DCRP Part A_1112016

18

SUPPORTING STATEMENT

The Bureau of Justice Statistics requests a new clearance for the prison portion of the Deaths in Custody Reporting Program (DCRP). The DCRP has a current clearance (OMB clearance number 1121-0249), which is valid through December 31, 2016. However, BJS is proposing the three components of the DCRP, prisons, jails, arrest-related data (ARD), be fielded under separate clearance numbers. The jail portion of the DCRP will become part of a larger jail collection that will also include the Annual Survey of Jails and the Survey of Jails in Indian Country collections (request submitted under OMB clearance number 1121-094). The ARD portion of the DCRP has been inactive since 2014 due to data quality issues, and BJS will seek a separate clearance for the ARD in 2016.

The DCRP was established in 2000, after the passage of the Death in Custody Reporting Act (DICRA) of 2000 (‘2000 DICRA’, P.L. 106-297). BJS began collecting data on deaths in state prisons in 2001. The prison portion of the DCRP collects data on prisoners who died while in the custody of a state prison.

The DCRP prison collection is a national database containing information about each death occurring in state prisons. Individual-level data about the criminal justice system’s involvement with decedents are not available in other national databases on mortality. The DCRP prisoner data are part of BJS’s core correction statistics, as they contribute significantly to BJS’s mission of describing transitions and movements of offenders through the criminal justice system. In the case of the DCRP, it describes a subset of release from correctional populations, (e.g. deaths).

Prisoner death data are submitted annually to the DCRP by state department of corrections, which provide individual-level data for each prisoner who died while imprisoned. All 50 departments of correction have submitted data to the DCRP since 2001.

The prison portion of the DCRP collects data on the following elements:

• Aggregate counts - yearend counts of the total number of custodial prisoner deaths;

• Identifying characteristics - name of inmate, name and location of the correctional facility involved;

• Demographics - date of birth, date of death, race and Hispanic-origin;

• Correctional characteristics - date of admission to the correctional facility, offenses for which the inmate was imprisoned, whether the prisoner stayed overnight in a mental health facility since admission, whether the death was due to a pre-existing medical condition, and what kind of medical treatment was provided for the underlying pre-existing condition;

• Characteristics and circumstances surrounding the death - cause of death, autopsy status, location of death, location of the incident that contributed to unnatural deaths (e.g. suicides, accidents, and homicides), and time of death;

• Any other relevant information - the form has a ‘notes’ section that allows the respondent to record any additional information.

A. Justification

1. Necessity of the Information

In recent years, the size, costs and societal impacts of prisons in the United States has become a national issue. State prisons currently hold 86% of the prison population and 61% the entire incarcerated population in the United States.^1,2 The prison population has grown 7% since 2001, and although growth has slowed over the last 5 years, the prison population increased between 2012 and 2013, before decreasing by 1% in 2014.

While the majority of prisoners will eventually be released at the end of their sentence, a small subset, approximately 3,200 prisoners annually, are released via death. The aging of the prison population and the cost of caring for an older inmate population have become an increasing priority for correctional administrators. The majority of prisoner deaths (90% annually) are due to natural causes including heart disease and cancer.³ The prisoner mortality rate has increased 13% since 2001.⁴

The DCRP was originally authorized by the Death in Custody Reporting Act (2000 DICRA) of 2000 (P.L. 106-297, see appendix C, attachment A). The 2000 DICRA was a Congressional mandate, which encouraged state prisons, state and local law enforcement agencies and local jails to report quarterly to the Department of Justice information pertaining to the circumstances of each death occurring in the process of arrests or while offenders were in custody. The DCRP prison collection began in 2001 and has continued uninterrupted since.

The legislation was motivated by a focus on conditions in state prisons and local jails and the problem of inmates dying in custody. Congressional interest in oversight of this issue arose primarily because of the rising tide of expensive wrongful death lawsuits brought in relation to these deaths. Press reports in the 1990s concerning prison abuses and deaths of those incarcerated being attributed to suicide led Congress to develop the 2000 Act in response to this problem.

After the 2000 DICRA expired in 2007, reporting to the DCRP was voluntary for about 7 years. BJS collected, analyzed and reported on prisoner deaths, despite the lapse, as public concerns about the safety and humane treatment of suspects, defendants, and offenders while in contact with or under the control or supervision of criminal justice agencies remained. Prisoner deaths continued to receive national attention, with media outlets like the Washington Post and the New York Times writing a series of stories on the topic. Organizations, such as Human Rights Watch, have long maintained that the failure of criminal justice agencies to maintain safe and secure detention and confinement facilities violates the human rights of persons to be treated with respect and to be free from cruel or inhumane punishment.

In 2014, a number of lawsuits involving inmate deaths, in addition to a rash of high-profile police-involved homicides increased public interest. Coupled with the perception that data were not being collected or were not sufficiently transparent, culminated in the passage of the Death in Custody Reporting Act (P.L. 113-242) of 2013. (See Appendix C, Attachment B). The Act, which reauthorized DICRA’s reporting mandate, became a law on December 18, 2014. The 2013 DICRA (‘DICRA 2013’) maintained the major components of the original 2000 law.

The DCRP is the only national database that can inform the issue of mortality in prisons. Prior to the DCRP, the public’s knowledge of prisoner deaths was limited. Prison mortality data were collected annually, in aggregate, at the state level via the National Prisoner Statistics program (OMB clearance number 1121-0102), and were typically reported nationally as a type of prisoner release. Furthermore, the data were limited to sex-based summary counts and a limited number of manners of death (e.g., homicide, suicide, natural causes, AIDS-related). In addition, the level of aggregation precluded analysis of subpopulations, e.g., age, race, and ethnicity.

