The Death in Custody Reporting Act of 2013 (DCRA) is meant to produce information that will inform efforts to prevent future deaths in custody. To that end, the law requires states to collect data about people who die in the custody of state and local law enforcement and corrections agencies, and to report that data to the U.S. Department of Justice. Each state’s commitment and effort to fulfill this mandate is an essential part of whether DCRA can successfully yield complete, accurate information about people who die in the government’s custody.
The Latest From POGO
Sent Saturdays
In 2016, during its initial attempt to implement DCRA, the Justice Department announced itwould require each state to submit a written plan detailing how it would meet the law’s reporting requirements. However, those requirements (along with many others) were rolled back and watered down in 2017 before ever being implemented. Restoring the state plan requirement became a central recommendation of the Project On Government Oversight’s 2023 report, co-authored with The Leadership Conference on Civil and Human Rights, about the importance of DCRA implementation.
Not long after we published that report, the Justice Department announced it was restoring the state implementation plan requirement. To remain eligible for federal assistance through the Edward Byrne Memorial Justice Assistance Grant program — a federal program that has awarded over $7 billion to states and localities since 2005 — each state must now submit a written plan outlining the infrastructure and methods they have in place to gather and report data about those who die in custody.
The Justice Department approves the plans, and it published the first batch in early 2024. Every state’s plan was approved and published as of September 2024.
The implementation plan requirement has multiple benefits. It forces states to formalize their compliance efforts and allows the Justice Department to identify and help address shortcomings. And because the plans are public, it allows for comparisons to evaluate which practices seem more or less likely to result in the accurate and complete data that is essential for DCRA to meaningfully inform efforts to prevent deaths in custody.
What Has to Be Reported Under DCRA
DCRA mandates reporting of all deaths in state and local custody. This covers situations ranging from police encounters where a person is not free to leave (such as a traffic stop) through post-conviction imprisonment. In general, deaths in custody can be grouped into three major categories: police interactions, including stops, arrests, and transport; jails and lockups, where people are detained prior to a trial; and corrections facilities, where people are incarcerated after a conviction. These categories involve different agencies: law enforcement such as local or state police departments, sheriffs (who often function as both police and jail operators), other jail operators (which are typically local or county entities), corrections departments (which are state agencies), and private companies with government contracts for incarceration.
DCRA also requires reporting on deaths in federal custody, including people who die during encounters with law enforcement agencies like the FBI, incarceration in federal prisons, and detention in immigration facilities. However, this analysis does not look at the federal component, which is managed separately from the state component and does not have the same requirement for implementation plans.
Challenges in DCRA Compliance
The central challenge of DCRA compliance is ensuring that every death in custody is counted. That means state plans need to be robust enough to capture all required data from the highly fragmented and decentralized criminal legal system, which has thousands of separate agencies and departments often operating with little oversight.
The common thread is that each state designates a central agency to administer its DCRA program, which includes the responsibility for gathering data and reporting it to the Justice Department. Beyond that, there is enormous variation in how states attempt to comply with DCRA.
While the Justice Department has approved each state’s plan, some states are clearly lagging in their DCRA compliance. This resource provides insights and context based on our review of the plans from all 50 states and the District of Columbia.
The inescapable conclusion from reviewing these plans is that DCRA is still not resulting in complete and accurate data about who dies in custody and why. Many states’ plans reveal that they are struggling to create processes to gather data on the full scope of deaths in custody. Some are producing next to no information at all. And even the states that have comprehensive plans on paper are not able to ensure that their data is correct.
Finding 1: Many States Have Agencies Self-Report, with Mixed Results
The most straightforward (though not necessarily most effective) approach to DCRA data collection is for the state agency to rely on all the agencies in the state with custodial power to report deaths in their custody.
In practice, this “self-reporting” method means each agency is responsible for filling out the Justice Department’s data form. Among states that take this approach, some are more proactive than others in notifying agencies when it is time to report. However, a number of the states relying on this method have found that a low response rate undermines the quality of their data.
For example, Rhode Island indicated that its corrections department is the only agency in the state that has responded to requests to submit data, with no responses from any police department or jail.
Idaho noted plans to improve reporting, including having the state police chiefs’ and sheriffs’ associations encourage members to report, but flagged that “the average quarterly response rate has been 30%.”
Indiana sends an email reminder for police departments to report each quarter, but struggles to maintain an accurate contact list and often resorts to checking media reports to identify deaths that departments failed to report.
State Reporting Laws Are Beneficial but No Panacea
Relying on self-reporting seems most successful when there is a state law that requires death in custody reporting.
