Novel Method for Measuring Ambient Heat Exposure–Acute Healthcare Utilization Associations Within a Safety Net: A Retrospective, Longitudinal Study

Background

Extreme and inequitable heat exposures cause weather-related deaths. Associations between maximum daily temperature and individual-level healthcare utilization have been inadequately characterized.

Objective

To evaluate and compare demographic and clinical associations for an individual’s healthcare utilization between high- and low-temperature periods.

Design

Retrospective, 5-year longitudinal study of acute care utilization comparing high-temperature periods (HHP) and low-temperature periods (LHP) defined by local maximum daily temperature. Using duration of observation, cases served as their own controls. Temperature-dependent utilization was reported as unadjusted incident rate ratio (IRR) using Poisson regression and log-transformed variable coefficients. IRRs were adjusted (aIRR) for demographic characteristics, heat-sensitive conditions/diagnoses, and neighborhood heat vulnerability score; false discovery rate p-values were adjusted for multiple comparisons.

Subjects

Patients aged ≥ 4 years visiting Denver Health between 4/10/2016 and 12/31/2020, with ≥ 2 visits over ≥ 365 days.

Main Measures

Comparison of an individual’s acute care visit rates in HHP versus LHP, stratified by demographic characteristics and heat-sensitive clinical conditions.

Key Results

While acute care utilization occurred at similar or higher rates during LHP compared with HHP, certain groups (i.e., Native Americans and those with congestive heart failure, liver failure, and/or alcohol use) had higher rates of utilization during HHP. Significant associations existed for acute care utilization by age, sex, racial and ethnic groupings, clinical characteristics, and neighborhood heat vulnerability. Adjusting for demographic and environmental covariates, individuals with any heat-sensitive clinical condition had higher HHP vs LHP utilization compared to those without (aIRR = 1.93).

Conclusions

Significant heat-related utilization occurred among individuals with heat-sensitive clinical conditions compared with those without. Demographic characteristics (e.g., older) and specific clinical conditions (e.g., liver failure) demonstrated higher utilization. In real-time, chronic disease management programs could proactively identify at-risk individuals for interventions which reduce heat-related morbidity and healthcare utilization.