Heat-Associated Death Surveillance, What's All the Heat about? Comparison of Methodologies

BACKGROUND:   Maricopa County Department of Public Health (MCDPH) tracks heat-associated deaths year-round using a combination of International Classification of Disease-10 (ICD-10) codes and key word searches for heat-associated causes of death (COD) for all deaths occurring in Maricopa County. This methodology differs from that recommended by the Council of State and Territorial Epidemiologists (CSTE) which is based on ICD-10 codes alone, and excludes cases that reside out-of-jurisdiction or cases occurring outside of May-September. By comparing methodologies, surveillance systems can be improved so that the true effects of heat can be evaluated.

METHODS:   SAS Enterprise Guide 5.1 was used to query Maricopa County death certificates from 2009-2012 using the following criteria: CSTE: ICD-10 codes X30 or T67.0 – T67.9 in underlying and contributing COD fields for Maricopa County residents and deaths that occurred in May-September. MCDPH: ICD-10 codes X30 or T67.0 – T67.9 in underlying and contributing COD fields and a free-text search of COD fields using heat-associated key words (heat exposure; environ; exhaustion; sun; heat stress; heat stroke; hyperthermia), which have been previously validated in Maricopa County, for all death occurrences, year-round. In addition, a review of the COD was performed to confirm the finding.

RESULTS:   The MCDPH method identified 370 heat-associated deaths compared to 259 using CSTE. All but three of the CSTE cases were identified using the MCDPH method; these cases were excluded by the MCDPH method because, for example, they were associated with exposure to hot bath water.  The CSTE method missed 30% of heat-associated deaths that were identified by MCDPH because they had non-heat ICD-10 codes (25%), were non-residents (81%), or occurred out of the heat season (9%); categories are not mutually exclusive. Of cases that were not included in the CSTE methodology because of ICD-10 coding, residency, or out of season, 48%, 82%, and 80%, respectively, had an environmental heat term listed in part I of the death certificate.

CONCLUSIONS:   Evaluation of methodologies helps to enhance sensitivity for capturing the true impact of heat-associated deaths. For jurisdictions like Maricopa County where excessive heat is a chronic environmental condition, it is imperative to monitor cases year-round. Additionally, heat is a public health threat that affects residents and visitors alike. By excluding cases that are non-residents, surveillance systems can underestimate the true effect of heat in the community. Because heat-associated death data are largely used for prevention efforts, an accurate depiction of the burden of heat effects is essential.