A.9
Prototype Indicator Set: Violence
BACKGROUND
Violence is one of America's most challenging and concerning public health problems. More than 2 million Americans are victims of violent injury each year (USDHHS, 1995), and violence is increasingly identified as one of the major concerns cited by citizens in urban and nonurban communities around the country (Rosenberg and Fenley, 1991). Although the public fears random violence, most violent acts are committed by individuals well known to the victim.
Violence is an issue that involves society at large; its causes and consequences permeate all sectors of a community. The committee acknowledges that violence is an issue for which measurement strategies are not standardized and health professionals have poorly defined roles. The prototype indicator set is intended to move the discussion of violence forward, not to be a final word on the issue. The committee chose to include this issue because the health effects of violence are clearly a major area of public concern.
The violent death rate (including both homicide and suicide) in the United States exceeds that of all other industrialized nations and constitutes the fourth leading cause of years of potential life lost prior to age 65 in the United States. Homicide is the leading cause of death in African American males aged 15–34 and the
second leading cause for young white men aged 15–24. Violence leads to many injuries as well as deaths, some of which require health care in hospital emergency rooms and other settings. In 1992, about 1 per 10,000 Americans suffered assault injuries (USDHHS, 1995).
The costs of violence are high—estimated at $54,000 per attempted or completed rape, $19,200 per robbery, and $16,000 per assault. A portion of these costs are financial, but the majority reflect pain, suffering, the risk of death, psychological damage, and reduced quality of life (NRC, 1993). Although violence affects all segments of American society, minorities are substantially more likely to be victims of violent crime. In 1990, for instance, African Americans were 41 percent and Hispanics 32 percent more likely than whites to be victims of violent crime (NRC, 1993).
Violent actions including rape, domestic violence, drive-by shootings, and terrorist attacks are highlighted daily in all forms of media. The issues of violence in our society are also highlighted in the content of television fiction and nonfiction programming. It is estimated that by age 18, children will have been exposed to 18,000 televised murders and 800 suicides; most will have seen 100,000 acts of violence on television by the sixth grade (Canterwall, 1992).
Substance use and psychosocial, family-mediated factors are major risk factors for violent behavior. Social factors associated with violence include concentration of poor families in geographic areas, income inequality, population turnover, community transition, family disruption, housing density and other aspects of social disruption, and opportunities for violence such as illegal markets in drugs and firearms (NRC, 1993). Dysfunctional family life (e.g., absent or divorced parents, turmoil and fighting between family members) is an early predictor of violence, especially for youth. Witnessing violence in the home and community is harmful to children and youth exposed to the violent event, even if they are not victims of the violence. Thus, as more violence occurs in homes, in neighborhoods, and in the media, greater percentages of the population are at risk to experience violence.
Criminal justice, health care, education, social services, and other community institutions can address the effects of violence and help prevent it. Strategies range from passing and enforcing strong gun control laws, to more punitive sentencing guidelines and building more jails, to community education and outreach programs aimed at violence prevention. Others focus on address-
ing the underlying causes of violence such as substance abuse and social disintegration.
Violence has been widely viewed as a health issue since the 1985 Surgeon General's Workshop on Violence and Health encouraged health professionals to address violence of all types. Violence is also one of the priority targets identified in Healthy People 2000: National Health Promotion and Disease Prevention Objectives (USDHHS, 1991). Specific objectives are included for homicide, suicide, weapons-related deaths, assault injuries, rape, adolescent suicide attempts, adolescent weapons carrying, and other issues. In 1995, ''weapons" was changed to "firearms" and an additional objective was added regarding laws requiring the proper storage of firearms to minimize access and the likelihood of discharge by minors (USDHHS, 1995).
"FIELD" SET OF PERFORMANCE INDICATORS
As indicated above, the causes and consequences of violence are broad and diverse, and cover many domains of the field model. Thus, it is possible to measure many dimensions of violence, its consequences, and its risk factors that shed light on the efforts of various community entities to address it.
Health and Function, Disease
The direct effects of violence include physical injury, death, and psychological problems. These can be measured through vital statistics (indicators 1 and 2, below), hospital discharge data (indicator 3), or survey data (indicator 4).
-
Homicide rate per 100,000 people.
-
Firearm-related deaths per 100,000 people.
Both of these indicators, which are included in Healthy People 2000 , can be calculated from vital statistics even at the local level. For small communities, however, the number of cases will be small and the number rather than the rate should be reported.
-
Number of emergency department visits related to violence, particularly gunshot wounds.
