Creating a National Violent Death Reporting System

Abstract

Following the urgings of many advocates, an IOM report in 1999 called for a national data system on violent death, modeled after that existing for motor-vehicle crashes. The National Violent Death Reporting System now provides rich data on the circumstances of violent deaths in the U.S.; it is the first national system to collect in-depth information on suicides. In November 2008, the system went online (www.cdc.gov/ncipc/wisqars/), making data from 17 states available to researchers and the general public. The successful development of this system was the result of a 10-year effort begun with the support of a half-dozen foundations. Key components of the success included a high-quality pilot version of the system and the mobilization of a large, bipartisan coalition of individuals, institutions, and agencies that advocated for federal funding for the system.

Introduction

In November 2008, the website of the CDC's Web-based Injury Statistics Query and Reporting System (WISQARS)—the online injury-data querying system for the U.S.—began offering easy access to data from 16 of the 17 states that currently participate in the National Violent Death Reporting System (NVDRS). The NVDRS links detailed information on suicides and homicides from a variety of law-enforcement and public health sources to provide rich, incident-level data on violent deaths in the U.S.¹ Placing this complex data set at the fingertips of researchers, reporters, policymakers, and the general public alike is a notable milestone. It is thus an appropriate time to step back and describe some of the factors that led to the successful creation of this public health data system.

Motor-vehicle crashes are the leading mechanism of injury death in the U.S. Data systems created by the National Highway Traffic Safety Administration (most notably the Fatality Analysis Reporting System [FARS]) have allowed for the evaluation of many interventions, including vehicle-safety standards and motorcycle-helmet laws.²

Over the past 20 years, there have been many calls for the creation of a similar data system for injuries and deaths caused by firearms, which until 2004 were the second-leading mechanism of injury mortality. These calls included articles in the New England Journal of Medicine,³JAMA,⁴ and the American Journal of Public Health.⁵ A key document that advocated such a data system was the 1989 cost-of-injury report to Congress that recommended the development of a national fatal firearm-injury reporting system, comparable to FARS, with sufficient data for documenting the firearm problem and designing prevention strategies.⁶

In 1994, a good first step toward the development of a firearm-injury data system was taken when the CDC funded the development or enhancement of more than a half-dozen firearm-injury surveillance systems, primarily at state health departments.⁷ Unfortunately, this initiative ended abruptly 3 years later. When Congress—urged by gun-advocacy groups—reduced CDC's funding by precisely the amount ($2.6 million) it devoted annually to firearm research and data collection, the CDC halted virtually all activities related to firearm injuries.8, 9

Several states managed to continue their local firearm-injury surveillance, but the attempt to create a national firearm-injury data system appeared to have been arrested. In 1998, however, a group of private foundations—Atlantic Philanthropies, the Annie E. Casey Foundation, the John D. and Catherine T. MacArthur Foundation, the David and Lucile Packard Foundation, the Irene Diamond Fund, and the two leaders of this ad hoc consortium, the Joyce Foundation and the Open Society Institute—stepped up, providing temporary funding to re-invigorate the stalled initiative.

The foundations funded three groups to serve different roles in the development of a national firearm-injury surveillance system: (1) the Harvard Injury Control Research Center (HICRC) worked with a dozen collaborating institutions to create a pilot version and test its utility; (2) Fenton Communications produced compelling brochures and other publications for a “Just the Facts” campaign designed to communicate the need for such a system to the public and to policymakers; and (3) the Handgun Epidemic Lowering Plan (HELP) network (a firearm-injury–prevention coalition) mobilized its institutional members (e.g., medical and public health societies) to actively support the development of a national system funded by the federal government. The foundations were farsighted in understanding that it was not enough to use the pilot version to demonstrate that the system was possible; without an education-and-advocacy effort, the pilot would have remained only a pilot.

Many decisions about the development of the future surveillance system were made at HICRC-sponsored stakeholder conferences. For example, at a May 2000 meeting in Washington DC of researchers, surveillance experts, and representatives from federal agencies and local law enforcement as well as private institutions, conference participants agreed to a set of ten principles for establishing a federal surveillance system, including provisions that the system should be state-based and administered by the CDC. Most important, the participants concurred that the scope of the system should be (1) narrowed from all injury to fatal injury only, to reduce cost and logistic barriers, and (2) broadened to include all—not just firearm—deaths from suicide and homicide, as well as unintentional shooting deaths and deaths of unknown intent, to increase the system's scientific merit, public health impact, and political acceptability.¹⁰

This latter change was key. It increased the utility of the system and expanded its base of support to those more broadly interested in suicide and homicide prevention. It also created a stronger system with which to test whether firearm policies were associated with changes in overall—rather than just firearm-specific—rates of suicide and homicide. Government officials welcomed this broader system, as they saw a firearm-specific reporting system as not only less useful but also less politically tenable.

With funding from the foundations, HICRC provided funding to a dozen universities, health departments, and medical centers located in ten states, including sites that already had developed extensive local firearm-data systems, such as the Medical College of Wisconsin. The group of collaborators met weekly via conference calls to develop the system—including creating uniform data elements, reporting protocols, software, and training manuals—and piloted the system by collecting up to 3 years of data from each site.