The DCRP prison collection provides a unique opportunity to study inmate mortality in-depth because it collects individual-level data on deaths in state prison. With information collected from the DCRP, BJS can track changes in mortality rates of imprisoned persons and improve the public’s understanding about deaths of persons under the control of the correctional system by reporting on cause of death and the circumstances surrounding the death. For example, through the DCRP it is known that the leading cause of death in prisons is cancer, followed by heart disease, and that the majority of prisoners who die in custody have medical conditions that pre-date their arrival into prison, as opposed to contracting or developing a fatal disease while in custody⁵.

Detailed information about prisoner deaths cannot be obtained through other, existing collections. The National Center for Health Statistics (NCHS) mortality files and the National Death Index (NDI), which rely on death certificates files in the U.S., do not identify the criminal justice system involvement of the deceased. Other sources of data on mortality in the U.S. prison system also have significant limitations (see item 4 below, Efforts to Identify Duplication).

BJS authority for the DCRP comes from the Omnibus Crime Control and Safe Street Act of 1968, as amended (42 U.S.C. 3732), which established BJS and authorized it to collect and analyze statistical information concerning the operations of the criminal justice system at the federal, state, tribal and local levels §3732(c)(4) (see Appendix C, attachment C). The U.S. Attorney General’s Office initially tasked BJS with collecting custodial death data after the passage of the 2000 DICRA. With the passage of the 2013 DICRA, BJS will continue to collect DCRP data as it has done since 2000.

Under Title 42 of the United States Code, Section 3789g (see appendix C, attachment D) BJS collects DCRP prisoner data for statistical purposes only, does not release data pertaining to specific individuals in the DCRP, and has in place procedures to guard against disclosure of personally identifiable information. DCRP data are maintained under the security provisions outlined in U.S. Department of Justice regulation 28 CFR §22.23, which can be reviewed at: http://bjs.ojp.usdoj.gov/content/pub/pdf/bjsmpc.pdf.

2. Needs and Uses

The prison portion of the DCRP is part of BJS’s effort to measure conditions of confinement and mortality throughout the criminal justice system and to integrate the analysis of mortality as a key criminal justice statistic. The DCRP enables BJS to track and report on all causes, and cause-specific mortality in prison populations. As a result, BJS annually tracks and reports on changes in mortality and assesses whether the changes in mortality arise from changes in the demographic and crime-type composition of populations or changes in the age, sex, race, or offense-specific mortality rates.

Recurring uses of the DCRP prisoner collection

Through the prison portion of the DCRP, and in conjunction with its other statistical collections BJS obtains and uses data that are necessary to understand the incidence and prevalence of mortality occurring while inmates are in prison. Through the DCRP collection, BJS produces statistics that provide policymakers, public health officials, and correctional administrators with information to track changes in mortality in prisons.

BJS describes the circumstances surrounding deaths, such as pre-existing medical conditions that highlight how changes in prisoner attributes may impact the overall prison mortality rate. In addition, this information informs key stakeholders on the medical needs of the prison inmates.

Since 2013, BJS has reported 2 and 3-year moving averages (depending on the number of years in the reference period) by cause of death and inmate characteristics, which led to the finding that the prevalence of prisoner cancer deaths began to increase in 2008.⁶ In contrast, mortality due to liver disease, which is the third leading cause of prisoner deaths, remained flat, with the liver mortality rates being nearly equal between each interval. By decomposing changes in overall mortality rates into compositional and group- and cause-specific rates, BJS can identify sources of changes in mortality rates.

The instruments used to collect the DCRP data include the following forms (see appendix C, attachment E):

• NPS-4: Deaths in Custody, 2015 Annual Summary of Inmate Deaths in State Prisons

• NPS-4A: Deaths in Custody, 2016 State Prison Inmate Death Report

Manner and cause of death are the cornerstone of all mortality studies, and the DCRP enables BJS to produce annual reports on how prisoners die. With data elements on sex, race/ethnicity, age at admission and age at death, BJS makes comparisons across demographic subgroups and standardizes the prisoner mortality rate to the U.S. resident population. The relationship between criminal behavior and mortality can be studied by comparing offense type and cause of death. For instance, BJS can assess any mortality differentials between violent offenders by type of offense (e.g. homicide offenders versus rape offenders) using the DCRP data. Medical treatment variables, including mental health treatment, give BJS a sense of the health care provided to inmates prior to death.

Information about where within the prison system these deaths occurred, e.g. facility name and location of death, permits comparisons among facility types and linkage to the census of prison facilities. Date of admission to the facility are used to calculate time served before death and can be used to calculate hazard rates, which address the problem of exposure time in prisons. The location and time of death allow BJS to describe the circumstances surrounding unnatural deaths (e.g., suicides, accidents and homicides) in prisons by location and the time of day.

The pre-existing medical conditions and medical examiner questions allow BJS to examine differences in the types of deaths that occur in prison and those in jail, and the rate of autopsy for different causes of death. Specifically, two-thirds of prisoners are autopsied whereas nearly all jail inmates (95%) are autopsied⁷ .

BJS has published special topic reports on medical causes of death in state prisons, in addition to an annual series on trends in correctional mortality. BJS has analyzed several years of data to describe in depth the nature of mortality in these settings. Since producing its first mortality report from the DCRP in 2005, BJS has covered a variety of mortality-related topics, including suicides and homicides in correctional facilities, natural deaths in prisons, as well as routine statistical tables highlighting recent trends in mortality among jail and prison inmates. These reports are available to the public through the BJS website, at http://bjs.ojp.usdoj.gov/index.cfm?ty=dcdetail&iid=243#Publications_and_products.