Data to measure this indicator may be available from hospital emergency department databases in states that E-code injury visits (i.e., Missouri and Nebraska). In areas that have not yet implemented E-coding of emergency department visits, information on
-
the number of gunshot wounds may be available from police departments, since the reporting of such incidents is generally required by law. Police data, however, may underestimate the number of visits. A review of emergency room logs and records would provide this information.
-
Number of assault injuries among people aged 12 and over.
This measure, which appears in Healthy People 2000, can be ascertained from police records, program data from rape crisis centers and shelters for victims of domestic violence, or population surveys.
Well-Being
Well-being can depend directly on the physical or psychological consequences of violence, but these dimensions are probably best measured directly. The one dimension that is not captured in health-related statistics is the effect on well-being of concerns about individual safety. A community might reduce fear of violence in three ways: (1) by reducing the actual risk of violence, (2) by helping people adapt to the possibility of violence by giving them information on how to alter their behavior to remain safe, and (3) by making information available regarding the true risks, which might be lower than people think. One way to measure this dimension is as follows:
-
Restriction of activities due to fear of violence.
Although not now generally available, population surveys could include questions to determine whether people fear violence enough to change their behavior or to measure specific activity restrictions (e.g., not going out at night or alone, not going to certain neighborhoods).
Individual Response
Violence is a complex problem of human behavior, so there are many ways to measure "individual response" to the social and physical environments associated with it. Some possible measures are as follows:
-
Suicide rate per 100,000.
-
Number of suicide attempts by adolescents aged 14 to 19.
Suicide is an important component of violence, which is re-
-
flected in the fact that both of these indicators are included in Health People 2000. The first indicator can be obtained from vital statistics even in small communities, but the number of cases may be small. If so, the number rather than the rate should be reported. Suicide attempts could be tracked through a hospital-based surveillance system for those cases serious enough to require inpatient or emergency care.
-
Number of rapes and attempted rapes.
Rape is an extreme form of sexual violence, and is addressed by Healthy People 2000. Data on reported rapes and rape attempts can be obtained through police reports and crime victim surveys, but these are thought to underestimate the true extent of the problem.
-
Prevalence of child maltreatment.
-
Prevalence of physical abuse of women by male partners.
-
Prevalence of elder abuse.
The three measures reflect the most common forms of domestic or family violence. Although there are mandatory reporting requirements for child abuse in most states, actual reports of domestic violence to police and child protective services probably reflect only a small portion of the events that take place. Furthermore, although there are multiple reports for some children, the available statistics have usually not been analyzed to indicate how many individual children have been involved. Unlike child abuse, health professionals and others are not required in most states to report cases of domestic or elder abuse, so data are likely to be even more incomplete. Domestic violence can be detected in population surveys, and indeed Healthy People 2000 relies on survey data for baseline estimates for the first two of these indicators.
-
Number of confirmed child abuse cases reported to authorities; percentage of confirmed cases receiving child protective services and appropriate medical care.
-
Number of restraining orders to protect women from domestic violence.
-
Number of batterers tried by the court system; percentage convicted or referred to treatment.
The courts and child protective systems can provide information on two dimensions of violence. First, numbers of violence-related cases in the courts and in child protective services indicate something about the prevalence of violent acts in the
-
community. Only a small proportion of violent events ever reach the courts and protective services, however, so if the percentage changes over time because of changed social norms about reporting, increased efficiency of the system, or other factors, changes in numbers of cases may not reflect changes in the prevalence of violence in the community. Data from the courts and protective services also can indicate something about the performance of those systems: Do investigators appropriately identify true cases of abuse, and does the system take the proper protective action? However, since reporting patterns may change over time (leading to higher or lower proportions of "true" cases among those reported) and courts and protective service determinations are not infallible, trends in the number or percentage of "positive outcomes" of these systems should be interpreted with caution. In addition, rates of substantiation or service referrals depend on the availability of resources for support services.
-
Incidence of physical fighting among adolescents.
-
Incidence of weapons carrying among adolescents.
-
Prevalence of substance abuse among youth and adults.
Violent behavior is more common among youth than at older or younger ages, so some of the measures of individual response ought to focus on youth violence. Substance abuse is a major contributor to violent behavior in a community, and can be measured through school-based as well as population surveys.
Health Care
Two types of measures can be used to address the role of health care in relation to violence: injuries resulting from violence that require health care and opportunities to prevent violence. Some possible measures are as follows:
-
Cost of care for intentional injuries (total and firearm injuries).
