Skip to Main Content

It’s virtually impossible to get information from states and correctional facilities about why people die in prison.

For more than two years, STAT endeavored to document the number of incarcerated people who died due to complications from hepatitis C, part of a broad investigation into prisons’ failures to prevent avoidable death and suffering related to the condition.


Prison systems fought our attempts at every turn.

This undertaking underscores how easy it is for prisons to hide the true reason why people die behind bars, and how useless the existing data is for determining whether people in prison are dying from preventable conditions. The overwhelming difficulty of the task raises questions about whether and how family members or community advocates could ever use the information to sound the alarm about especially dangerous facilities, help families and friends figure out if a death was preventable, or pinpoint inadequate care for certain diseases.

By law, states are supposed to record the details of every death that occurs in prison. The Death in Custody Reporting Act, a federal law on the books for more than 20 years, requires states to fill out a four-page form that lays out basic information about every incarcerated person’s death, including the name, age and ethnicity of each person who died, and their cause of death.

The Department of Justice uses that data to compile high-level reports about deaths in prison, but that data is often reported at such a macro level that it’s not useful. Those reports, for example, only provide a tally of how many people die of “liver disease,” which could have been caused by everything from hepatitis A, B, and C to liver damage from alcohol use.


When STAT filed public record requests seeking the information, states gave multiple reasons for hiding it. Arizona, for example, said it didn’t maintain death data in the format STAT requested. When STAT sent a follow-up request demanding the exact forms the state already submits to the federal government, the state then said it couldn’t provide those documents because they were owned by the Department of Justice. (Other states provided the exact same forms.)

Arkansas is one of several states that declined to provide STAT with any data at all because the state’s records law doesn’t require it to provide data to out-of-state residents. (Arkansas Gov. Asa Hutchinson wrote the Death in Custody Reporting Act when he was in the U.S. House of Representatives.)

Other states, like Texas and Alaska, told STAT the forms they are required to submit to the federal government simply don’t exist. Texas added that it doesn’t have access to the forms because they’re submitted via a “data upload.”

Iowa said the information was confidential. Others, like Connecticut and Georgia, simply ignored STAT’s request altogether, in clear violation of public records laws.

Some states, like Indiana, shared data that were largely useless: The state only categorizes deaths as “homicide,” “use of force,” “suicide,” “accident,” or “natural causes,” making it nearly impossible to assess the adequacy of the prison medical system there.

In fact, STAT was only able to document the number of people who died of hepatitis C after crafting an agreement with the federal government to analyze the death data submitted by state prisons pursuant to the federal Death In Custody Reporting Act. That data is housed at the University of Michigan and is typically only available to Ph.D. researchers who are approved by the Department of Justice — an option unavailable to the average American. For our research, the DOJ withheld names and other demographic information.

Even that data was highly imperfect. Often, it was not detailed enough to discern whether someone died of hepatitis C. We excluded any such death from our tally, meaning it is likely an undercount of the actual death toll from this disease.

Prisons are only required to submit one cause of death to the federal government. What that means in practice is that a person who died of a hemorrhage due to hepatitis C-induced bleeding is often listed as dying of a hemorrhage, making it impossible to know if that bleeding was caused by something like hepatitis C or a homicide.

The problem is especially acute when it comes to hepatitis C-induced liver cancer. The Centers for Disease Control and Prevention estimates that nearly 50% of liver cancer cases are attributable to hepatitis C, but prisons across the country only report when people die of liver cancer, not what actually caused the cancer. STAT counted 416 deaths since 2014 that listed liver cancer as the cause of death with no additional causes; those are not included in our count of 1,013 hepatitis C-related deaths in prison.

STAT’s tallies also only include deaths from 2014 to 2019 because responsibility for collecting this data was transferred in 2019 to another office in the federal Justice Department, which is known as the Bureau of Justice Assistance.

There are growing signs that the reporting of prison deaths has worsened under the new office. The Government Accountability Office, the federal government’s oversight arm, found earlier this year that 70% of the records submitted to that office in 2021 were incomplete. The Department of Justice itself also issued a lengthy report this year acknowledging it had limited “capacity to collect accurate and complete information.”

The Bureau of Justice Assistance declined to share more recent data.

“We know now that the Department has received underreporting of deaths in custody since late 2019,” a spokesperson said. “The Department is not releasing [death] data from this period as it is taking steps to improve the quality and completeness of state reporting.”

The reporting of deaths in custody is so shoddy that academics and advocacy groups have launched their own projects to cobble together information on how people die in custody.

“The way that we have to do it is frankly unacceptable,” said Lauren Brinkley-Rubinstein, an associate professor of population health sciences at Duke University who heads the Third City Project. “It shouldn’t fall on curious citizens to go dig and find what they can.”

Andrea Armstrong, who compiles and releases data on deaths in Louisiana jails and prisons, noted a number of shortcomings with the federal data, including that it typically is released years after deaths actually occurred, and that the data does not specify the facility or even region of the state where the person died. That, she said, makes it impossible to pinpoint problematic prisons in the state.

Armstrong’s project, which also memorializes people who die by name, also couldn’t be compiled using the federal data, because the federal government doesn’t release identifiable information about those who died in custody. Her team relies on media reports, litigation, and public record requests to get this more detailed information.

“We made that commitment to say the names out loud, and therefore some of the other datasets [wouldn’t work],” Armstrong, a ​​distinguished professor of law at Loyola University New Orleans, explained.

Some of the privacy around deaths in custody is supposed to protect people’s personal health information. Under the federal privacy law, known as the Health Insurance Portability and Accountability Act, information about a person’s health is confidential for 50 years following their death. States like Texas, which report the cause of all prison deaths, are able to disclose this information because their state legislatures have passed specific laws allowing them to do so.

Privacy lawyers say it’s reasonable, generally, for states without clear-cut laws to be cautious about disclosing anything that could be considered protected health information because they could be dragged to court for an invasion of privacy.

“You can’t unring the bell if [private information] leaves and you shouldn’t have let it go,” said Barry Herrin, a health privacy lawyer, who noted officials usually won’t give up potentially protected information unless someone sues for it.

But there are ways to disclose why a person died without divulging identifying information. States like California, for example, release a yearly report outlining how many people die of specific medical conditions in the state, including hepatitis C. That sort of disclosure, HIPAA experts told STAT, would not violate privacy rules.

Create a display name to comment

This name will appear with your comment

There was an error saving your display name. Please check and try again.