Proposed improvements and expansion of the DCRP

The following sections describes BJS proposed changes within the next three years:

Cause of death item revision

The White House Office of Science and Technology Policy and Domestic Policy Council working group on medicolegal death investigation data stated during their October 2015 meeting that improving data quality for underlying cause of death is a top priority in improving understanding how and why persons die in custody.⁸ BJS is proposing to revise the current cause of death item to more closely resemble the standard U.S. death certificate to better capture final and contributing/underlying causes of death. The goal of the revision is to not only improve how leading causes like cancer and heart disease are reported, but to improve data quality for less common causes of death among inmates, such as AIDS, homicide, suicide, and accidents (including accidental intoxication). For example, AIDS-related deaths are currently captured via a check box and no additional information is captured.

BJS is seeking to test the feasibility of revising the cause of death item so it will more completely record both final and contributing causes, which is in line with death certificates (see appendix C, attachment F). BJS will address the proposed work for improving cause of death for the DCRP in detail in BJS’s generic clearance (OMB clearance number 1121-0339). BJS proposes to pilot the revised cause of death item to assess respondent understanding of the item, ability to answer a revised item, and potential burden. If successful, the revised item will appear on the 2018 death forms.

The Committee on National Statistics (CNSTAT) of the National Academy of Science reviewed BJS programs and data collections in 2009 and released a report that made several suggestions for corrections data collection.⁹ Of these, the suggestions of particular relevance to the prison portion of the DCRP include (1) facilitate linkage to existing datasets, and (2) expand coverage to include prison reentry and recidivism issues.

BJS has worked to implement these suggestions and proposes the following to improve the DCRP:

Linking DCRP records to the National Death Index (NDI)

BJS intends to continue work first undertaken in 2013 to match DCRP prison death records to the NDI. BJS linked DCRP 2007, 2008, 2009, and 2010 death files to the NDI to determine how many DCRP deaths had a matching record in the NDI and to assess cause of death as reported in both collections. Nearly all (94%) DCRP records had a match in the NDI, and the final cause of death was accurately reported in most cases. However underlying causes were often missing in DCRP cases because the DCRP respondents listed only the final cause of death.¹⁰

Appending U.S. Census Bureau codes

The 2013 DICRA requires the Department of Justice to produce a report on how facility-level characteristics of prisons may affect prisoner mortality rates. Currently, DCRP does not have a unique identifier for each prison reporting a death. A unique identifier has not been necessary to produce state and national level statistics on prisoner deaths for BJS prison studies.

In 2013, BJS attached the 21-digit Governments Integrated Directory (GID) code, as established by the U.S. Census Bureau, to all the DCRP files spanning from 2000 – 2013. The GID code is a unique prison facility identifier. This work was undertaken in order to link the DCRP death files to the Census of State and Federal Adult Correctional Facilities OMB clearance number 1121-0147). The goal is to match prisoner deaths to the facilities where the inmates were housed prior to death. The process was significantly complicated because it was done retrospectively and the assignment of the GID was contingent the facility name in DCRP. Since facility name is a descriptive literal entered by the DOC respondent, there is variation in how facility name is reported across the years. As a result, we were unable to assign a GID code in about 4% of death cases, or about 1,700 deaths. By obtaining the GID from respondents as part of the data collection, BJS can assign the GID at the time of data processing in the future to produce more accurate data related to facility location.

Other proposed changes to the DCRP prison collection include minor modifications to instructions that add clarity or assist in completing inmate death records.

Use of the DCRP by other entities

Other entities rely on the DCRP data for research, planning, and programmatic purposes. DCRP data and statistical reports are used by the U.S. Department of Justice, U.S. Congress, local, and state correctional administrators, public health officials and practitioners, researchers, and special interest groups.

The Office of Justice Programs (OJP), within which BJS is located, has devoted effort to coordinate the work of all of its bureaus on a suicide prevention effort. The Federal Partners in Suicide Prevention, a federal working group sponsored by the Center for Disease Control and Prevention and Health and Human Services in order to report on suicide awareness and prevention in resident and incarcerated populations. The group is particularly interested in incarcerated populations because the crude suicide rates for prisoners is higher than the crude suicide rate for the U.S. resident population. This difference, however, disappears once the rates are adjusted to account for age, sex and race difference. This adjustment would not have been possible without data from the DCRP.¹¹

Congress used BJS reports as part of the justification for the passage of 2013 DICRA. The legislation’s main sponsor cited BJS statistics on the decline in mortality in prisons, arguing that with detailed statistical data, policymakers at the local, State, and Federal levels can make informed judgments about the appropriate treatment of prisoners and develop ways to lower the prisoner death rate.

Additional examples of groups and individuals who have used DCRP to address policy issues related to mortality include -

Public Health, Legal and Justice Policy Researchers:

• Dr. Ingrid Binswanger of the University of Colorado-Boulder used DCRP data to publish an article on smoking cessation policies and cancer deaths in prisons.¹² BJS Visiting Fellow Dr. Christopher Wildeman of Cornell University used DCRP and NCRP data to analyze state-level differences for prisoners and their resident population counterparts after controlling for sex, age and race-ethnic origin differences. Both Drs. Binswanger and Wildeman will continue their work with the DCRP death data (see section 16. ‘Plans for Publication) in 2016.