In their plans, only eight states reported that they had their own laws requiring police, jails, and prisons to regularly report deaths in custody to state authorities. A somewhat larger set of states reported having laws that require a subset of deaths in custody be reported, most often focusing on deaths in jails.
“While the Justice Department has approved each state’s plan, some states are clearly lagging in their DCRA compliance.”
The eight states that referenced comprehensive reporting laws all have plans in place that are sufficient, at least on paper, to capture data on the full scope of DCRA-reportable deaths, without the sorts of categorical gaps or major noncompliance that other states are struggling with.
However, state laws are not a guarantee of complete or accurate data. Some states noted reporting delays and data quality issues. In addition, several state laws and policies did not match up well with DCRA’s requirements.
Texas requires all deaths to be reported to the state attorney general’s office. The state noted in its plan that reporting is frequently delayed, and there is sometimes confusion about which agency is responsible for reporting certain deaths.
California, for instance, uses different categories for race and ethnicity on its state form than the Justice Department uses, so the data collected under its program has to be reclassified when submitted to the Justice Department.
The deaths that must be reported in states with partial reporting laws vary considerably.
In New Mexico, for instance, only deaths in private correctional facilities must be reported.
In Maryland, deaths “involving a police officer” must be reported annually. Maryland’s reporting includes incidents beyond the scope of DCRA.
Finding 2: Some States Collect Data from Health Agencies, but There Are Drawbacks
Separate from state laws that require death-in-custody reporting, some states have laws that require medical examiners or coroners to examine deaths in custody. Some states use records from medical examiners and coroners as an additional source of data for DCRA reporting.
A handful of state plans rely primarily on data from death certificates or what is generated by medical examiners, coroners, or public health agencies rather than requiring the agency that had custody of the person who died to report the necessary information.
Medical Examiners and Coroners Are a Potential, but Flawed, Alternate Source
These records may not capture all DCRA-reportable deaths. For years, media reports have found that medical examiners frequently misclassify deaths in custody, such as by categorizing a violent death as accidental or natural.
Another risk in relying on medical examiner data is that they may not be made aware of in-custody deaths in certain circumstances.
For example, Oklahoma DCRA reporting relies primarily on data from the state’s medical examiner, but while the state medical examiner’s statutory authority requires them to investigate all deaths in prisons and jails, there is no explicit requirement to investigate all deaths that occur during interactions with police.
Kansas law, meanwhile, does not require medical examiners to investigate deaths deemed “natural.”
Finding 3: Some States Effectively Combine Approaches
Other plans combine approaches, using agency self-reporting and public health data to complement one another. Additionally, some states monitor open-source data to supplement reporting. In practice, this typically means using news reports to identify unreported deaths or to check details submitted by agencies.
“For years, media reports have found that medical examiners frequently misclassify deaths in custody, such as by categorizing a violent death as accidental or natural.”
A combined approach may provide advantages over relying on self-reporting or death certificates alone. Even if all agencies report, or medical examiners are aware of all reportable deaths, gathering data from both sources should increase the odds that all necessary information is captured. Using open-source data to inform DCRA reporting is a good additional safeguard, but it cannot replace collecting data from official sources. If media reports alone were sufficient to shed light on every death in custody, there would be no need for a program like DCRA.
Vermont contacts agencies quarterly to report. It also has a detective detailed to the state medical examiner’s office; the detective is trained in DCRA requirements to identify reportable cases the medical examiner is investigating.
South Carolina uses direct reporting from its Department of Corrections and state police, as well as some local police departments. It augments that reporting with information from county coroners and state data on traffic fatalities, and uses a variety of open sources including police press releases, organizations that track deaths in custody, and news reports.
Ohio receives direct reporting on deaths in prisons and jails. State researchers often rely on open sources to identify deaths in police custody, but then obtain official records like autopsies or incident reports from police departments for details. This is preferable to relying on open sources to provide all details, but runs the risk that deaths that are not reported in the media will not be counted.
North Dakota is entirely reliant on media reports for its data, with no official data from agencies. This is insufficient to ensure full and accurate reporting.
Finding 4: Several State Plans Overlook Entire Categories of DCRA Data
Several states’ plans do not provide systems for capturing certain kinds of deaths, resulting in categorical gaps such as deaths in police custody or deaths from particular causes. In these states, deaths that fall into these categories may be caught by open-source reviews or may go entirely unreported.
For instance, several state plans appear to overlook deaths from natural causes, which is concerning both because DCRA requires all deaths in custody be reported and because “natural” deaths in custody often are the result of mistreatment or neglect and could have been prevented.
Other states, which admit no plan for gathering information about deaths of children in juvenile custody — the most vulnerable people who can be caught up in the criminal legal system — deserve particular criticism.
