By Debra Karch, Lead Behavioral Scientist, Etiology and Surveillance Branch, Division of Violence Prevention, Centers for Disease Control and Prevention, Atlanta, Georgia; and E. Lynn Jenkins, Branch Chief, Etiology and Surveillance Branch, Division of Violence Prevention, Centers for Disease Control and Prevention, Atlanta, Georgia
s law enforcement officers know all too well, no one is immune to violence. Violence takes its toll on individuals, families, and communities throughout the United States and affects people of all ages, from infants to the elderly, regardless of sex, race, ethnicity, or socioeconomic status. Each year, about 50,000 violent deaths occur in this country, the equivalent of the active rosters of 2,000 major league baseball teams.1 Suicides are nearly twice as common as homicides. Homicide is in the top four leading causes of death for persons aged 1 to 34 years; suicide is in the top three for persons aged 10 to 34 years.2 Historically, society has reacted to violence with law enforcement and criminal justice identification, apprehension, and punishment. But in recent decades, criminal justice and public health professionals have recognized the importance of preventing violence before it occurs.
A public health focus on preventing violent death was spurred by at least three trends: (1) homicide and suicide becoming more prominent causes of death as deaths by infectious and other diseases decreased; (2) the epidemic increase of homicides and suicides during the 1980s; and (3) an increased focus on preventing disease by changing behaviors (for example, a low-fat diet and exercise are prevention tools for heart disease), suggesting the same approach could be used to prevent violence.3 Understanding the background and circumstances associated with violence is key to being able to prevent deaths.
Achieving the goal of preventing violent deaths has been hampered by fragmented and incomplete information on the circumstances surrounding homicides and suicides. Currently available data do not always provide the information needed to accurately assess the factors associated with violent death and identify potential prevention strategies. For example, death certificates provide data on the victim’s demographics, along with the cause and the manner of death, but do not provide information on the perpetrator; the incident characteristics such as alcohol involvement; detailed weapon data; the type of location where injured, such as in a home or a park; and, of critical importance, the circumstances leading up to the death. This information is more commonly found in law enforcement reports and coroner or medical examiner data. Bringing these data sources together creates a single, comprehensive source of information.
The Evolution of Violent Death Data Collection
The Division of Violence Prevention (DVP) at the Centers for Disease Control and Prevention (CDC) is committed to preventing violent deaths in the United States and has been engaged in numerous efforts over the past two decades to improve the availability and utility of routinely collected information on violent deaths. For example, the CDC has been collecting data on school-associated violent deaths since 1992.4 This reporting system has provided valuable information, unavailable from either national death or law enforcement data, which has helped to monitor the problem and better understand its magnitude. CDC has also worked to improve data collection on fatalities associated with intimate partner violence, child abuse, and suicide. Each of these concerns could be more efficiently addressed in the context of a national violent death reporting system.
In 1999, the Harvard Injury Control Center launched the National Violent Injury Statistics System (NVISS) with funding from five major private foundations to support local violent injury data-gathering efforts and to build a model national reporting system.5 Members of NVISS worked with the Medical College of Wisconsin, state and local recipients of violence prevention grants, and others to design and pilot this reporting system modeled on the multiple source system for tracking motor vehicle fatalities, the Fatality Analysis Reporting System. The model system pulled data from four major reporting sources: death certificates, coroner and medical examiner reports, police incident and supplementary homicide reports (SHRs), and crime laboratory data. Administered through Harvard University, NVISS initially focused on collecting information on firearm-related injury, but expanded in its second year to collect information on homicides and suicides regardless of method in 13 sites, including several major metropolitan areas. The pilot system demonstrated that compiling data on violent injuries was both feasible and valuable for a better understanding of violence, and it provided the foundation to establish the National Violent Death Reporting System (NVDRS).
The National Violent Death Reporting System
In 2002, the CDC received federal funding to establish the NVDRS. The NVDRS collects data on suicides, homicides, and unintentional firearm deaths, as well as deaths of undetermined intent. Multiple deaths that occur during a single incident are linked to help identify risk factors for multiple homicides or homicides followed by suicides. Data are obtained from a variety of sources, including death certificates, law enforcement reports, medical examiner and coroner reports, crime laboratories, SHRs, and child fatality review teams. Individually, these sources provide fragmented data specific to their roles in violent death investigation and explain violence in only a narrow context. Together, these sources offer a more comprehensive picture of the circumstances surrounding a homicide or a suicide.
