Data Sources and Methodology

5.0 Data Sources and Methodology

This help file section provides a detailed description of the purpose of and data sources for WISQARS NVDRS as well as the procedures involved in processing the data.

5.1 Purpose and Intended Audience of WISQARS NVDRS

WISQARS NVDRS provides data on violence-related incidents and injury deaths. These data are intended for a broad audience--the public, the media, public health practitioners and researchers, and public health officials--to increase their knowledge of violent injury. The data provide information on major causes of violent deaths, how common they are, and whom they affect. In addition, these death data help federal, state, and local public health officials to

• Characterize and monitor injury trends.
• Identify persons at risk.
• Provide reliable surveillance data for program and policy decisions.

5.2 Data Sources for Violent Deaths and Methodology

NVDRS is a population-based, active surveillance system that provides a census of violent deaths that occur among both residents and nonresidents of funded U.S. states. There is no sampling involved and thus all violent deaths are included. The system defines a death due to violence as "a death resulting from the intentional use of physical force or power against oneself, another person, or against a group or community," which is the World Health Organization (WHO) definition of violence. The case definition includes suicides, homicides, deaths from legal intervention (a subtype of homicide), deaths from undetermined intent, and unintentional firearm fatalities. Deaths of undetermined intent are included because this category includes deaths with some evidence of intent, but without enough to definitively classify the death as purposeful. In addition, information about unintentional firearm injury deaths (i.e., the individual did not intend to discharge the firearm) is collected, although these deaths are not considered violent deaths by the above definition.

Legal executions, which are considered part of deaths from legal intervention, are excluded from NVDRS as they are beyond the scope of public health. Deaths due to acts of war are also excluded. The system is coordinated and funded at the federal level and depends on separate data collection efforts in each funded state managed by the state health departments.

Unlike most public health surveillance systems that are victim-based, the NVDRS is incident-based and reports all victims and alleged perpetrators (suspects) associated with a given incident in one record. Decisions about whether two or more deaths belong to the same incident are determined by the timing of the injuries, rather than the timing of the deaths, and are based on a 24-hour rule (as stated below) and source documents indicating a clear link among the deaths.

Examples of a violent death incident include, but are not limited to:

• One isolated violent death.
• Two or more homicides, including legal interventions, when the deaths involve (1) at least one person who is a suspect or victim in the first death, (2) a suspect or victim in the second death, and (3) the fatal injuries are inflicted less than 24 hours apart.
• Two or more suicides or undetermined manner of deaths, when there is some evidence that (1) the second or subsequent death was planned to coincide with or follow the preceding death, and (2) the fatal injuries are inflicted less than 24 hours apart.

Each incident record includes information about victims, suspects, victim-suspect relationships, and any weapon(s) involved in the incident. To fully characterize the incidents, states collect information about each incident from numerous data sources.

The primary sources are:

• Death certificates (DC)
• Coroner/medical examiner (CME) records
• Police records (PR)
• Data abstractor input

Secondary or optional sources are:

• Child Fatality Review Team data (CFR)
• Crime lab data
• Supplementary homicide reports (SHR)
• Hospital data
• Alcohol Tobacco, Firearms and Explosives (ATF) trace information on firearms

Data collection is done by either abstraction from the records maintained by the primary sources at their offices or by transfer of data from the primary sources to the health department’s NVDRS office. Data may be manually entered into the software or electronically imported. Data collection is staged so that basic demographic information is available for early analyses and more detailed information about potential causal factors can be analyzed later. Death certificates often provide the earliest information in most states, but other states may identify incidents through the coroner/medical examiner or law enforcement offices. Regardless of the source of an identified incident, information is typically available to the health department and entered into the system within six months. Police and CME data are expected to be available within 18 months of the occurrence of the death.

The NVDRS database contains about 700 data elements. For deceased persons (i.e., deceased victims and suspects), NVDRS collects demographic data, as well as information on other personal characteristics such as marital and pregnancy status. It also collects data on the injury event (e.g., date, time, and place of injury), the occurrence of the death (e.g., time, place, and cause of death), as well as other related factors such as toxicology findings. NVDRS makes a unique contribution in that it captures information on circumstances for suicide and undetermined deaths, homicides, and unintentional firearm injury. The circumstances on suicide and undetermined death relate to mental health history and status, whether the person disclosed intent to die by suicide and precipitating factors. For homicide, circumstance information is captured on felony-related and nonfelony-related circumstances. Circumstance variables for unintentional firearm are related to the context and specific use of the firearms. Details are collected on relationships between victims and suspects, as well as whether there was a history of abuse or whether the suspect was a caregiver of the victim. The system also collects data on mechanisms leading to injury. The data are sorted by source document so the source of each entry can be determined.

WISQARS NVDRS provides access to a selected set of all data elements collected in states participating in the NVDRS, including characteristics (age, sex, race, ethnicity) of the victim or suspect, intent and cause of injury, and circumstances of the injury incident. Some of these data elements contain data from multiple data sources. For these data elements, CDC has set up computer algorithms to assign the most probable values for each victim or suspect. CDC also set up a primacy rule for which data source to use to assign a value for each victim or suspect to provide consistency across all participating states.

5.3 Population Estimates

Resident population data used to calculate crude death rates and age-adjusted rates are from the U.S. Census Bureau. Population data are from Bridged-Race Postcensal Population Estimates of the United States for July 1, 2003, and beyond. These estimates were produced by the U.S. Census Bureau under a collaborative arrangement with the National Center for Health Statistics (NCHS). These estimates are relatively consistent with population estimates used by NCHS for calculating mortality rates based on data from the National Vital Statistics System. For more information about these population data, visit the National Center for Health Statistics website at http://www.cdc.gov/nchs/about/major/dvs/popbridge/popbridge.htm . Please note that postcensal population estimates are updated annually, which means that death rates from WISQARS NVDRS prior to the update may be different.

5.4 Crude and Age-adjusted Violent Death Rates

WISQARS NVDRS Violent Death Rates: Death rates are calculated based on all resident deaths that occur within the participating states. Therefore, nonresident deaths are not included.

Crude Death Rate: The crude death rate identifies the number of violent deaths one would find among 100,000 people of that population. For instance, there were approximately 135,000,000 males in the U.S. in 2000. If there were 1,350 violent deaths in a particular category, the crude death rate would be (1,350 / 135,000,000) * 100,000 = 1.0 per 100,000 population.

Age-Adjusted Death Rates: Some injuries occur more often among certain age groups than others. For instance, suicide rates are generally higher among the elderly than among any other age group. Age adjustment enables one to compare injury rates without concern that differences are due to differences in the age distributions between populations or for the same population over time.

Because of the way age-adjusted rates are computed, WISQARS NVDRS may not be able to display age-adjusted rates for all categories on a report. For instance, a report by sex will have age-adjusted rates for males and for females, not both sexes combined. If you desire an age-adjusted rate for both sexes, a second request is required. Also, because of the way age-adjusted rates are computed, you cannot get age-specific rates and age-adjusted rates at the same time. If you select a report sorted by age group AND choose a standard year for age-adjusting, WISQARS will ignore the request for age-adjusted rates.