• Three doctoral candidates have received enclave licenses from the University of Michigan’s National Archive of Criminal Justice Data to use DCRP prisoner data in their dissertation work.

• The American Psychiatric Associations’ American Psychiatric Publishing Textbook of Suicide Assessment and Management, Second Edition (2012) used DCRP data in a section on suicides in special populations.

• Psychiatrists from several teaching hospitals, such as Bellevue, St. Vincent’s and St. Luke’s Hospitals in New York City, have requested data on suicides in prisons (and jails).

• Staff at NCHS reported that the DCRP collection is important and useful for understanding certain types of mortality that they cannot address with their national mortality files. (See item 4 below, “Efforts to Identify Duplication.”)

• The National Association of Medical Examiners (NAME) has endorsed the DCRP as being the sole source of identifying inmate deaths. NAME used data from the DCRP to urge members to revise state death certificates to indicate whether the deceased had recently been in law enforcement or correctional custody. Washington, DC is planning to revise its death certificate accordingly, with the expectation that other states will follow suit.¹³

Correctional administrators and researchers

• Three Departments of Corrections recently requested DCRP data regarding inmate homicides (Georgia) and suicides (Wisconsin and Mississippi) to inform annual reporting.

• The Washington State Department of Health routinely incorporate DCRP tables in its own annual reports.

• The Oklahoma State Department of Health Injury Prevention Service uses DCRP tables in regular “Injury Updates” as part of its Injury Surveillance Program.

3. Use of Technology

The current DCRP web reporting system includes a limited set of online edit checks to identify data entry errors by respondents. BJS has also implemented computer methods for reviewing and editing the entered data in more detail. A set of decision rules are coded and the code is run against the data to identify out-of-range or erroneous values and to assess the impacts of out-of-range values on quantities to be estimated. These methods are used to make decisions about priorities for follow-up contact with respondents.

Based on analytics and comments obtained from respondents during follow-up, the bulleted items that follow are some of the ways in which BJS has used information technology to reduce burden and improve data quality, agency efficiency and responsiveness to the public.

• Real time, “always on” data collection. With the start of a calendar year, respondents will have access to the web collection instrument and can report their data as soon as the 2016 data collection year opens on the DCRP website. Historically, the web collection has a ‘soft’ opening in December, so respondents can begin entering records before the official opening of the collection year.

• Improvements in identifying the reporting year. Because the DCRP collections are “always on,” the Web site will include buttons that allow DCRP respondents to select the relevant year and access multiple years (including previous years) if needed.

• Pre-filled forms. Forms are prepopulated with respondent contact information, eliminating the need for respondents to enter this information unless there has been an update to contact information.

• Improvements in survey flow. The online data collection tool follows the paper forms, but enhancements to the web tool facilitate navigation through the form. Respondents will be led through the items in smaller segments rather than scrolling through the items on a single screen, minimizing the possibility of them inadvertently skipping an item. This simultaneously enhances data quality while reducing the burden of future data quality follow-up.

• Enhanced capacity to add death reports. Respondents are able to create new blank death records for data entry simply by choosing an “Add a Death Report” option button. This is especially helpful for prisons, which often submit multiple deaths at a time.

• Enhanced capacity to review existing death reports. Respondents are able to easily locate existing records based on a combination of identification criteria (e.g., date of death, date of birth, etc.) or by a list of prisoner names.

• Timeout warning so that important data are not lost. Warnings are sent to respondents to if a web session is about to automatically timeout due to inactivity. This warning prevents data loss and eliminates re-entry of data.

• Real-time prompts that alert respondents of potential errors. The functionality of the Web forms alerts agencies to potential data problems. This includes soft prompts when respondents report improbable values on the prison death forms, ex. A birth year greater than admission to prison year. This systems reduces data errors and item non-response.

• On-screen reporting capabilities. Upon completion of each form, respondents receive an on-screen report that summarizes their responses. This enhancement was designed in response to respondents’ interest in wanting to assess the completeness of their submission while allowing them to review and edit their entries prior to final submission.

• Explicit confirmation of form completion following online form submission. Upon submitting their data, respondents receive confirmation that their submission is complete for the relevant reference year.

• Paper versions of submissions for web respondents. Many respondents have expressed that, while they prefer to use the Web-based tool to enter their data, they also need paper copies for documentation. As result, the web-based tool automatically generates .pdfs of completed survey forms that can be printed or saved.

• Real-time reporting to data collection agent of errors encountered by respondents. In 2009, DCRP data collection agent RTI introduced an error log, which notifies RTI of errors encountered by respondents as they maneuver within the DCRP web site. This allows RTI to identify and correct systemic issues, which in turn, has resulted in increased user satisfaction with web reporting. (This is a behind-the-scenes enhancement that does not affect the content of the instrument.)

• Continual additions to the frequently asked questions (FAQs) document. The FAQ is available on the public-facing DCRP Web site (http://bjsdcrp.rti.org) and can be accessed without a user credentials. BJS and RTI modify the content in response to evolving needs.

As outlined in the 2012-2015 DCRP OMB package, BJS introduced a plan to phase out the use of paper forms. Prior to 2012, respondents received a mailing packet that included a cover letter, reporting instructions and copies of the survey forms. Unlike earlier years, paper surveys were excluded in the annual mailings announcing the start of the 2012 data collection, although respondents still received a cover letter and accompanying instructions. The goal of the experiment was to encourage respondents to report their data via the DCRP web tool. Immediately post-experiment, web-based data submissions increased 10% for DOC DCRP respondents. By 2013, 43 DOCs submitted their data via the web tool, and four DOCs submitted their files in an electronic format (e.g. Excel, Access). As a result, the paperless model became permanent for the DCRP.

