Models and input parameters for the estimation of heat-related excess mortality: a scoping review

This scoping review aimed to identify existing methodological approaches and input parameters to estimate heat-related excess mortality based on real-case human populations from studies conducted worldwide.

Human-induced climate change is increasing the frequency and severity of heat waves, leading to a rise in heat-related deaths globally. As direct measurement of heat-related mortality is challenging, many countries have implemented a statistical model to estimate excess deaths, with ongoing efforts to refine it. An overview of methodological approaches and input parameters used should help to improve existing models.

The included studies featured excess mortality due to high temperature as an outcome, and specifically named and defined parameters used in their model(s). Observational studies in English, German, Spanish, or French were eligible for incluison. Studies on non-humans, studies focused on cold or indoor temperatures, studies using prognostic-only models, and all types of reviews were excluded.

The main information sources searched were PubMed, GreenFILE, and Web of Science Core Collection, with the final search conducted in August 2023. Study selection involved dual independent screening, with data extracted via a standardized form and partly supported by artificial intelligence. Data were analyzed descriptively and qualitatively, with results presented narratively and visualized through graphs.

This review analyzed 197 studies on heat-related excess mortality, highlighting that most used daily death (97%) and daily temperature data (77%), with relative risk being the most common outcome measure. Various temperature indicators and environmental parameters were included. Most studies used more than 1 temperature indicator: mean temperature was used in 123 studies, daily maximum in 104 studies, daily minimum in 79 studies, diurnal temperature change in 8 studies, and other temperature indicators in 20 studies. Environmental parameters were used in 58% of the studies, among them relative humidity, ozone, fine dust, and wind. Statistical models such as distributed lag nonlinear models were used in 68 studies, generalized linear models in 52 studies, and generalized additive models in 28 studies (some studies used multiple models), with distributed lag models gaining popularity since 2010. The review also identified a wide range of alternative modeling approaches, emphasizing the diversity in methods and inputs used across studies.

A wide range of studies on heat-related excess mortality highlights the diversity in data sources, temperature indicators, environmental parameters, and modeling approaches, underscoring the need for standardized, adaptable methods to inform effective heat-health interventions.

OSF https://osf.io/v356m/overview.

A German-language version of the abstract of this review is available: http://links.lww.com/SRX/A175 .