Violence in the United States Status, Challenges, and Opportunities

Interpersonal violence is a pervasive public health, social, and developmental threat. It is a leading cause of death in the United States, particularly among children, adolescents, and young adults. Exposure to violence can cause immediate physical wounds that clinicians recognize and treat but can also result in long-lasting mental and physical health conditions that are often less apparent to health care providers. Violence directly affects health care expenditures. Indirectly, it stunts economic development, increases inequality, and erodes human capital.

Interpersonal violence is defined by the World Health Organization as the intentional use of physical force or power, threatened or actual, against another person or against a group or community that results in or has a high likelihood of resulting in injury, death, psychological harm, maldevelopment, or deprivation.¹ Although violence has a long history of study by various fields, a focus on public health approaches to prevention has largely emerged over the past 3 decades. In 1992, the Centers for Disease Control and Prevention (CDC) established the National Center for Injury Prevention and Control as the focal point for advancing a public health approach to violence prevention in the United States.

Homicide rates have varied over the past 50 years. Beginning in the 1960s, the homicide rate in the United States increased steadily from about 4 to 5 deaths per 100 000 persons to a peak of 10.7 deaths per 100 000 in 1980.² Homicide rates remained markedly elevated through the mid-1990s and raised the profile of violent crime in political and social discourse. Rates of aggravated assault were 86 per 100 000 in 1960, peaked at 442 per 100 000 in 1992, and decreased to 242 per 100 000 in 2012 (Figure).

The homicide rate is from the National Center for Health Statistics3 and the assault rate is from the Federal Bureau of Investigation (FBI).4 The case-fatality rate (or lethality rate) was conceptualized by Harris et al5 and is calculated herein as the homicide rate divided by the homicide rate plus the assault rate (multiplied by 100 to display percentage). An aggravated assault is defined by the FBI as an unlawful attack by one person on another for the purpose of inflicting severe or aggravated bodily injury, usually accompanied by the use of a weapon or by means likely to produce death or great bodily harm.

The increase in homicides from the late 1960s to the 1990s (Figure and Table 1) has been attributed to factors such as the elevated proportion of youth among the population resulting from the post–World War II baby boom; the spread of multiple drugs of abuse; the proliferation of more powerful firearms; and rapid changes in family structures, cultural norms, and societal dynamics.⁶ Since the early 1990s, homicide rates have declined, but they still exceed rates in other high-income countries. The World Health Organization’s Global Status Report on Violence Prevention places the 2012 US homicide rate at 5.4 per 100 000, whereas the rate for Canada was 1.8; for the United Kingdom, 1.5; and for Australia, 1.1 per 100 000.⁷

For many forms of nonfatal violence—including child maltreatment, youth violence, intimate partner violence, sexual violence, and elder abuse—progress has also been made, but the burden remains high (Table 1). For example, from 1992 to 2012, official reports by child protective service agencies of substantiated sexual abuse declined by 62%, physical abuse by 54%, and neglect by 14%; however, an estimated 12.5% of US children still experience confirmed child maltreatment by age 18 years.^10,11Furthermore, a recent analysis of national crime survey data indicates that from 1995 to 2010, rates of rape or sexual assault among females decreased 58% from 5.0 episodes to 2.1 episodes per 1000 population; however, nearly 1 in 5 women (19.3%) have experienced rape (completed or attempted unwanted penetration) at some point in their life and experience the sequelae of such violence.^15,17

The tables and figure in this report synthesize information from multiple, national violence data systems maintained by the CDC, the US Department of Justice, the US Administration for Children and Families, and other partners, including the National Vital Statistics System, the Uniform Crime Reports, the National Crime Victimization Survey, the National Survey of Children’s Exposure to Violence, the National Child Abuse and Neglect Data System, the National Intimate Partner and Sexual Violence Survey, the Youth Risk Behavior Surveillance System, and the National Electronic Injury Surveillance System–All Injury Program. These data sources are diverse and range from household surveys to official administrative data (see eTable 1 in theSupplement for a description of the major data systems used).