BJS will expand the paperless feature further in 2016 by sending half of DOC respondents their survey invitation by e-mail instead of the traditional hardcopy mailing. Similar to the current system, respondents will be given the option to request a hardcopy mailing. Respondents in this group who have not responded by the time of the standard replacement mailing (typically in mid-March) will receive a hardcopy mailing.

The e-mail sent to agencies in the truly paperless group will resemble the normal cover letter, i.e., it will be on electronic OJP letterhead, and it will include a URL to the web site with login credentials. Additionally, URLs will be enclosed in the invitation e-mail which will allow the agency to access a special Web page providing electronic versions of all typical enclosures in the January mailing (e.g., reference year reporting instructions, DCRP handout).

Analysis will assess the impact on cost, unit response rate, time/speed of response, and mode of response. If successful, the email notification model will take the place of paper mailing packets.

The online system also allows BJS’s contractors to generate weekly progress reports which allows BJS to assess response rates on a weekly basis and is able to determine the completeness of each DOC. These reports allow BJS to have a real-time look at the progress of data collection and can identify response rate issues early in the collection cycle.

4. Efforts to Identify Duplication

The DCRP is not duplicated by any other federal agency or program. While there are other sources of mortality data related to the topic of prisoner deaths, none are as comprehensive as the DCRP. Since the beginning of the DCRP, BJS has undertaken efforts to identify other national data collections that could be redundant with the DCRP. While states like California, Texas and Washington have state-level laws that mandate the collection of this information and the Arizona DOC issues a press release in the event of a prisoner death, the DCRP is the only national database that collects prisoner deaths for all state correctional systems in the United States.

BJS has identified two national databases with death data, but there are significant differences between these systems and the DCRP. The national collection systems that capture death data are the NCHS’ NDI (as mentioned in section 2, under ‘Necessity of the Information), and the National Violent Death Reporting System (NVDRS).

The NDI captures all deaths in the United States, as it is a clearing house of death certificates filed in the U.S. However, with a few locality-based exceptions, e.g., Washington, D.C., death certificates do not indicate whether the deceased had been in correctional custody and in turn, the NDI is unable to independently identify persons who were in custody at the time of their death.

The NDI is a subcomponent of the National Vital Statistics System, and has total coverage of known deaths in the United States. The NDI is based on death certificates. However, death certificates currently do not have a flag or other indicator identifying custodial deaths. BJS has had discussions with both NCHS and NAME about the possibility of adding such an identifier to the U.S. Standard Death Certificate, and while states plan to incorporate such a flag on an ad hoc basis, any proposed national changes have to be approved by the World Health Organization and changes are proposed and voted on decennially. Per NCHS, the next opportunity will not present itself for several years.¹⁴

The Center for Disease Control and Prevention’s National Violent Death Reporting System (NVDRS) tracks homicides and suicides in 32 states in the U.S. The NVDRS is a state-based surveillance system that triangulates data from death certificates, medical examiner/coroner reports and police reports to create a database on violent deaths. The NVDRS excludes deaths by suicide or homicide in correctional settings in its reporting, but even if these deaths were included, they would only be capturing less than ten percent of prison deaths in two-thirds of states.

5. Impact on Small Businesses

Not applicable. This statistical collection does not involve small businesses or other small entities.

6. Consequences of Less Frequent Collection

The primary concern of a less frequent collection is that it would make respondents noncompliant under the DICRA 2013. Under DICRA 2013, states are required to report inmate death data to the U.S. Attorney General on a quarterly basis (BJS will fulfill this requirement by pulling the information from our data quality reports rather than burdening respondents). Collecting data on a less frequent basis could result in a reduction in state funding as outlined in the 2013 DICRA.

Collecting death records on a less than an annual basis would also compromise BJS’s capacity to report in a timely manner on trends in deaths in custody. It would also pose challenges for data collection because of the respondents store older data. It would impose additional costs associated with restarting the collection at various intervals.

Less than annual collection would delay publication of mortality data and collection on other than an annual cycle would make it difficult for BJS to maintain the high levels of participation. Respondents know that the collection is annual and over the years have developed internal procedures to facilitate responding to the DCRP. Every year since collection began, BJS has collected data from all 50 state Departments of Corrections. Due to the sensitivity of the information collected, it is likely some drop in participation would take place if collection ceased and did not resume for 2 or more years.

Turnover among respondents to the collection would also potentially negatively impact response rates and increase follow-up costs if the collection were fielded less frequently. With annual collection, BJS learns about pending turnover during routine data collection and verification calls (see Part B, section 2 for more information) and can plan for it. With less frequent collection, each effort to obtain data from the approximately 50 state departments of correction would require extra effort to implement the collection. Were the collection to be done every 2 years, BJS would incur additional costs associated with identifying new respondents and in all likelihood the data would not only be delayed but of poorer quality.

Finally, were the collection done on less than an annual basis there would be a loss of information. DCRP respondents have relayed that medical records and death certificates are often shipped off site within a comparatively short period of time, usually within a year of the death. If the data were collected on less than an annual basis, some respondents would no longer be able to access this critical piece of data. Other respondents would be required to go to off-site storage to obtain records, typically at an additional cost to the respondent. This would likely result in a negative effect on participation in the collection.

7. Special Circumstances Influencing Collection

Not applicable. These data will be collected in a manner consistent with the guidelines in 5 CFR 1320.6.

