Outbreaks of Covid-19 Variants in Prisons: A Mathematical Modeling Analysis of Vaccination and Re-Opening Policies

Ryckman, T. S.; Chin, E. T.; Prince, L.; Leidner, D.; Long, E.; Studdert, D. M.; Salomon, J. A.; Alarid-Escudero, F.; Andrews, J. R.; Goldhaber-Fiebert, J. D.

Background: Residents of correctional facilities have experienced disproportionately higher rates of SARS-CoV-2 infection and Covid-19-related mortality. To protect against outbreaks, many prisons and jails imposed heavy restrictions on in-person activities, which are now beginning to lift. Uncertainty surrounds the safety of these moves. Methods and Findings: We obtained system-wide resident-day level data for the California state prison system, the nation's third largest. We used the data to develop a transmission-dynamic stochastic microsimulation model that projects the impact of various policy scenarios on risks of SARS-CoV-2 infections and related hospitalization among residents after an initial infection is introduced to a prison. The policy scenarios vary according to levels of vaccine coverage, baseline immunity, resumption of activities, and use of non-pharmaceutical interventions (e.g., masking, physical distancing). The analyses were conducted across 5 types of prisons that differed in their residential layouts, security levels, and resident demographics. If a viral variant is introduced into a prison that has resumed pre-2020 contact levels, has moderate vaccine coverage, and has no baseline immunity, 23-74% of residents are expected to be infected over 200 days. High vaccination coverage coupled with use of non-pharmaceutical measures reduces cumulative infections to 2-54% of residents. In prisons consisting mostly of dormitory housing, even with high vaccine coverage and non-pharmaceutical interventions, resumption of in-person activities is associated with substantial risk, unless there is high baseline immunity (e.g., [≥]50%) from prior outbreaks. In prisons consisting mostly of cell housing, <10% of residents are expected to become infected, even with no baseline immunity. However, hospitalization risks are substantial in prisons that house medically vulnerable populations, even for prisons consisting mostly of cells. Risks of large outbreaks are substantially higher if there is continued introduction of infections into a prison. Some findings may not be transportable to other carceral settings, and our assumptions regarding viral variants will not be accurate for all variants. Conclusions: Balancing the benefits of resuming normal in-person activities against the risks of Covid-19 outbreaks is a difficult challenge for correctional systems. The policy choices are not strictly binary. To protect against viral variants, prisons should focus on achieving both high vaccine coverage and maintaining widespread use of non-pharmaceutical interventions. With both in place, some prisons, especially those with lower room occupancy that have already had large outbreaks, could safely resume in-person activities, while continuing testing and measures to protect the medically-vulnerable.