Modeled on FARS, the system linked data from four data sources that already contained information about violent deaths: death certificates, law-enforcement data, medical examiner/coroner reports, and crime lab information. The first two sources of information already had their own national data systems—vital statistics collected by the National Center for Health Statistics and the supplementary homicide reports collected by the Federal Bureau of Investigation—but these had not been linked. Data from the other two sources had never been standardized or assembled together in a consistent way. The linked data sets allowed, for the first time, a more comprehensive picture of violent death, including data on victims' occupations, marital status, blood-alcohol content, military service status, mental health diagnosis, and abuse history, as well as the type of firearm used.

The HELP network aided in bringing together scores of organizations—medical associations, suicide-prevention groups, child-protection advocates, injury- and family violence–protection organizations, and others—to actively support the creation of the system. The organizations included the State and Territorial Injury Prevention Directors Association, the American Medical Association, the American Bar Association, the American Public Health Association, and the Suicide Prevention Action Network. The organizations wrote letters that were presented to key members of Congress, and a coalition met with Senate and House appropriators. HICRC personnel alone spoke before more than 80 national groups to raise awareness about the system, resulting in wide press coverage, including editorials in The New York Times, The Washington Post, The Miami Herald, U.S. News & World Report, and The Boston Globe.

All this was aided by a series of effective publications developed by Fenton Communications and HICRC. One was a “pop quiz” on important information about violent death (e.g., What percent of women killed in domestic violent attacks had a restraining order against the offender? What percent of youth suicide victims are intoxicated at the time they kill themselves?). The answer to all the questions was We don't know—because there is no national system in place that gathers and links the information into a useable database. Other highly readable and effective brochures were “A Public Health Approach to Saving Lives,” which emphasized the importance of a good data system, and “Linking Data to Save Lives,” which emphasized that what we don't know is killing us. (These publications are still available at the National Violent Injury Statistics System website [www.hsph.harvard.edu/hicrc/nviss/about_parent_nvdr.htm].) A website currently supported by the Joyce Foundation (www.preventviolence.net/index.html) continues to help promote the NVDRS message.

National reports emphasized the importance of creating an NVDRS. In 1999, for example, an IOM report called for a national data system for reporting suicides and homicides to provide objective data with which to monitor trends and evaluate the effectiveness of prevention programs and policies.¹¹ In 2001, the Surgeon General set as a national suicide-prevention objective the implementation, by 2005, of “a national violent death reporting system that includes suicides and collects information not currently available from death certificates.”¹² In 2002, an IOM report on suicide recommended that “federal funding should be provided to support a surveillance system such as the NVDRS that includes data on mortality from suicide.”¹³

The HELP network was crucial in working with legislative liaisons from the American Public Health Association, the American Academy of Pediatrics, the American College of Preventive Medicine, and others in educating Congressional staffers about the overwhelming support for and benefits of this data system. The skill of these lobbyists was critical in helping to shepherd the system through the appropriations committees. The need to persuade Congress about the importance of the system helped to ensure that the movement's educational message was not only scientifically sound but also short, clear, and consistent.

With growing support for the NVDRS, in 2002 Congress appropriated funds to the CDC to begin such a data system.¹⁴ The HICRC-led collaboration turned over the system to the CDC, assisting in the development of a CDC implementation plan, writing the extensive NVDRS coding manual, and training the sites on coding cases. Within 3 years, the CDC had provided funding for 16 states and part of California; these areas account for about one third of U.S. violent deaths. It is estimated that a complete system for all 50 states will cost the federal government some $20 million per year to operate.¹⁵

While a 2004 report from the National Research Council recommended support for the development and maintenance of the NVDRS,¹⁶ the number of states that participate, as of the beginning of 2009, remains at 17. Support for all states would make the NVDRS a more effective surveillance system and provide local data to help prevention efforts in every jurisdiction.¹⁵ A 2006–2007 HICRC survey of statewide suicide-prevention groups found that, among those in NVDRS states, most had used NVDRS data in planning their activities.

The surveillance system is providing useful data. For example, while it is relatively rare for a homicide to be followed by a suicide, data show that a man's murder of his female intimate partner is commonly followed by his suicide.17, 18 For such cases, making mental health services more readily available might not only reduce the suicide risk but also save the lives of many women. In addition, data from NVDRS have already been used for specific suicide-prevention activities. For example, after the Oregon NVDRS found that 37% of older adults who died by suicide had visited a physician in the preceding month, the state launched a plan to improve physicians' skills in identifying and treating potentially suicidal older adults.¹⁹ More recently, the U.S. Department of Veterans Affairs has been using NVDRS data to identify suicide risk among veterans, allowing the department to target more effectively its suicide-prevention activities.²⁰

Easy access to data from NVDRS will be a key to the system's utility. In 2004, a CDC description of the new system mentioned that “plans eventually call for a web-based, interactive system to create basic queries of aggregate data.”¹⁴ That such a web-based system is already up and running reflects well on the hard work and dedication of the CDC, its state partners, and those who supply the data: law-enforcement personnel, medical examiners, coroners, and vital statistics registries.