Nearly all homicides are reported to health and safety officials and are counted in public data sources, but nonlethal violence is often unreported (eFigure 1 in theSupplement). For example, in 2011, approximately 1570 children younger than 18 years died from child maltreatment.¹⁸ In that year more than 3 million children received an investigation or response from state child protective services departments,¹⁸ but it is estimated that most episodes of child maltreatment were not reported. National survey data from 2011 estimate that 13.8% of children (calculated to be >10 million youth), had experienced some form of maltreatment by a caregiver.¹⁹ Child maltreatment can go undetected for a considerable amount of time; clinicians must be vigilant to avoid missed opportunities for detection and referral by recognizing symptoms and signs of maltreatment in clinical encounters.²⁰

Violence toward adults is also underreported. Survey data from the National Intimate Partner and Sexual Violence Survey estimate that more than 12 million women and men report experiencing some form of violence (rape, physical violence, or stalking) by an intimate partner each year, but only approximately 480 000 injuries are reported to police annually and 150 000 injuries receive medical treatment (eFigure 2 in the Supplement). The US Justice Department surveillance systems indicate that other forms of violent crime are underreported as well (eFigure 3 in the Supplement).

One reason that violence is underrecognized is that violence reporting systems are compartmentalized; the CDC’s National Violent Death Reporting System synthesizes information from several sources of data, but it only covers deaths and is currently operational in only 32 states. Medical, public health, police, judicial, child welfare, educational, correctional, and community agencies and organizations have also not yet built mechanisms for comprehensive, coordinated responses to violence.

Rates of violence vary by age group, geographic location, sex, race, and ethnicity. For example, homicide is the leading cause of death for non-Hispanic blacks from age 1 through 44 years, whereas it is the fifth most common cause of death among non-Hispanic whites in this age range. This high rate among African Americans is primarily driven by exceptionally high rates among males between the ages of 15 and 34 years.³ Child protective services reports from 2013 indicate that the rate of child maltreatment is 8.1 per 1000 children among white, 14.6 among black, and 8.5 among Hispanic children.¹⁸ Differences in child maltreatment rates as well as other forms of violence are attributable to underlying risk factors, such as poverty.³⁹Table 3 displays disparities in violence by race/ethnicity; disparities in rates of violence increase with the severity of violence (ie, homicide vs nonfatal violence).

Violence disproportionately affects younger individuals—rates of experiencing homicide among those between the ages of 15 and 39 years are more than 2-fold higher than those among individuals 40 years or older.³ The increase in homicide rates from 1985 to 1993 was largely accounted for by an increase in homicide among individuals aged 15 to 24 years.⁴⁰ Rates of violence also vary by location; for example, metropolitan areas have greater homicide rates than suburban or rural areas.⁸Variation in rates of violence by geographic area may be attributable to social constructs such as inequality and poor social cohesion, which are elevated in urban areas.⁴¹

The proportion of assaults resulting in the death of an individual (ie, the case-fatality rate) has declined markedly since the 1960s (Figure). This decline in the lethality of assaults is believed to be largely attributable to improvements in the quality and availability of trauma care. Over the past several decades, the number of hospitals, hospital capacity, physician availability and specialization, and technology used in caring for individuals who are critically ill has improved and such measures have been associated with lethality reductions on the county level.⁵ Injured patients receiving care at trauma centers have a lower risk of death than those treated at other facilities.⁴² However, since the mid-1990s, decreases in case-fatality rates have slowed (Figure). It appears that there is a limit to which skilled critical care can save patients’ lives; clinicians and health care systems should consider ways to become more involved in violence prevention.

Beyond physical injuries, which are the most apparent consequences of violence, the association between violence and infectious diseases, especially sexually transmitted infections (STIs) such as human immunodeficiency virus (HIV), is well established.^43,44 For example, although forced sexual intercourse can directly transmit infectious agents, sexual and nonsexual violence are also associated with subsequent early sexual debut, multiple partners, failure to use condoms or other forms of protection, and other behaviors that increase the risk of STIs. In a recent meta-analysis, relative to children experiencing no abuse, the odds of contracting STIs or engaging in risky sexual behaviors ranged from 57% higher among those children experiencing neglect to 78% higher among children experiencing physical abuse.⁴⁴Beyond the risk of STIs, exposure to abuse as a youth is also associated with adverse reproductive health outcomes, including fetal death and postpartum depression.⁴⁵

