associations were largely similar when estimated from weather station observations versus population-weighted temperature estimates. However, spatially refined exposure data may be more appropriate for analyses seeking to elucidate local health effects.

Evaluating the association between extreme heat and mortality in urban Southwestern Ontario using different temperature data sources

“…Weinberger et al examined the relationship between temperature and mortality using the Parameter elevation Regression on Independent Slopes Model (PRISM) vs. weather stations across the United States. In the majority of counties, PRISM led to slightly larger relative risks of death with high temperature compared with weather stations 14 . Lee et al examined the association between modelled daily mean air temperature (1 km resolution aggregated to the zip code level) and non-accidental mortality in urban and rural regions of the Carolina and Georgia states (2007–2011).…”

“…Reasons for observed differences between our studies and the latter studies might be manifold. Studies were conducted outside of Canada or not focused on cities, and there are known area-level differences in temperature susceptibility due to geographic, built environment, demographic and the social characteristics of populations 14 , 47 – 50 . Studies used different methodologies (i.e.…”

“…We conducted population-based time series analyses to investigate the association between temperature and mortality in urban Ontario, Canada. Given heat-health risks vary regionally 14 , we focused upon Southwestern Ontario 18 , a 21,639 km 2 region, bounded North to South by the Bruce Peninsula and Lake Erie (latitudes 41.92–45.23), and West to East by Windsor and Niagara Falls (longitudes − 83.11 to − 78.93). Southwestern Ontario has moderate humid continental climate, with warm to hot humid summers 19 .…”

Spatial and intraseasonal variation in changing susceptibility to extreme heat in the United States

“…We used spatially resolved temperature estimates because our geographic units of analysis covered large areal extents, which may not adequately be represented by observations at a single weather station. 21 , 22 We chose to use T max instead of T mean primarily to maximize the applicability of the findings to policy interventions—particularly heat alerts, which are generally issued based on forecasted high temperatures or heat indices. To achieve spatial consistency with the metropolitan-scale mortality data, we calculated population-weighted averages of the T max estimates for each county based on the census tract population centroids from the 2000 Census: 23 we extracted the TopoWx pixel value underlying each census tract centroid in the county of interest and weighted it by the proportion of the county population living in that tract.…”

Time-series analysis of daily ambient temperature and emergency department visits in five US cities with a comparison of exposure metrics derived from 1-km meteorology products

“…Accurate exposure assessment is an important component of population-based health studies, in part because the health association estimates are frequently used in subsequent risk assessment and health impact analyses [ 10 , 11 ]. While the use of spatially-resolved exposures has become common in air pollution epidemiology, only a few studies on temperature and mortality have considered heat exposure’s spatial variation in short-term health effect studies [ 12 – 14 ]. Though spatially-resolved estimates have been utilized in studies that focus on spatial exposure contrasts [ 15 – 18 ].…”