8. Federal Register Publication and Consultations Outside the Agency

The research under this clearance is consistent with the guidelines in 5 CFR 1320.6. The 60- and 30-day notices for public commentary have been published in the Federal Register. In renewing the data collection procedures, BJS has consulted with various experts to obtain their views on the instruments. BJS consulted the following correctional officials and public health experts:

• Dr. Roger Mitchell, Chief Medical Examiner, Washington, D.C.

• Dr. Margaret Warner, National Center for Health Statistics

• Dr. Ingrid Binswanger, MD, University of Colorado

BJS maintains frequent contact with DCRP data providers and data users in an effort to improve data collection, reporting procedures, data analysis and data presentation.

BJS received no comments relevant to the prison portion of the DCRP during the 60-day comment period following the publication of this proposed information collection in the Federal Register.

9. Payment or Gift to Respondents

Participation is without direct payment or compensation.

10. Assurance of Confidentiality

Respondents will be advised of Title 42 USC, Sections 3735 and 3789g, which establishes the allowable use of data collected by BJS. Under these sections (see Appendix C, attachments D and G), all information which can identify individuals is to be held confidential by BJS and shall be used only for statistical or research purposes, and shall be gathered in a manner that precludes their use for law enforcement or any purpose relating to a particular individual other than statistical or research purposes (Section 3735). As the collection agent for the DCRP, RTI will collect these data with assurances pursuant to these sections.

As required under these sections, any BJS report using these data will be only statistical in nature and will not identify individual subjects.

12. Estimate of Respondent Burden

The prison portion of the DCRP collects data from prison administrators in all 50 states. Estimates of annual burden on respondents are based on the number of hours required to review the instructions associated with the instruments, search existing data sources, obtain information necessary to complete data collection instruments, and provide data quality follow-up responses and verification.

Burden estimates are based on data from reported by prison respondents for years 2010 through 2013. A general summary of how burden estimates were calculated are provided in Table 1, with more detailed text below.

Table 1. Summary of Total Respondent Burden for DCRP Data Collection

Reporting Method	Type of Data Supplier	Number of Data Suppliers	Number of Responses	Average Reporting Time	Total Burden Hours†
Mail and Online Data Entry	State Prison – Death Records1	50	3,400	30 minutes per death	1,700 hours
Mail and Online Data Entry	State Prison – Annual Summary2	50	50	5 minutes	4 hours
Data quality follow-up	State Prison respondents	50	46	15 minutes	12 hours
Telephone	State Prisons – Verification Call	50	50	8 minutes	7 hours
TOTAL		50	3,646	73 minutes	1,723 hrs

† Hours rounded to the nearest whole number.

1. The form associated with the state prison death records is form NPS-4A.

2. The form associated with the state prison annual summary form is form NPS-4.

State Prison Respondent Burden. Reporting time estimates from state prison authorities are based on previous DCRP data collection cycles. Based on 2010 through 2013 data, 50 DOCs submitted data on 3,400 inmate deaths. The average response time for the death report form (i.e., NPS-4A) is 30 minutes per death, including follow-up time for data quality checks. Given these assumptions, we expect that in any data collection year:

• 50 state prison respondents will submit 3,400 death reports, with an average response time of 30 minutes per death report form, for a total burden of 1,700 hours.

Each state prison respondents will need to fill out a one-item annual summary form (i.e., NPS-4) with an estimated reporting time of 5 minutes. We expect the response time to remain constant:

• 50 state prison respondents will each submit an annual summary forms, with an average response time of 5 minutes per annual summary form, for a total burden of 4 hours.

The 50 state prison respondents will be called to participate in the verification call (with an estimated reporting time of 8 minutes). We expect that in any data collection year:

• 50 state prison respondents will have an average response time of 8 minutes per verification call, for a total burden of 7 hours.

Data quality follow-up occurs between May and July of each collection year and involves confirming, clarifying or correcting data reported on the NPS-4 and NPS-4A forms submitted by the respondent. Not all DOCs or records submitted require data quality follow-up. An average of 46 respondents require some level of follow-up annually, which averages about 15 minutes per respondent, for a total burden of 12 hours.

Thus, the total burden of hours associated with the prison data collection is 1,723 hours, with an average burden of 34 hours per year across the 50 state prison respondents.

13. Estimate of Cost Burden

We do not expect respondents to incur any costs other than that of their time to respond. The information requested is of the type and scope normally collected as part of their operations and no special hardware or accounting software or system is necessary to provide information for this data collection. Respondents are not expected to incur any capital, startup, or system maintenance costs in responding. Further, purchasing of outside accounting or information collection services, if performed by the respondent, is part of usual and customary business practices and not specifically required for this information.

14. Estimated Cost to Federal Government

Total cost to the federal government for all aspects of the program will be $105,823 annually.