Experiencing violence (physical, sexual, psychological) is associated with increased risks of mental health and behavioral disorders such as depression, posttraumatic stress disorder, personality and conduct disorders, anxiety, sleep and eating disorders, substance abuse, and suicide and suicide attempts.⁴⁶ For example, meta-analyses indicate that children who were physically abused have a 54% increased odds of depressive disorders and a 92% increased odds of drug use.⁴⁴ Children experiencing emotional abuse or neglect can have even higher rates of psychological comorbidities (>300% increased odds of depressive disorders from emotional abuse and >200% increased odds from neglect, relative to children without abuse exposure).⁴⁴ Psychosocial outcomes in adulthood associated with past experiences of childhood violence may also include problems with finances, family, jobs, anger, and stress.⁴⁵

Lastly, violence is also associated with the development of major noncommunicable diseases, such as cardiovascular disease, cancer, chronic lung disease, and diabetes as well as key risk factors for several chronic diseases, including harmful alcohol use, tobacco use, physical inactivity, and obesity.^44,47 For example, data from meta-analyses, indicate that children who have experienced physical abuse have a 55% increased odds of tobacco smoking and a 32% increased odds of obesity, relative to nonabused children.⁴⁴ Although the relationship between childhood violence exposure and many noncommunicable diseases is still emerging, a growing number of prospective, longitudinal studies are establishing strong associations.

Violence prevention programs are often developed to address certain forms of violence (eg, child maltreatment, partner violence, youth violence) rather than multiple forms of violence. By focusing on those strategies that can reduce multiple forms of violence, practitioners have the potential to maximize gains in violence prevention. In Table 4 and Table 5, we highlight violence prevention strategies with some evidence for an effect on multiple forms of violence and that have been evaluated by major violence prevention evaluation bodies or possess other strong experimental evidence. The oldest and most tested violence prevention strategies are largely parent- and family-focused programs, early childhood education, therapeutic or counseling interventions, school-based programs, and public policy approaches.

Parent- and family-focused programs provide education and training to parents with the goal of improving emotional bonds between parents and children and teaching participants how to effectively discipline, monitor, and supervise children as well as strengthen access to social support and other resources. Programs for young children with components that teach parents communication skills and positive parent-child interaction skills and that include active role play and practice produce greater preventive effects than those without these components.⁶² Although parenting and family focused programs can vary substantially in content, method, and setting, a systematic review of early childhood home visitation programs detected a 38.9% reduction in episodes of child maltreatment in intervention participants compared with control participants.⁶³ The earlier parenting and family programs are delivered in a child’s life, the greater the benefits; however, significant benefits have also been demonstrated when delivered to adolescent populations.⁶⁴

The evidence also suggests that early childhood education can prevent future violence involvement. Early childhood education programs are associated with positive social and emotional development, lower rates of official reports of child abuse and neglect, less aggression and child behavior problems, and higher rates of secondary school completion and have also demonstrated long-term effects on violent and criminal behavior.^65,66 One example is the Chicago Child-Parent Center program, which provides a comprehensive educational program beginning in preschool, as well as other family support services to economically and educationally disadvantaged children.⁶⁷ A 15-year follow-up of preschool program participants across 25 sites in Chicago demonstrated statistically significant lower rates of juvenile arrests in the intervention vs control group (16.9% vs 25.1%, respectively).⁶⁷

There is also substantial evidence for therapeutic approaches and universal school-based violence prevention programs. Not only do therapeutic programs reduce the trauma-related harms of experiencing violence, but they also can prevent subsequent violence involvement. Cognitive behavioral therapy (CBT) and Multisystemic therapy are 2 examples of therapeutic approaches. A Campbell Collaboration meta-analysis of CBT programs for criminal offenders demonstrated a crime recidivism reduction of 25% (recidivism rate of 0.30 in the intervention group compared with 0.40 in the control group over approximately 12 months after the intervention period).⁵³ As for universal school-based programs, such approaches are associated with a 15% relative reduction in violent behavior in students across all school years participating in these programs and a 29% reduction in violence among students in high school.⁶⁸ School-based programs have demonstrated reductions in both peer violence and teen dating violence and are also cost-effective.^48,69

Although many policy-level interventions for violence remain to be tested, there is some evidence for policies that aim to reduce alcohol-related harms. Given the strong association between alcohol use and violence involvement, reductions in alcohol consumption are expected to be associated with reduced levels of multiple forms of violence. Based on systematic reviews, the Community Preventive Services Task Force recommends increasing alcohol prices, limiting days and hours of sale, regulation of alcohol outlet density, upholding liability of establishments for alcohol-related harm committed by customers (dram shop liability), enhanced enforcement of laws prohibiting the sale of alcohol to minors, and screening and brief interventions for problem drinkers.⁴⁹

