Emptying Prisons to Prevent the Spread of Coronavirus Will Save Lives on the Outside, Too

This piece is a commentary, part of The Appeal’s collection of opinion and analysis.

There is a catastrophe on the horizon. COVID-19 is showing up in prisons and jails across the country, and the number of cases inside will soon explode. There have been numerous calls from a range of voices to dramatically reduce the number of incarcerated people in order to avoid a humanitarian disaster. At the same time, people have noted how the flow of people in and out of jails will increase the number of cases on the outside because jails are the perfect environment for transmission, lacking sanitary conditions and space for those on the inside to socially distance.  

The New York Times reports that the Cook County Jail, in Chicago, is now the nation’s largest source of new infections—but the worst is yet to come, both there and in the other jails and prisons across the country. We need to release people before the rate of infection inside gets higher, but politicians have been slow to let people out, and when they have done so, far too few people have been freed. Perhaps what the politicians haven’t yet realized is that holding people in these conditions has dire consequences for unincarcerated people as well. 

A model of the spread of COVID-19 inside prisons from the nonprofit Recidiviz suggests that hospitals will soon be overwhelmed by a deluge of sick prisoners—further burdening already overwhelmed healthcare systems, and preventing sick people on the outside from accessing care when they desperately need it.

Looking at current counts of people in jails and prisons alongside this model, it seems that more than one-third of the public hospital beds in the nation could be in use by prisoners in less than three weeks. This prediction is based on a worst-case scenario model outcome, which assumes that jails and prisons don’t release many more people and don’t take many measures to stem the spread of the disease inside—in other words, if they largely keep doing what they’re doing now. Using up-to-date incarcerated population counts, known rates of infection spread in a confined population, and rates of hospitalization after infection, we can estimate the number of hospital beds that will be needed by prisoners. Predictions look less dire when the number of susceptible people is reduced or the rate of spread is decreased. Decarceration accomplishes both. 

Like many models charting the pandemic’s trajectory in the general population, this one estimates the number of susceptible, exposed, infected, recovered, and dead based on past rates of spread, hospitalization, and death in an analogous situation—in this case, using  data from the spread of the 1918 influenza outbreak in San Quentin. It predicts that in about four weeks the number of prisoners who need to be hospitalized would peak at more than 285,000.

Worst-case model estimates predict that almost two-thirds of hospital beds in Louisiana (8,000 of 12,000) and Georgia (14,000 of 23,000), two-fifths in California (28,000 of 65,000), and more than three-quarters in the most vulnerable state, Arkansas (7,000 of 9,000), will be required by prisoners between 14 and 22 days from now. If politicians decide that the prisoners won’t be hospitalized—a likely scenario given the lack of action so far—there will be thousands of avoidable deaths in some states. Nationwide, judges, governors, and the president will be faced with a choice between calamity inside and risking a collapse of the healthcare system outside.

But politicians can still avoid the worst of this crisis if they immediately release many, many more prisoners. Thus far, the longstanding misconception that released prisoners are a risk to public safety has been a major obstacle standing in the way of more rapid decarceration. By demonstrating the future impact of this developing crisis on hospital populations, it is now clear that by keeping people inside jails and prisons, politicians are creating a much more dire threat to public safety: prisoners will get sick and flood an already overburdened healthcare system. As this happens, healthcare will suffer for everyone, inside and out. 

The model as used here has known limitations. For example, parameters on the rate of infection spread and hospitalization are taken as averages of a range of observed values; using parameters at the high value of the range predicts worse outcomes than reported here, while using the low end of the range predicts less dire outcomes. We can’t use this model to estimate the death rate if people aren’t hospitalized when they need to be. And the rate at which hospitalization is necessary is taken from populations outside jails and prisons. It’s likely that incarcerated people are more vulnerable than those outside due to underlying health issues, poor sanitation and quality of care in normal times. Also, the model does not take gender into account, and many more men, who are more vulnerable to this virus, are incarcerated than women. What is more, the only data we have on the spread of infection comes from a different disease inside one specific prison more than 100 years ago. 

All of that comes in addition to the general difficulty of making accurate predictions about the trajectory of COVID-19, a new disease that the world’s top scientists still know little about. Projections about the spread of the virus in the general population, like the much-cited ones produced by the University of Washington, have changed a lot over time—a reflection of the challenges inherent in using models to predict the future.

Despite these limitations, this model is the best tool we have to predict the effects of the COVID-19 pandemic on prisoners, and it can still help us weigh the effects of different choices. It predicts that Louisiana could reduce its prisoner hospital bed usage to about one-eighth of total beds at peak if it released half of its prisoners now, a move which would help flatten the curve inside by reducing crowding and allowing for more social distancing. This estimate starts from a reduced susceptible population, and also employs a lower rate of spread than the worst case rate of spread. Also important is that this move would delay peak bed usage by about 10 days, giving the state’s hospitals crucial additional time to prepare. An even lower rate of spread could be achieved by releasing three-quarters of those in Louisiana jails and prisons now. That would mean that at the peak need, less than 5 percent of the state’s hospital beds would be used by prisoners. The model predicts that delaying these releases by two weeks substantially reduces the beneficial impact they would have on hospital bed capacity. Releasing half in two weeks would still result in more than two-fifths peak hospital bed usage.

It has never been more obvious that immediate decarceration is needed for everyone’s sake. The more people who are released now, the fewer people in hospitals in a few weeks, and more free people will have access to the healthcare that they will certainly need.

Oliver Hinds is a senior data scientist at the Vera Institute of Justice and a member of End Incarceration. He is a formerly incarcerated researcher working to end incarceration in the U.S.