ACLU advances model to prevent COVID death surge in prisons and jails

OKLAHOMA CITY — COVID-19 could claim the lives of approximately 100,000 more people nationwide than current projections stipulate if jail populations are not dramatically and immediately reduced, according to a new epidemiological model released yesterday by the American Civil Liberties Union and academic research partners.
The findings indicate that even if communities across the United States continue practicing social distancing and following public health guidance they will still experience much higher death rates if no substantial action is taken to reduce jail populations.

“We are likely facing massive loss of life — both in Oklahoma jails and in our communities — if dramatic steps aren’t taken to reduce our incarcerated population,” said Nicole McAfee, ACLU of Oklahoma Director of Policy and Advocacy. “Mass incarceration was a major public health crisis before the outbreak of COVID-19, but this pandemic has pushed it past the breaking point.”

In some analyses, the U.S. is the largest incarcerator in the world. About 40 percent of all incarcerated people suffer from at least one chronic health condition, such as asthma or diabetes. This means the U.S. faces a unique challenge in mitigating the spread of COVID-19 and is likely facing a much higher death count than models based on data from other countries predict.

For this study, the ACLU model pulled data from more than 1,200 midsize and large jail systems around the country, whose surrounding communities account for 90 percent of the U.S. population. It found that keeping people out of jail saves lives — both inside the jail and in the surrounding community.
Other key findings from the model include:

    • If a model that doesn’t account for jails predicts the U.S. death toll will be 100,000, the ACLU model shows that that projection undercounts deaths by 98 percent. Actual deaths, once accounting for jails, could be more than double, rising to 200,000, the study’s analysts conclude.
    • Implementing swift, bold reforms to reduce arrests by 50 percent can save 12,000 lives in jails, and 47,000 lives in the surrounding communities.
    • Aggressive action and policy change could save as many as 23,000 people in jail and 76,000 in the broader community if arrests end for anything but the five percent of crimes defined as most serious by the FBI — including murder, rape, and aggravated assault — and double the rate of release for those already detained.
    • States that have begun to reduce their jail populations are quantifiably saving lives. Colorado, for example, has so far achieved a 31 percent reduction in jail population. The ACLU analysis/model found this likely will save 1,100 lives — reducing total lives lost in the state by 25 percent.
    • Delaying action for a week could mean a difference of 18,000 lives lost in the U.S., the ACLU analysts believe.

“The prevailing epidemiological models largely fail to take into account our incarceration rates and the complete absence of social distancing in our jails — which is why we had to build our own model,” said Lucia Tian, chief analytics officer, ACLU. “We can’t save our community while ignoring our jails.”

The original model was developed by Dr. Nina Fefferman at the University of Tennessee, Dr. Eric Lofgren at Washington State University, and Dr. Kristian Lum from the University of Pennsylvania, in collaboration with Aaron Horowitz and Brooke Madubuonwu of the ACLU’s data analytics team, experts from the ACLU and other corrections organizations contributed expertise.

Recently, the ACLU of Oklahoma submitted a draft Executive Order to Governor Kevin Stitt 
( with recommendations on steps to take to address concerns around prison and jail populations. 
Additionally, the group joined eight other organizations in a joint statement ( to encourage swift action to manage the serious threat of a COVID-19 outbreak in Oklahoma Corrections facilities, and are continuing the push for commutations and a medical needs docket.

The white paper with key results outlined can be found here ( ).
The original academic paper on Allegheny County can be found on MedRxiv, here:
( ).