There are also other policy- and community-level approaches that represent promising strategies to prevent multiple forms of violence. Violence is higher in communities where there are limited economic opportunities; where there are high concentrations of poor and unemployed people; where people move frequently; and where there are limited public, mental health, and social services available to residents and fewer civic and voluntary associations.³¹ Consequently, the evidence for other policy- and community-level approaches to address these characteristics points to income-strengthening approaches (eg, subsidies or cash transfers), urban upgrading (eg, improved transportation, lighting, buildings, green space), economic development strategies (eg, business improvement districts), and residential mobility programs that enable families living in disadvantaged environments to resettle in more advantaged neighborhoods.^70– 72 Increasing family income through subsidies or cash transfers, for example, has been shown to reduce child abuse and neglect, youth violence, and partner violence.^58,59,71

Lastly, bystander strategies are increasingly being used to prevent sexual violence, teen-dating violence, and bullying. With this approach, participants are trained to recognize potentially violent situations or behaviors that they see occurring and are taught skills to safely intervene. This approach also seeks to change underlying social norms that promote violence. A meta-analysis of bystander education to prevent sexual assault on college campuses found moderate effects on both bystander efficacy and intentions to help others at risk; smaller but significant effects were seen on increasing actual bystander helping behaviors, decreasing acceptance of rape myths, and decreasing rape proclivity.⁵⁶ Additionally, a recent meta-analysis of bullying programs detected an increase in bystander interventions among program recipients (an increase of 20% of 1 standard deviation more than control group participants).⁵⁷

Although less is known about the modifiable factors that serve as protective buffers in the face of disadvantage, previous research suggests that connectedness and community-level collective efficacy are protective factors that may offset many of the negative influences in disadvantaged environments. These factors also seem to be protective across multiple forms of violence, including child maltreatment, youth violence, intimate partner violence, and suicidal behavior.⁷³ Although, to date, few interventions in these areas have been tested and evaluated, it is a promising area for future research. It is also important to acknowledge the significant contributions to violence prevention and response elucidated by other domains, such as criminology. Strategies such as problem-oriented policing are also associated with important effects on violence reduction. The science of violence prevention has made significant progress over the last 2 decades. Widespread dissemination and adoption of available evidence-based strategies, however, is still needed.

New incentives and opportunities have been created to improve identification and assistance of patients at risk of violence and those presently experiencing violence. With the passage of the Affordable Care Act, screening adolescent and adult women for interpersonal violence and counseling must now be covered by many health plans with no cost sharing for patients.⁷⁵ Reimbursement rates and implementation details are likely to differ across plans, but this may facilitate identification of such patients and support on-going efforts. Large trials have shown little efficacy if the intervention for patients who screened positive for intimate partner violence involved providing them a resource list or informing their clinician with subsequent discussion and referral left to the discretion of the clinician.^83,84 Programs that have followed a positive screen result with more intensive support, such as a clinic-based CBT intervention, have noted improved outcomes.⁸⁵ One such program, which screened minority women at prenatal care clinics and then provided CBT, demonstrated a reduction in subsequent intimate partner violence (23.3% in intervention group vs 37.8% in control group over all follow-up interviews; intervention participants also gave birth to a lower proportion of infants with very preterm birth [1.5% vs 6.6%]).⁸⁵More research is needed to better identify elements of programs that most effectively assist individuals found to be experiencing intimate partner violence or to identify other novel strategies. For example, medical-legal partnerships, an approach that attempts to better integrate legal services for patients into clinical care, can potentially assist clinicians with interventions not typically considered by the medical community. For example, in 1 retrospective cohort study, at the 12-months follow-up, women who had obtained permanent protection orders had a rate of police-reported physical abuse that was 20% of those women without protection orders (rate of 2.9 incidents per 100 person-years in intervention vs 14.0 incidents per 100 person-years in control).⁸⁶

Some health systems have also attempted to address underlying risk factors by various forms of community engagement. For example, hospital systems have funded school-based educational programs that attempt to modify early risk factors for violence.⁷⁶ Other health care systems have begun to intentionally purchase produce, goods, or services from local businesses in high-risk communities in an attempt address underlying economic risk factors for violence involvement.⁷⁶ These strategies represent early engagement of health care systems with more population-level preventive approaches; evaluation of such strategies is needed to best guide efforts.