The ultimate goal of the NVDRS is to provide communities with a clearer understanding of violent deaths so that these deaths can be prevented. NVDRS works to achieve this goal by
- collecting and analyzing timely, high-quality data that monitor the magnitude, characteristics, and precipitating circumstances of violent death at the national, state, and local levels;
- ensuring that violent death data are disseminated routinely and expeditiously to public health officials, law enforcement officials, policy makers, and the public;
- ensuring that data are used to develop, implement, and evaluate programs and policies that are intended to reduce and prevent violent deaths and injuries at the national, state, and local levels; and
- building and strengthening partnerships among organizations and communities at the national, state, and local levels to ensure that data are collected and used to reduce and prevent violent deaths and injuries.
NVDRS began data collection in 2003 with seven states participating (Alaska, Maryland, Massachusetts, New Jersey, Oregon, South Carolina, and Virginia); six states joined in 2004 (Colorado, Georgia, North Carolina, Oklahoma, Rhode Island, and Wisconsin), and four more joined in 2005 (California, Kentucky, New Mexico, and Utah). In 2010, Ohio and Michigan successfully competed for funding and began collecting data. California collected data in only four counties from 2005 to 2009. The remaining states collect data statewide. Currently, the CDC funds 18 states to participate in the NVDRS (see figure 1). The CDC anticipates that ultimately the NVDRS will expand to include all 50 states, the District of Columbia, and the U.S. territories. The IACP Research Advisory Committee has developed a resolution to be presented during the annual IACP conference, October 23-27, in Orlando, Florida, that promotes increased law enforcement awareness of and agency participation in the NVDRS as well as expanded federal funding to enable every state, territory, and the District of Columbia to be included in the NVDRS.
The NVDRS collects approximately 250 unique variables for each death, including information on demographic characteristics such as age; sex; race; cause and manner of death; location, date, and time of injury and death; toxicology results; bodily injuries; decedent-suspect relationship; method of injury; precipitating circumstances; and a brief written narrative of the incident.6 The circumstances that precede a fatal injury are coded based on the content of law enforcement reports and the medical examiner or coroner record. Different sets of circumstances are coded for suicides, homicides, and unintentional firearm deaths. The following is a sampling of the circumstances captured in the NVDRS. If a specific circumstance is present but not a part of the current NVDRS variables, it can be noted in the narrative of the incident.
The system is coordinated and funded at the federal level, but depends on individual data collection efforts in each funded state. When a violent death occurs, the event is initiated in the NVDRS in one of two ways (see figure 2). A few states identify the death through direct communication with the medical examiner or coroner. Other states—the majority—identify deaths utilizing the death certificate issued by the medical examiner or coroner and submitted to the state vital statistics office. Once the incident is initiated in NVDRS, state health departments work with law enforcement, crime labs, toxicology labs, child fatality review teams, and other partners to collect the remaining data. Data collected each month by every state are forwarded to the CDC with all personally identifying information removed.
The Role of Law Enforcement
Law enforcement reports are critical to the success of the NVDRS in a number of ways.
First, the richness of law enforcement information, particularly as it pertains to circumstance information, is crucial to a complete understanding of these events. For a circumstance to be endorsed, that circumstance must be confirmed in either the law enforcement or the medical examiner or coroner report. For example, a medical examiner report on a homicide may not indicate that the death was gang related; however, the law enforcement report may clearly demonstrate that it is. In studies of data presented to lawmakers or communities, this death would not be reported as gang related without the contributions of the law enforcement report.
A study of the precipitating circumstances coded for each death in the NVDRS indicated that the cases identified as gang related increased by 209 percent when the law enforcement report was added to the medical examiner report.7 Similarly, the identification of deaths that were drug related increased by 95 percent when the law enforcement report was added. The opposite is true as well. Circumstances dealing with mental health, physical health, and substance abuse problems are more readily identified in medical examiner reports than in law enforcement reports, due to the former’s access to medical records. In fact, identification of suicide deaths involving premortem mental health problems increased by nearly 54 percent when the medical examiner information was added to law enforcement data; similarly, the identification of physical health problems as a contributing factor for suicides increased by nearly 35 percent, and substance abuse problems increased by nearly 60 percent.