As indicated above, the costs of the physical and psychological consequences of violence are high. One measure of these costs might be obtained from emergency department and hospital discharge records, but this information is not available on a regular basis. Furthermore, such information must be interpreted carefully, since health care costs are only a small part of the costs of violence (NRC, 1993), and records for an initial hospital visit usually do not reflect follow-up treatment.
-
Number of people in substance abuse programs.
Because substance abuse is a major contributor to violence in a community, utilization of substance abuse programs is one way that a community can measure the contribution of the health sector to the prevention of violence. The number of people in such programs, however, may be due to both the availability of programs and the needs of the population. Ideally, it would be preferable to measure the proportion of those that need substance abuse programs that utilize them, but denominator data for this measure are generally not available. Thus, trends in utilization per se should be interpreted with caution.
-
Existence of protocols for health care professionals to identify, treat, and properly refer for further services individuals who attempt suicide; victims of sexual assault; and victims of spouse, elder, and child abuse.
-
Existence of child death review systems.
-
Proportion of children identified as neglected or physically or sexually abused who receive physical and mental evaluation with appropriate follow-up.
Injuries, including death, resulting from violence are not always easy to identify. Careful review of cases that present to health officials can help identify situations in which violence occurs, ensure that the victims and their caretakers receive the services they need, and provide an opportunity to prevent further violence in the future. These measures suggest actions that the health sector can take to respond to violence and contribute to its prevention.
Social Environment
As indicated above, the social factors associated with violence include concentration of poor families in geographic areas, income inequality, population turnover, community transition, family disruption, housing density and other measures of social disruption, and opportunities for violence such as illegal markets in drugs and firearms (NRC, 1993). Only some of these factors are amenable to public health or public policy interventions, at least in the short run, but communities might want to monitor all of them to help understand the causes of violence, and to target interventions to the areas in which they are most needed.
-
Concentration of poor families in geographic areas.
-
Of all of the underlying social factors associated with violence, this might be easiest to measure in communities because of income data available from the decennial census, even at the census block level. It would be important to measure not just level of income, but the concentration of poverty in small areas and disparities with other areas. It might also be possible to measure such items as vacant housing units and the proportion of families that own their own homes.
-
Opportunities for violence such as gangs, illegal markets, and firearms in the community.
Gangs, illegal markets, and firearms are typical of the situations in a community's social environment that provide opportunities for violence. Although the presence of such situations is difficult to represent in statistical terms, the use of focus groups and other observational techniques for documenting the existence of contributors to violence in a community would be an important measure of the performance of the police and criminal justice system and of local policymakers concerned with housing, economic development, and so on.
-
Use of community policing techniques.
"Community policing" is the policy by which police are assigned to specific neighborhoods and spend most of their time there solving problems before they lead to serious trouble. The premise of this policy is that if police are thought of as members of the community, they can contribute to an improved sense of security and intervene early to prevent violence and other crimes before they occur. Initial studies suggest that this approach can be effective (Fleissner and Heinzelmann, 1996).
-
Number of hours of violence-related programming on television most watched by children and youth.
Violence on television contributes to a social norm that condones or even glamorizes violence. As such, violence on television is a measure of the private sector's concern for the community, as well as the community's political will to control the forces that impact the health of its members.
-
Gun control laws.
The availability of guns in the community is an important proximal risk factor for violence. Gun control laws can address the sale, possession, storage, or use of guns and can address
-
various types of weapons from assault rifles to concealable handguns.
-
Availability of shelters for battered women and their children.
Shelters provide opportunities for women in abusive relationships to remove themselves and their children from a situation in which violence is likely to continue or to increase. The availability of shelters, which can be represented by the number of beds or perhaps the number of women turned away in times of need (USDHHS, 1991), is one measure of the contribution of community-based, often religious, organizations that typically provide these shelters.
Physical Environment
-
Percentage of youth reporting carrying weapons to school.
The availability of weapons is an important situational factor in the causal chain leading to violence (NRC, 1993). From a public health perspective, reducing the number of weapons in the school environment is analogous to removing the vector that transmits disease. Data on youth possession of weapons, including guns, at school could be obtained from school records of children found with weapons or from school-based surveys.
Genetic Endowment
Although there are some preliminary reports of genetic predisposition to violent behavior, not enough is known at this time to suggest performance measures in this area.