Here are some reporting deficiencies that stand out:
Several States Overlook Deaths in Juvenile Custody
Oregon has no process for gathering information on deaths in its juvenile system, and its reporting law only covers deaths in police custody if officer use of force was the cause of death.
West Virginia does not have a system in place to collect data from police departments or its juvenile corrections system. Its plan seeks to address one of those gaps by outlining steps to begin direct reporting from law enforcement and the state Office of Vital Statistics, but only provided general information about its intent to gain access to information about children in custody.
Several States Overlook Deaths Classified as “Natural”
Kansas state law does not require reporting of jail deaths deemed “natural,” and its plan contains no processes for gathering information on deaths during arrests.
Iowa uses information from the medical examiner system, but deaths in custody “attributed to a natural disease process” do not trigger investigations by medical examiners.
Minnesota relies on records from the state Bureau of Criminal Apprehension as a major source of its data on deaths in police custody. But state law does not require local police to report to the bureau deaths that are not caused by officer use of force, leaving the state reliant on media reports to capture those deaths.
Nevada’s state plan clearly focuses on deaths in carceral facilities, making no specific mention of reporting from police departments.
Kentucky notifies state and local law enforcement that receive subgrants from the state of their obligations under DCRA, but did not have a system in place for collecting data from departments that did not receive those grants.
Finding 5: DCRA Reporting Plans in High-Incarceration States Fall Short
The United States has one of the highest incarceration rates in the world, including both prisons and jails. The incarceration rates of nine states in particular are higher than any country’s but El Salvador’s. The findings from our examination of the three highest of those — Louisiana, Mississippi, and Arkansas — cast doubt on whether their current efforts under DCRA can yield complete or accurate data.
Louisiana relies solely on the state’s roughly 301 police departments, jails, prisons, and other facilities to self-report deaths in their custody. While the state is making some efforts to increase awareness of the reporting requirement, it is unaware of the response rate and cites timely reporting as a significant challenge.
Mississippi likewise relies primarily on self-reporting. There is an agreement between the Division of Public Safety Planning, which manages DCRA, and the state corrections department to regularly provide data on deaths in the state’s six prisons. However, there is no similar agreement for reporting from jails or other lockups, and the state’s own plan indicates that there is widespread resistance to reporting from police and sheriffs.
Arkansas’s plan has the most potential. It has recently developed a system for agencies to self-report deaths in custody, and is also planning to coordinate with the State Crime Lab, which houses the medical examiner. This is a promising component, as state law requires the state medical examiner or a county coroner to investigate deaths in police, jail, and prison custody. However, it remains to be seen if those changes will be fully implemented or succeed in practice.
Wyoming relies primarily on self-reporting from police departments. The Department of Criminal Investigation also supplements agencies’ self-reported data if it has information from its own investigations into deaths in local custody, but these investigations are not mandatory. The small number of agencies in the state may make self-reporting more viable than in larger states.
Montana has the most robust plan of the three, using multiple data sources for deaths in police custody. The state receives quarterly reports on death certificates that indicate law enforcement involvement, and the state Division of Criminal Investigations also searches state data for additional instances. When the state identifies unreported cases, it contacts the relevant agency.
New Mexico is in the process of rebuilding its reporting infrastructure. Until 2023, the state received aggregated data on deaths in custody from the State Police, Department of Corrections, Department of Health, and state medical examiner. However, it indicated in its plan that it is switching to direct outreach to each law enforcement agency in the state; it is unclear if it will continue to use medical records as an additional source.
Studying these plans is valuable. State officials should be aware of the practices that represent more robust approaches to DCRA compliance. If the lack of a state reporting law is a major impediment to compliance, officials or activists should approach state legislators to address that. If a state health department has records that could serve as an independent data source, the agency administering DCRA collection should seek a data-sharing agreement.
It is clear from comparing the plans that some states seem better equipped to fully meet the DCRA reporting requirement than others. It is safe to conclude that states that have their own reporting laws, have formal data-sharing agreements among agencies, or leverage data from health agencies or medical examiners are in a stronger position to gather accurate and complete data.
As POGO was preparing this analysis, the department did take an important step forward, publishing some initial tables showing aggregated data from DCRA. However, because the Justice Department has yet to release DCRA data broken out by state, it is impossible to evaluate how well any state's plan translates into results.
While this is an important transparency measure, even the department itself recognizes that, until states improve reporting, we cannot have a full understanding of deaths in custody. In fact, the tables published by the Justice Department on its own website includes a disclaimer that the department makes "no representations about the completeness or accuracy of this information." While having states complete implementation plans is a step in the right direction, the fact remains: Nobody in the United States knows exactly how many people die on our government’s watch.
Related Content
DOJ Can and Must Improve Data on Deaths in Custody
