Little is known about the prevalence of COVID-19 in immigration detention facilities run by Immigration and Customs Enforcement (ICE). Yet, conditions of confinement put people at high risk for contracting the virus. New estimates the Vera Institute of Justice (Vera) will publish this month suggest ICE is severely underreporting the prevalence of COVID-19 in detention. Indeed, the true number of people who have been infected in ICE detention may be 15 times higher than official numbers and is still increasing.

As of May 31, 2020, ICE reported testing 2,781 people in its custody; 1,461 positive cases were disclosed at just 61 of the more than 200 facilities it uses to detain people. ICE has not clarified whether the remaining tests are confirmed negative or still pending, meaning a minimum of 52 percent of people tested have received positive test results. Two people have died in ICE custody as a result of COVID-19, and a third person reportedly died of COVID-19 shortly after release from detention.

The number of tests conducted is vastly eclipsed by the number of people who are held in the custody of ICE. Though ICE reports 25,911 people in detention at the time of this writing, more than 58,000 people have passed through detention since ICE’s first known case of COVID-19.¹

In response to an urgent need for better information, researchers at Vera’s Center on Immigration Justice built an epidemiological model to demonstrate how the coronavirus could be spreading through detention centers. Our model is based on methods developed by other teams modeling COVID-19 in congregate settings. The model specifically explores how ICE’s ongoing apprehensions—as well as book-ins to and frequent transfers between immigration detention centers—may be contributing to COVID-19’s spread.

Our approach is a scientific way of projecting what infection rates might be, using the best data available to the public.² This model simulates a 60-day period following a “Day 0,” on which there were no COVID-19 cases in detention facilities.

Our simulation projects that, by the end of an initial 60-day period:

• The number of COVID-19 cases would be 15 times higher than ICE had reported by mid-May (and eight times higher than the number of tests ICE has even administered according to its own accounting).
• Approximately 20 percent of people detained would have become infected with COVID-19. However, the number of cases among people in detention would not yet have peaked. This means that the risk of infection remains high and the cumulative number of cases will be far greater than it is today.
• More fatalities will occur if current trends continue.
• New book-ins and frequent transfers of people who are not being tested contribute to COVID-19’s spread through detention centers across the country.

In our model (what epidemiologists call a Susceptible, Exposed, Infections, Removed (SEIR) model), we assumed that COVID-19 cases originate from ICE book-ins of people whose probability of having been exposed to COVID-19 is constant. In the simulation, the virus spreads to other people detained in that facility and to people in any other facility to which ICE transfers an infected person.

Future versions of the model will expand our predictions from 60 days to 90 days. Our initial simulation is a simplified version of how COVID-19 may actually spread, as it accounts for spread only among people who are detained. The complex task of modeling transmission to and from others who come into contact with people in detention remains the subject of further research.

Vera’s research underscores that ICE continues to withhold critical information from the public, in this case with potentially grave public health consequences.

In an effort to model the spread of COVID-19, we bumped up against many of the same transparency issues advocates have been noting for years.

For example, even as the number of people ICE detains has increased in recent years (before declining in the fall of 2019 due to policies that make it harder for people to enter the country while seeking protections), ICE has not shared detailed patterns of its facility use with the public, whose taxpayer dollars fund these operations. ICE continues to book thousands of people into custody each month, and transfer people across its vast network of facilities, seemingly without proper public health precautions and without sharing detailed information about where people are being apprehended or moved.

ICE has reported COVID-19 cases in only a fraction of the facilities it uses. It is unclear how the agency is coordinating with the many contracted private facilities and local jails it relies on to ensure they are properly testing and protecting people held for ICE. The agency has not revealed how many people in detention have sought to be tested for COVID-19, how it makes decisions about who to test, or how it ensures proper treatment for people with confirmed COVID-19 in its custody—including how many people have been hospitalized or recovered, if this is even being tracked. It does not report how it is ensuring sick staff do not enter its facilities, though it has reported that 44 staff working at detention facilities tested positive for COVID-19.

In short, ICE’s lack of transparency makes it hard to predict the risks confronting people being held for immigration reasons. However, as our model demonstrates, the true scale of the spread of COVID-19 in ICE detention is likely to be shockingly high. ICE has caused preventable harm and continues to endanger people’s lives and wellbeing with its policies and ongoing operations.

¹ We calculated the number of people passing through ICE detention using data from the ICE detention statistics page over time.

²The estimates shown here are the median results of 500 runs of the model simulation. Due to the modeling assumptions necessary in the absence of transparency from ICE, the results should not be interpreted as precise predictions, but indicative of the order of magnitude of the spread of COVID-19 among people ICE has detained.