Prosperity
As indicated above, a community's prosperity both depends on and affects violence. Violence carries high economic costs, not only in treating its direct effects, but in terms of pain and suffering, psychological damage, and reduced quality of life. Similarly, economic factors, especially income inequality, are important risk factors for violence. These factors, however, are difficult to measure on a community basis and are reflected in other measures proposed above, so no specific measures are proposed in this area.
SAMPLE SET OF INDICATORS
From the preceding list of potential indicators, the following 12 are proposed as a primary set that might be used to monitor the performance of the community as a whole and of specific segments within it. Their selection reflects the committee's best judgment, but individual communities must consider available resources, including the availability of data. Comments on other uses of the measures (e.g., in Healthy People 2000), data and measurement issues, and suggestions of where accountability may lie are presented for each of the measures. The relatively large number of indicators in this suggested set reflects the varied forms of violence, causal factors for violence, and possible preventive and remedial interventions.
Of the five measures of the physical consequences of violence, one (firearm-related deaths) can be obtained from vital statistics in every community. The others (assault injuries, rapes, abuse of women by partners, and suicides attempts) might be obtained from a combination of police reports (e.g., through the Uniform Crime Reports System [FBI, 1994]) and state-based hospital discharge data systems. Data from these sources are also available at the local level, but both may underrepresent the actual number of cases for a variety of reasons. A population survey will be necessary to measure one of the proposed indicators (restriction in activities due to fear of violence) and could supplement official records of the nonfatal consequences of violence.
-
Firearm-related deaths per 100,000 people.
This indicator is the most reliable and valid of the proposed set. It is easily obtainable in a timely manner from public health departments, but it may be difficult to use as a performance measure in small communities because the number of firearm-related deaths will usually be small. With small numbers, changes are difficult to attribute to any of the accountable entities in the community. This indicator measures not only the outcome of violence, but also something about the risk factors for violence in the community and interventions such as gun control laws to control these risks. This indicator is included in Healthy People 2000.
-
Number of assault injuries among people aged 12 and over.
Data for this indicator should be available from hospital discharge data, police records, and population surveys. Each source is likely to be incomplete, but together they should yield an accu-
-
rate representation of the situation in the community. The E-coding of hospital discharge data (to indicate the cause of injuries) would help to identify assault cases, but it is not uniformly done throughout the country.
-
Number of rapes and attempted rapes.
-
Prevalence of physical abuse of women by male partners.
-
Number of suicide attempts by adolescents aged 14 to 19.
Data for these three indicators can be obtained from police and medical records, as well as from a population survey. Care must be taken to preserve the confidentiality of the individuals concerned.
-
Restriction of activities due to fear of violence.
This indicator, for which data must be obtained from a population survey, measures adverse effects on well-being that are not captured in health-related data. Although not now generally available, population surveys could include questions either to determine whether people fear violence enough to change their behavior or to measure specific activity restrictions (e.g., not going out at night or alone, not going to certain neighborhoods).
-
Number of hours of violence-related programming on television most watched by children and youth.
There are no standard measures of this indicator now, but the voluntary classification system agreed to by the television industry in 1996 will provide a starting point for definitions for measuring the amount of violence-related programming.
-
Percentage of youth reporting carrying weapons to school.
This measure reflects both individual response and the physical environment. Data can be obtained from school and police records and from student surveys. A version of this indicator appears in Healthy People 2000 .
-
Number of confirmed child abuse cases reported to authorities; percentage of confirmed cases receiving child protective services and appropriate medical care.
Although not reported in a consistently reliable and valid manner, the number of confirmed child abuse cases in a community is an indicator of the level of violence in families. Data on the number of cases are available from child protective service systems (NCCAN, 1996). The number of confirmed cases reflects the abil-
-
ity of the health care and education systems, as well as others in the community, to detect possible cases as well as the true incidence of child abuse. The second part of this indicator—the percentage of abused children receiving appropriate social and medical services—measures the ability of these systems to respond to identified needs. Determining whether the services provided are adequate and appropriate is a judgment that must be made in each community, informed by scientific evidence about the kinds of programs that are effective.
-
Existence of protocols for health care professionals to identify, treat, and properly refer for further services individuals who attempt suicide; victims of sexual assault; and victims of spouse, elder, and child abuse.
This indicator measures the capacity of the health care system to appropriately identify, treat, and refer victims of violence for further treatment and other services that may prevent future occurrences. There are no specific quantitative measures of the availability of these protocols, so implementation of this indicator will depend on judgment relevant to local circumstances.
-
Use of community policing techniques.
