Identifying Temperature Thresholds for Heat Morbidities and Comparison of Effects with Various Extreme Heat Indicators in New York State

BACKGROUND: Extreme heat events can cause or exacerbate several health conditions. While studies have examined the effect of heat on mortality and morbidities, only few have studied the heat-heat morbidity relationship. None have identified heat-heat illness threshold effects or compared effects of heat on heat morbidity using different heat indicators. Objectives of this study included determining 1) heat-threshold effect on heat morbidity; 2) most sensitive indicator of the heat-heat morbidity relationship in New York State (NYS); 3) subgroups of this population vulnerable to extreme heat.

METHODS: This bi-directional, time-stratified, case-crossover study assessed the heat-heat morbidity relationship using several temperature metrics and indicators. Hospital admissions for heat stroke, heat syncope, heat exhaustion, heat edema and other heat morbidities among NYS residents during June 28th-August 31st, 1991-2004 were extracted from NYS’s Statewide Planning and Research Cooperative System (SPARCS) database and geocoded to assign meteorological data from National Center for Atmospheric Research. Heat-threshold on heat morbidity was identified as “turning points” where temperature-morbidity relationship changed significantly. Heat-health effects were assessed and compared using daily temperature, apparent temperature, and three categorical temperature indicators: 1) > 90th percentile of daily mean temperature; 2) Heat waves (definition 1): daily maximum temperatures >= 90°F for >=3 consecutive days; and 3) Heat waves (definition 2): daily maximum temperatures >= 97th percentile for >=2 consecutive days. Conditional logistic regression was used to analyze effects up to six lagged days for different heat indicators. Effect modification by age, gender, race and ethnicity was assessed. 

RESULTS: Compared to other metrics daily average temperatures and apparent temperatures showed highest odds ratios with strongest effects on lag day 1, followed by a trend of decreasing magnitude of association with increasing lag. Temperature threshold effects on heat morbidities were seen at 78°F. Differences were seen across strata with all heat wave indicators but statistically significant differences were seen with 90th percentile temperature indicator (lagged 1 day) among blacks (OR 20.79; 95%CI:12.8, 33.53) vs. others (OR 11.82; 95% CI:9.62, 14.51); females (OR 17.27, 95% CI:13.03, 22.9) vs. males (OR 10.04, 95% CI:7.85, 12.84) and those >= 65 years (OR 16.56, 95% CI:12.93, 21.21) vs. 20-64 years (OR 8.94, 95% CI:6.60, 12.11)

CONCLUSIONS: Heat-heat morbidities threshold was observed at 78°F. The 90th percentile temperature indicator (lag day 1) identified blacks, females and people >65years as vulnerable populations to extreme heat. This study can help policy makers plan adaptation strategies targeting vulnerable populations in their community.