BJS Cost Estimate Summary

These costs include $47,500 for overall program management, data analysis, publication review, and dissemination by BJS:

Staff costs:

All data collections, rounded to the nearest $100 (forms NPS-4, NPS-4A)

• 20% GS-13, Statistician ($30,000)

• 2% GS-14 ($2,800)

• 2% GS-13, Technical Editor ($2,000)

• 1% GS-12, Production Editor ($900)

• 1% GS-13, Digital Information Specialist ($1,000)

• Benefits (subtotal @ 20%, $4,300)

• Other administrative costs ($6,500 subtotal @ 30%)

Total, BJS costs: $47,500

Data Collection Agent Cost Summary

RTI International is BJS’ data collection agent for all three DCRP collections. The annual cost estimates are as follows (see Tables 2):

Table 2. DCRP Collection Annual Estimated Cost Estimate

BJS ASJ—DCRP collection cost estimate
20% GS-13, Statistician	$30,000
2% GS-14, Supervisory Statistician	$2,750
2% GS-13, Technical Editor	$2,000
1% GS-12, Production Editor	$850
1% GS-13, Digital Information Specialist	$1,000
Benefits ($21,600 subtotal @ 20%)	$4,320
Other administrative costs ($21,600 subtotal @ 30%)	$6,480
Data collection agent (RTI International) cost estimate
Personnel, benefits	$3,895
Indirect costs	$5,900
Computer expenses, materials, services, travel and other fringe costs	$263
Shipping, postage, telecommunications, reproduction, other	$100
Subcontractor expenses	$765
Total costs	$58,323

15. Reason for Change in Burden

The total respondent burden has decreased by 2,406 hours since the last OMB clearance was obtained for these activities. The reason for this decrease is that the arrest-related death collection and the jail collection will be seeking separate clearances.

16. Project Schedule and Publication Plan

BJS’s plans for products and publications from DCRP data over the next 3 years fall into three categories: Technical reports, BJS bulletins, and BJS special topic reports. The products include the following (a calendar of proposed publications is located at the conclusion of this section):

Technical reports

BJS plans to undertake methodological work to improve the cause of death variable and to expand BJS’s understanding of inmate mortality by linking the DCRP to other data sources. BJS plans to produce technical reports on these issues, each of which will describe the problem, the approaches taken to analyze the problem, the results, and recommendations for improving the utility of the DCRP that will address:

• Contingent on the success of assigning GID codes to each prison facility reporting a death and on the completion of the Census of State and Federal Adult Correctional Facilities (OMB clearance number 1121-0147), which is scheduled to be fielded in 2016, BJS proposes a report on prison facility characteristic and prisoner mortality. The report will examine issues related facility characteristics, such as staffing composition, programs offered to inmates, inmate characteristics (e.g. race/ethnicity, number of juveniles held and custody classification), age of facility and whether the facility has a geriatric wing in relation deaths occurring at the facility. The report will also examine what percentage of facilities did not report a death in recent years and how the characteristics of these facilities differ from facilities that consistently report deaths to the DCRP. (Expected release: 2018).

• BJS would like to explore options for improving death data for persons on probation and parole. In order to assess the extent of post-custodial deaths, BJS will release a report on community corrections deaths using data from the NCRP, the Annual Surveys on Probation and Parole (OMB clearance number 1121-0064), and the DCRP. The report will compare aggregate community correction death counts and crude mortality rates to the death counts and mortality rates. The aim of the report is to assess mortality differentials between custodial and post-custodial deaths. (Expected release June 2016).

BJS Bulletins

Annually, BJS releases statistical tables that report on trends in deaths in local jails and deaths. These data provide a “first cut” from an annual collection, and as described above, focus on the effects of changes in the composition of criminal justice populations on the overall change in mortality rates. More specifically, the bulletins/statistical tables will report, annually, on changes in the overall mortality rate and number of deaths, as well as mortality rates by characteristics such as age, sex, race, and jurisdiction in which deaths occurred. BJS will continue to use the DCRP data in the production of annual statistical tables:

• Mortality in state prisons and local jails, 2000-2014 (expected release August 2016)

• Mortality in state prisons and local jails, 2000-2015 (expected release August 2017)

• Mortality in state prisons and local jails, 2000-2016 (expected release August 2018)

The annual bulletins also serve as a vehicle for providing updates to previous years’ statistics by incorporating into the reports data from delayed data submissions. These statistics are consistent with the Committee on National Statistics (CNSTAT) recommendation to produce mortality data on a timelier schedule.¹⁵

BJS special topic reports

Periodically, BJS produces special topic reports from its data collections. These reports address a specific issue in more depth than can be addressed in the bulletins. BJS plans during the next three years for special topic reports from the DCRP data include:

• Drug and Alcohol Abuse and Mortality in State Prisons – This paper will examine the effects of drug and alcohol abuse and mortality in prisons and jails. Using data from the NDI-DCRP match, BJS will identify causes of death that are have a strong association with a history of drug or alcohol abuse. The analysis will use the NDI-DCRP linked files, because the linked death file has richer death data than the unlinked DCRP files. The analysis will select International Classification of Disease, 10th revision (ICD-10) codes that are indicate deaths associated with acute intoxication (ex. Cocaine toxicity) as well as deaths associated with chronic disease or conditions associated with a history of substance abuse (ex. Cirrhosis).¹⁶ (Expected release 2017).

• Mortality Among Incarcerated American Indians – This report will examine the mortality profile of American Indians incarcerated in both prisons and jails. Using data from 2000 through 2013, BJS will examine the demographics and cause of death of American Indians incarcerated in the U.S. correctional system. The report will compare American Indian inmate mortality outcomes to their counterparts in the resident population, as well as other non-American Indian inmates. (Expected release September 2016).

Online dissemination tools

Currently, DCRP prisoner death data are only available through the enclave at the National Archive of Criminal Justice Data (NACJD). Due to the individual-level records of DCRP and the sensitivity of the topic, NACJD has classified the archived DCRP data as a restricted dataset, available only through enclave. The categorization of the dataset as enclave means that any member of the public who wishes to get access to the data must obtain approval or a waiver from an institutional review board (IRB), demonstrate that they have adequate data security to view and use DCRP data, and write a description of their project to justify use of the data. Once a user license is granted, the research must travel to Ann Arbor, Michigan, and view the data onsite. Data sets cannot be removed and all output is reviewed by NACJD staff prior to removal from the enclave.