Second, the law enforcement report may be the only source document that can identify unique incidents such as multiple death events (for example, multiple homicides or an incident of homicide followed by suicide), firearm trace information, or details not collected in SHRs. Law enforcement reports may also be the only document describing the event if a medical examiner or coroner report is not available. In addition, other source documents may have a substantial amount of missing data that can be completed by including a law enforcement report.
Third, information collected from multiple source documents may not always be consistent.8 For example, if inconsistencies exist between the death certificate and the medical examiner report, the law enforcement report may be the only other document that can be used to resolve those inconsistencies. The true richness of the NVDRS data is dependent upon the partnerships between state NVDRS programs, medical examiners and coroners, and law enforcement; the benefit is the availability of the comprehensive data set for all law enforcement agencies and violence prevention experts.
Benefits to Law Enforcement Participation in the NVDRS
|Suicide Circumstances Collected in the NVDRS|
- Recent mental health problem
- Recent treatment for mental health problem
- Alcohol problem
- Drug problem
- Physical health problem
- Job problem
- Financial problem
- Intimate partner problem
- Criminal legal problem
- Civil legal problem
- History of suicide attempts
- Disclosed intent to take own life
- Left a suicide note
Homicide Circumstances Collected in the NVDRS
- Argument over money or property
- Argument over topic other than money or property
- Drug involvement
- Gang involvement
- Intimate partner involvement
- If crime was precipitated by another crime
- Nature of the crime
- Crime was in progress
- If the victim was a bystander
- If the victim used a weapon
- If the victim was a police officer on duty
Law enforcement agencies benefit from participation in the NVDRS in at least two ways. First, participating agencies will be provided access to a unique combination of state- and community-level data through state-specific and national NVDRS reports and electronic data sets, potentially, if that is of interest to the law enforcement agency. The data and analyses not only summarize the law enforcement information provided by a jurisdiction, but also provide a comprehensive record of the combination of death certificate, medical examiner or coroner report, and the law enforcement data associated with violent deaths. Second, participation provides an opportunity for law enforcement agencies to be part of a system that allows better understanding of the circumstances preceding deaths in their communities, and thus brings valuable law enforcement expertise to the development of prevention programs. Many opportunities and success stories have emerged using analyses of NVDRS data.
The data collected in the NVDRS can also contribute to understanding specific violence problems such as violence directed toward police officers. NVDRS data from 2003 to 2009 indicate that in the 17 states where data were collected, there were 86 homicides of law enforcement officers. A review of the narratives for these cases indicate that at least 15 of these homicides occurred during a motor vehicle chase or traffic stop, 8 occurred during arrest or apprehension, and another 8 occurred while the suspect was in custody, corrections, or court. There were 5 incidents in which two law enforcement officers were killed, 4 incidents in which one law enforcement officer was killed and at least one other officer was injured, and 1 incident in which a law enforcement officer was killed along with a court reporter and a judge. This kind of detailed information identifies potential directions for prevention efforts.
A Case Study: Oklahoma NVDRS Data Support Police Intervention Study
From 2004 to 2007, 134 Oklahomans (120 intimate partner victims and 14 bystanders) were killed in intimate partner violence (IPV) incidents. Intimate partner homicides accounted for approximately 16 percent of all homicides in Oklahoma. More than one-fourth (28 percent) of the incidents were homicide-suicides; 5 percent were multiple homicides; and 67 percent were single-victim, single-offender incidents. Seventy-three percent of intimate partner homicide victims were females and 27 percent were males.
The impact of efforts to reduce IPV homicides cannot be evaluated without detailed homicide data that include the nature of the victim-suspect relationship and information on the circumstances surrounding the incidents. The Oklahoma NVDRS homicide data collected since 2004, along with IPV homicide data collected from 1999 to 2003, helped to secure grant funding for a National Institute of Justice research study on law enforcement interventions for domestic violence–related calls.
