Lagged and non-linear effects of temperature extremes on mortality in US federal prisons: a case-crossover study, 2009-2021

Background: – Temperature extremes are becoming more common due to climate change. Due to infrastructural issues and the demographic characteristics of the incarcerated population, incarcerated people are at high risk of poor health outcomes during extreme temperature events. However, limited empirical evidence exists, with no prior studies on US federal prisons. Our goal was to estimate associations between ambient temperature and all-cause mortality in federal prisons. Methods: – Data on mortality in federal prisons from 2009 to 2021 were linked with daily mean temperature data. We employed a case crossover design and distributed lag non-linear models to estimate the temperature-mortality relationship for deaths occurring during the hot (May-September) and cold (November-March) months throughout the US. Odds ratios were estimated comparing the odds of mortality at extreme temperature percentiles (5th and 95 ^th percentiles) to the median temperature at each prison during the study period from 2009-2021. Results: – Of 5, 182 deaths from 2009 to 2021, most were recorded during hot (42%) and cold (41%) months. The cumulative effect of extreme hot (95 ^th percentile) and cold (5 ^th percentile) temperatures was protective, yet imprecise. At a the 2-day lag for extreme hot temperatures and a same-day lag for extreme cold temperatures, we observed increased odds of mortality in federal prison. Conclusions: – Our analysis of mortality in federal prisons yielded different results from previous analyses of mortality in state prisons. Future studies are needed to improve exposure estimation of indoor temperatures and quantify air conditioning access in US prisons.