Due to the stringency of the enclave file access, potential data user’s access to the data is significantly limited. BJS is working with NACJD to create restricted and public-use versions of the data files that will allow users who are not able to access enclave-level data access to masked DCRP data files. Although work on this effort will not begin until early 2016, the restricted dataset will be stripped of personally-identifiable information (PII) (e.g., inmates’ names and correctional facility they were housed in prior to death, as well as PII that may be in the notes or specify fields). Level of aggregation is still be determined, but it is likely that the restricted and public use file will be aggregated at the state-level, meaning users will not to get facility-level data, and data suppression rules for small Ns will have to be finalized to prevent the inadvertent release of PII.

In addition to the data files through the NACJD, BJS will begin work in 2016 to create an online query tool for data users who are interested more basic queries, such as count of deaths by state and year, demographics of inmates who died in prions, and other information surrounding the death event such as inmates’ most serious offenses and whether medical care was provided prior to death. Due to the open accessibility of the tool, it will be designed to produce only national distributions for demographic and criminal justice characteristics. The tool will be designed to address more basic queries that can be answered without accessing the NACJD data files.

Table 3. BJS Calendar for DCRP Publications/Products

Type of BJS publication	Title/topic of publication/product	Estimated publication date
Annual bulletin	Mortality in state prisons and local jails, 2000-2014	September 2016
Annual bulletin	Mortality in state prisons and local jails, 2000-2015	September 2017
Annual bulletin	Mortality in state prisons and local, 2000-2016	September 2018
Special topic report	Mortality and Incarcerated American Indians, 2000-2013	April 2016
Special topic report	Substance Abuse and Intoxication-Related Inmate Deaths, 2007-2010	December 2016
Technical report	Assignment of GID and matching to prison census (dependent on fielding of Census of State and Federal Correctional Facilities in 2016 by BJS)	December 2018

BJS will continue to archive DCRP data at NACJD on an annual basis, submitting the data files for a given year concurrent with its publication of statistical tables update for that year. Statistical tables for a given calendar year are published in the fall of the following calendar year, given the roughly 18-month period to collect DCRP data. (See Part B, Item 2, Procedures for Information Collection).

17. Expiration Date Approval

The OMB Control Number and the expiration date will be published on all forms given to respondents and on the web portal used for electronic submissions of prison death data.

18. Exceptions to the Certification

There are no exceptions to the Certification Statement. The Collection is consistent with the guidelines in 5 CFR 1320.9.

1 Carson, E.A. (2015). Prisoners in 2014. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. Washington, DC. NCJ 248955

2 Glaze, L.E. and Kaeble, D. (2014). Correctional Populations in the United States. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. Washington, DC. NCJ 248479

3 Noonan, M., Rohloff, H. and Ginder, S. (2015). Mortality in Local Jails and State Prisons, 2000-2013. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. Washington, DC. NCJ 248756

4 Chiu, T. (2010). It’s About Time: Aging Prisoners, Increasing Costs and Geriatric Release. New York: Vera Institute of Justice.

5 Noonan, M., Rohloff, H. and Ginder, S. (2015). Mortality in Local Jails and State Prison. U.S. Department of Justice. Office of Justice Programs. Bureau of Justice Statistics. NCJ 248756.

6 Noonan, M.E. and Ginder, S. (2015). Mortality in Local Jails and State Prisons, 2000-2013. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. Washington, DC. NCJ 248756

7 Zeng, Z. et al (forthcoming). Assessing Inmate Cause of Death: Deaths in Custody Reporting Program and National Death Index. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics, Washington, D.C.

8 Medicolegal Death Investigation Data Quality Workshop on Oct 22.

9 National Research Council (2009). Ensuring the Quality, Credibility, and Relevance of U.S. Justice Statistics. A Panel to Review the Programs of the Bureau of Justice Statistics. Robert M. Groves and Daniel L. Cork, eds. Committee on National Statistics and Committee on Law and Justice, Division of Behavioral and Social Sciences and Education. Washington, DC: The National Academies Press.

10 Zeng, Z. et al. (forthcoming) Deaths in Custody Reporting Program Technical Report: DCRP-NDI Data Linkage. U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. Washington, DC.

11 Noonan, M.E. and Ginder, S. (2015). Understanding Mortality in State Prison: Do Male Prisoners Have an Elevated Risk of Death? Justice and Public Policy. Vol. 16. P.65-80.

12 Binswanger, I.A., Carson, E.A., Krueger, P.M., Mueller, S.R., Steiner, J.F., Sabol, W.J. (2014) Prison tobacco control policies and deaths from smoking in United States prisons: population based retrospective analysis. British Medical Journal. Aug 5;349:g4542.

13 Dr. Roger Mitchell, Chief Medical Examiner, Office of the Chief Medical Examiner, Washington, DC. Personal conversation, July 2015.

14 Dr. Margaret Warner, personal conversation, August 2015.

15 Panel to Review the Programs of the Bureau of Justice Statistics, National Research Council. "Abstract." Ensuring the Quality, Credibility, and Relevance of U.S. Justice Statistics. Washington, DC: The National Academies Press, 2009.

16 Shield, K.D., Parry, C. and Rehm, J. (2013). Focus On: Chronic Diseases and Conditions Related to Alcohol Use. Alcohol Research: Current Reviews. 35(2), pp. 154-173.