Investigators with Oklahoma NVDRS are working with researchers from the University of Oklahoma Health Sciences Center, Arizona State University, and Johns Hopkins University to evaluate the effectiveness of a new, innovative intervention for police officers responding to domestic violence scenes. The study began in January 2009 and will be conducted in two phases. Phase 1, the collection of data from a comparison group, is currently under way. During Phase 1, police officers have changed nothing about how they respond to domestic violence calls. During Phase 2, the intervention phase, police officers will conduct a brief, 11-item lethality assessment to determine if the victim is at high risk for homicide. If they determine the victim is at high risk, immediate coordination with the local domestic violence service provider will occur.
Data collected from victims in both phases of the study will be compared to determine if the intervention was effective for increasing victims’ safety. Seven police departments in Oklahoma are participating in the study, along with the domestic violence service providers in their areas. The study is expected to be completed by December 2011. Oklahoma NVDRS data will be used to help determine the impact of the intervention by monitoring homicide rates in the study areas over time. The 2007 data show that 40 percent of all IPV homicides in the state occurred in the areas served by the law enforcement agencies participating in the study.
The Role of Law Enforcement Leadership in the NVDRS
Developing a partnership with the NVDRS program in each funded state is the greatest contribution law enforcement can make to the success of the system and its ability to provide comprehensive violent-death data to states, communities, legislators, violence prevention–funding entities and programs, and the law enforcement community. Participation is easy. Each state NVDRS principal investigator will work with law enforcement agencies to determine the method of data sharing that works best for the agency. The CDC can provide contact information for each state, or state profiles can be viewed on the NVDRS website at www.cdc.gov/ViolencePrevention/NVDRS/stateprofiles.html. For states where there is not currently an NVDRS program, law enforcement leaders can raise awareness of the system and advocate that such an approach is needed in their states and nationwide. ■
1National Association for Public Health Statistics and Information Systems, “National Violent Death Reporting System (NVDRS)” (2007), http://www.naphsis.org/index.asp?bid=1014 (accessed September 13, 2010).
2Centers for Disease Control and Prevention, Injury Prevention & Control, “Data & Statistics (WISQARS),” http://www.cdc.gov/injury/wisqars/index.html (retrieved May 27, 2010).
3Linda L. Dahlberg, PhD, and James A. Mercy, PhD, “History of Violence as a Public Health Problem,” Virtual Mentor: American Medical Association Journal of Ethics 11, no. 2 (February 2009): 167–172, http://virtualmentor.ama-assn.org/2009/02/mhst1-0902.html (accessed September 13, 2010).
4Centers for Disease Control and Prevention, Injury Prevention & Control, “Youth Violence” (July 12, 2010), http://www.cdc.gov/ViolencePrevention/youthviolence/index.html(accessed September 13, 2010).
5NVISS Workgroup, Uniform Data Elements: National Violent Injury Statistics System (NVISS), Release 2.0 (Cambridge, Mass.: Harvard School of Public Health, 2002), http://www.hsph.harvard.edu/hicrc/nviss/documents/Version2UDE3.pdf (accessed September 13, 2010).
6Centers for Disease Control and Prevention, National Violent Death Reporting System (NVDRS) Coding Manual, version 3, produced by the CDC NVDRS Team; Jennifer C. Friday, PhD; and the Carter Consulting Company (2008), http://www.cdc.gov/ncipc/pub-res/nvdrs-coding/vs3/NVDRS_Coding_Manual_Version_3-a.pdf(accessed September 13, 2010).
7Joseph E. Logan, Debra L. Karch, and Alexander E. Crosby, “Reducing ‘Unknown’ Data in Violent Death Surveillance: A Study of Death Certificates, Coroner/Medical Examiner, and Police Reports from the National Violent Death Reporting System, 2003–2004,” Homicide Studies 13 (November 2009): 385–397.
8Debra L. Karch and Joseph E. Logan, “Data Consistency in Multiple Source Documents: Findings from Homicide Incidents in the National Violent Death Reporting System, 2003–2004,” Homicide Studies 12 (August 2008): 264–276.
Please cite as:
Debra Karch and E. Lynn Jenkins, "The National Violent Death Reporting System: The Critical Law Enforcement Component," The Police Chief 77 (October 2010): 130–140, http://www.nxtbook.com/nxtbooks/naylor/CPIM1010/#/130 (insert access date).