Comparing COVID-19 between Prisons and the Community

ABSTRACT

Policy-makers need the best available evidence to inform policy decisions during public health emergencies. We illustrate this in the context of temporary prisoner release decisions due to the COVID-19 pandemic. We leverage parolee COVID-19 mortality rates in PA to show that prior estimates of the expected COVID-19 mortality rate for prisoners if released into the community, calculated from community rates, underestimate expected deaths by a factor of over 2.5. These results suggest a substantially smaller return to public health from temporary prisoner release programs than previously estimated and that, in some settings, such programs can increase COVID-19 risk. They also highlight the importance of selecting an appropriate comparison group when estimating counterfactuals to inform public policy.

KEYWORDS:

• Prisons
• COVID-19
• public policy
• counterfactual estimates

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. We focus on COVID-19 mortality rates (as opposed to COVID-19 positive case rates) to circumvent issues with COVID-19 testing variation between prisons and the community.

2. These time-sensitive decisions can vary widely from whether to expand promising correctional, educational, or housing programs to whether and for whom to target stimulus packages.

3. Following Adams et al. (2020) we consider an individual to have a health risk factor if they have been diagnosed with cardiovascular disease, diabetes, respiratory disease, hypertension, or cancer.

4. We use aggregate-level comparisons for this analysis, as opposed to individual-level comparisons, as only aggregate level LTCF death data is provided in Pennsylvania Department of Health reports.