Comprehensive Summary
This study asked whether county-level environmental and social exposome factors predict one-year readmission or mortality after hospitalization for heart failure. Researchers conducted a retrospective cohort study using the OneFlorida+ clinical research network, which includes electronic health records from 2016-2022. The sample comprised 63,940 patients (mean age 65 years, 48% women). Patient addresses were linked to 1,308 county-level variables across domains such as economic stability, education, healthcare access, and environmental conditions. Because of the high dimensionality, the authors used LASSO regularization for variable selection and then mixed-effects logistic regression under an exposome-wide association study framework. The primary composite endpoint was all-cause mortality or hospital readmission within one year of index heart failure admission. The key finding was that higher maximum May temperatures were associated with increased adverse outcomes (adjusted odds ratio 1.04, 95% CI 1.02-1.06). No other exposome factors remained significant after correction. The temperature effect was consistent across age, sex, race, socioeconomic, and rural/urban subgroups.
Outcomes and Implications
This study shows how computational modeling can extract meaningful predictors from electronic health record data. Specifically, LASSO regularization and mixed-effects logistic regression can be applied to areas of medicine outside of heart failure as well to identify new predictors of various conditions. Clinically, the results suggest that seasonal heat exposure increases risk for poor outcomes in heart failure, underscoring the need for closer monitoring and patient education during warmer months.