Throughout the pandemic, data about infections and deaths in jails have been extraordinarily sparse. Although a handful of large jails have posted dashboards, and state oversight agencies and state courts have mandated reporting in a few states, the vast majority of the approximately 3,000 county and city jails in the United States have operated in obscurity.

A new report released last month by the Department of Justice’s Bureau of Justice Statistics (BJS) contains previously unreleased data showing the spread of COVID-19 in the nation’s jails during the early months of the pandemic. It covers only about a third of the country’s jail jurisdictions and only from March through June 2020, before the winter surges. But because jails have otherwise been such a black box, this report offers the most comprehensive picture yet of the impact of the pandemic in these local facilities.

From the outset, testing has been inadequate.

Unlike prisons, jails — which holds individuals before their trial, and some after being sentenced — experience significant population “churn.” But according to the BJS report, just nine percent of the nearly 2.4 million people entering jails between March and June of 2020 were tested for COVID-19 at admission, even though the positivity rate of those tests was over 11 percent.

This finding — of such a high positivity rate paired with a low overall testing rate — suggests that because of their poor testing practices, jail systems across the country have been failing to detect untold numbers of COVID-19 cases introduced from outside communities into their facilities. It is to combat exactly this phenomenon that the CDC has since updated its guidance to jails and prisons to include testing all incarcerated or detained people upon intake (before March 2021, this recommendation was given as a “consideration” for facilities in communities with “moderate to substantial levels of community spread.”)

The report shows that in jail systems located in counties with higher community COVID-19 rates, more testing was done, but still, we believe, not nearly enough. In counties where at least one percent of the general population had, at some point April 2020, tested positive, just about 20 percent of people admitted to jails were tested and 14 percent of those tests were positive.

There were also striking regional differences in testing rates. In the South, which accounted for half of the country’s jail admissions, about seven percent of the more than one million people who were admitted to jail were tested. Of those tests, over 12 percent came back positive. In the Dallas County, Texas jail, where more than 5,000 people were held on a given day, just six percent of people admitted to jails were tested, and the positivity rate was a whopping 53 percent.

As in prisons, COVID-19 has taken a significant toll on staff in jails.

The report also found a high COVID-19 infection rate among jail staff. During the months from March to June alone, it found five percent of jail staff nationally tested positive for the virus. This data point in particular reaffirms other research findings identifying elevated rates of COVID-19 among staff in carceral settings, including a recent article published in the American Journal of Preventive Medicine that analyzed our data and found rates of COVID-19 among American prison staff were more than triple that among the U.S. population as a whole.

Cumulative infection rates among jail staff were especially high in larger jail systems and in systems located in counties with higher community infection rates. In New York City, then the national epicenter of the pandemic, the jail staff infection rate was just under fifteen percent.

People are dying of COVID-19 in jails, but death reporting is deeply inadequate.

BJS also includes data on suspected or confirmed COVID-19-related deaths. Although it reports 43 deaths among incarcerated people and 40 among jail employees, there are reasons to think this reporting is far from comprehensive.

Even with respect to the four-month period covered in the BJS report, death data were reported by only a subset of jurisdictions, holding approximately 69 percent of the country’s total jail population. During that period alone, therefore, extrapolating from the death rates from reporting jail facilities, it is fair to estimate that somewhere on the order of an additional 20 people died in non-reporting jails.

Furthermore, death tolls in prisons and in communities across the country were substantially lower in the early months of the pandemic than later in the year. In our prison data set, only about 25.2 percent of all COVID-19-related deaths among incarcerated people occurred between March and June 2020. If the same share of jails deaths occurred during this early period, the BJS report would have failed to capture about 128 subsequent deaths among reporting jails and 60 subsequent deaths among non-reporting jails.

Based on these two factors alone, it appears likely that the true death count in America’s jails has been closer to 251 incarcerated people.

Finally, there are reasons to question the overall reliability of the death figures reported by BJS. A recent investigation by Reuters revealed a number of flaws with death data reported by BJS prior to the pandemic, and it is likely that these issues have persisted, if not been amplified, as the number of deaths has grown during the COVID-19 pandemic.

According to that investigation, there are a number of ways by which jails avoid reporting accurate figures revealing the number of deaths in their facilities. Some, Reuters found, misclassify suicides and homicides. In other stances, sheriffs discharged from their custody individuals who were about to die while hospitalized. Upon their death, these individuals were then technically not in the custody of the jail facility, and thus not classified as a jail death.

In its own investigation, Reuters uncovered 59 deaths in 39 jail facilities that went unreported to government agencies and their own investigative team.

Rapid decarceration in jails is possible, but has not occurred on a large enough scale.

In addition to providing a glimpse into COVID-19 data in jails, the report highlights a significant drop in the country’s jail population in the early months of the pandemic. By mid-year 2020, the number of people in jail across the country appears to have been down 185,400 — or about 25% — from the same time a year earlier. That drop, the report notes, which brought the country’s jail population to its lowest levels since 1996, resulted from mostly reductions in jail admissions, but also the increased use of expedited releases.

However, according to the report, even after the early pandemic-related population decreases, there were still more than 500,000 people held in local jails, 380,000 of whom had not been convicted of a crime. As the Prison Policy Initiative noted, even after cutting 25% of its jail population, the United States still incarcerated far more of its own population in jails alone in the middle of the COVID-19 pandemic than peer countries do in any incarceration environment (jail or prison).

Further, since the report’s last date of data collection, jail populations across the country have been on the rise. The report itself documents a low point in April 2020, when the jail population was just about 71% of where it was in June 2019. By June, it had already crept up to 74.8%.

Data published by the Vera Institute of Justice reveal a steady rise in jail population continuing through the rest of the year. Vera has estimated that in the second half of 2020, the national jail population rose as much as ten percent.

The crisis exposed longstanding failures of transparency and the critical need for timely and reliable reporting.

Neither the public health crisis unfolding in jails nor their abject lack of transparency is new. American jails have long had staggeringly high death rates — over a thousand people died in jail in 2016 alone. Nor is the lack of transparency we have seen surrounding those deaths behind bars new — the federal government is extremely slow to release data on the number of jail deaths (the most recent data, from 2016, was released in 2020), and when it does, still does not show in which jail facilities those deaths occured.

It is critical for the federal government to present this data, more so even than for deaths in prisons, because there are thousands of individual jurisdictions to account for, rather than just 50 or so jurisdictions that can report individually. Many of these jurisdictions operate with little to no oversight — 17 states have no formal legal standards or oversight bodies

The pandemic is not over. Jail systems must continue to respond to the threat posed by COVID-19 by reducing admissions, expediting releases, and testing and vaccinating those in their custody. At the same time, however, federal, state, and local officials must make public information about who is sick and dying inside. Health and death data pertaining to people detained in jails is critical public health information and must be treated as such.