The implementation of community policing can be measured by the number or proportion of police on community policing beats or by the number of blocks of the community covered by a community policing approach.
-
Gun control laws.
The effect of gun control laws on the actual possession or use of guns will be difficult to measure, but their effect might be seen, for instance, in the proportion of homicides or robberies in which guns were involved. Rather than a statistical measure, an analysis of the strength of national, state, and local gun control laws and their implementation would be an appropriate indicator for the law enforcement sector.
The measures include the primary forms of violence that have been the focus of the public health community's concern and interventions: child abuse (indicators 9 and 10), violence against women (indicators 3, 4, and 10), youth violence (indicators 5, 7, and 8), and gun-related violence (indicators 1, 8, and 12). The proposed set combines measures of the direct effects of violence (indicators 1–5), one measure of the long-term social consequences
-
of violence (indicator 6), measures relating to predisposing factors for violence (indicators 7 and 8), measures of the quality of the treatment of its effects (indicators 9 and 10), and measures related to community efforts to prevent violence (indicators 11 and 12).
The direct measures of violence (indicators 1-6) reflect many factors in the community, and are indicative of the community's efforts as a whole to control violence. The indicators that measure the potential contributions of the schools (indicators 8 and 9), the media (indicator 7), the police and the criminal justice system (indicators 11 and 12), and the health care system (indicators 9 and 10), suggest specific actions for which these sectors in the community can be held accountable.
REFERENCES
Canterwall, B.S. 1992. Television and Violence: The Scale of the Problem and Where We Go from Here. Journal of the American Medical Association 267:3059–3063.
FBI (Federal Bureau of Investigation). 1994. Crime in the United States, 1994. Washington, D.C.: Government Printing Office.
Fleissner, D., and Heinzelmann, F. 1996. Crime Prevention Through Environmental Design and Community Policing. Research in Action Series. Washington, D.C.: U.S. Department of Justice, National Institute of Justice.
NCCAN (National Center on Child Abuse and Neglect). 1996. Child Maltreatment 1994: Reports from the States to the National Center on Child Abuse and Neglect. Washington, D.C.: U.S. Government Printing Office.
NRC (National Research Council). 1993. Understanding and Preventing Violence. A.J. Reiss, Jr., and J.A. Roth, eds. Washington, D.C.: National Academy Press.
Rosenberg, M.L., and Fenley, M.A., eds. 1991. Violence in America: A Public Health Approach. New York: Oxford University Press.
USDHHS (U.S. Department of Health and Human Services). 1991. Healthy People 2000: National Health Promotion and Disease Prevention Objectives . DHHS Pub. No. (PHS) 91-50212. Washington, D.C.: Office of the Assistant Secretary for Health.
USDHHS. 1995. Healthy People 2000: Midcourse Review and 1995 Revisions . Washington, D.C.: USDHHS, Public Health Service.
TABLE A.9-1 Field Model Mapping for Sample Indicator Set: Violence
Field Model Domain |
Construct |
Sample Indicators |
Data Sources |
Health and Function, Disease |
Violence-related mortality and morbidity |
Firearm-related deaths per 100,000 people |
Vital statistics |
|
|
Number of assault injuries among people aged 12 and over |
Police records, program data, population surveys |
Individual Response |
Frequency of sexual violence |
Number of rapes and attempted rapes |
Police records, surveys |
|
Frequency of domestic violence |
Prevalence of physical abuse of women by male partners |
State health department; police records |
|
|
Number of confirmed child abuse cases reported to authorities; percentage of confirmed cases receiving protective services and appropriate medical care |
Child protective services data |
|
Prevalence of self-directed violence |
Number of suicide attempts by adolescents aged 14–19 |
State health department, program data, hospital discharge databases |
Well-being |
General impact of violence in personal activities |
Restriction of activities due to fear of violence |
Survey required |
Social Environment |
Atmosphere that supports violence |
Number of hours of violence-related programming on television most watched by children and youth |
Television industry voluntary classification system |
|
Law enforcement prevention activities |
Use of community policing techniques |
Police department |
|
Legislative changes to decrease violence |
Gun control laws |
Police department, state and local statutes |
Health Care |
Professional awareness and support for victims of violence |
Existence of protocols for health care professionals to identify, treat, and properly refer individuals who attempt suicide; victims of sexual assault; victims of spouse, elder, and child abuse |
Surveys |
Physical Environment |
Environmental hazards |
Percentage of youth reporting carrying weapons to school |
State health department (Healthy People 2000 indicator data), surveys, education department, school districts |