Suicide, Violence, and Other Forms of Injury
Joseph E. Logan And James A. Mercy
Field epidemiology often is not considered applicable to investigations of suicide, homicide, acts of nonfatal violence, or other fatal and nonfatal injuries. Investigations for such injuries and deaths are commonly believed to be the responsibility of law enforcement or the local or state coroner or medical examiners and not of public health officials. In fact, not until the 1970s and early 1980s were suicide and interpersonal violence recognized as public health problems rather than primarily matters for mental health care and law enforcement (1). The late ‘80s and early ‘90s continued to mark a wakeful period when local, state, and national leaders recognized increases in adolescent and young adult suicide, adolescent homicide, and homicide inequities disproportionately affecting African Americans over other racial/ethnic groups (1–3). For example, suicide rates for adolescents and young adults almost tripled during 1950–1990, and rates of homicide among youth in their late teenage years increased by 154% during 1985–1991 (1–3). In the realm of unintentional injury, although rates of motor vehicle–related deaths have declined substantially since the 1960s, during the past 15 years, death rates attributed to drug poisoning have tripled in the United States, driven primarily by the use of prescription and illicit opioids (4,5). These trends have generated demand for new thinking on ways to prevent self-directed and interpersonal violent and unintentional injuries. During 1979–1989, multiple reports from the Surgeon General, the Office of the Secretary of Health and Human Services, and the Centers for Disease Control and Prevention (CDC) pushed the national agenda of making suicide, violence, and other injuries public health priorities (2,3,6–8). Collectively, these reports called for the use of multidisciplinary public health strategies that proactively prevented injury. They also stated that effective approaches needed to involve multiple sectors, including healthcare, mental health, education, social services, and criminal justice. During this period, beliefs began to fade about whether mental health, education, and law enforcement initiatives alone were sufficient to handle the national burden of injury (6–8).
Concurrently, as injury prevention efforts incorporated the public health approach, the use of field epidemiology also gained momentum. Field epidemiologists helped discover new patterns or epidemics of injuries, why injuries sometimes clustered, what protected individuals or placed them at risk, and how public health officials can control local and sometimes national epidemics.
Injury-related epidemic investigations can involve intentional or unintentional injuries, and sometimes both. The intent of injury often directs which agencies are involved in subsequent investigations and therefore usually establishes the initial scope of an epidemic early in the investigative process. For example, for fatal injury– related cases, the local medical authorities will establish a “manner of death” of either suicide, homicide, accidental (hereafter referred to as “unintentional”), or undetermined (9). This initial classification sometimes directs which agencies should handle succeeding action if an epidemic of fatal injury cases is perceived. For instance, epidemiologic investigations into clusters of homicides might require stronger coordination with law enforcement efforts. Similarly, the intent behind nonfatal injury cases of interest (e.g., assault cases) can also determine whether law enforcement should be involved and therefore direct the initial scope of an epidemic investigation of nonfatal injuries. Sometimes, field investigators might pursue cases across multiple manners of death (e.g., suicide, unintentional, undetermined) or both intentional and unintentional nonfatal injuries (e.g., self-directed violence, unintentional) if they are most interested in a specific mechanism of injury, such as with drug poisonings. This is described later in the chapter.
Field epidemiologists investigate both nonfatal and fatal injuries, and sometimes both simultaneously. For example, field epidemiologists might investigate patterns of nonfatal and fatal suicidal behavior, nonfatal and fatal unintentional drug overdoses, or homicides and nonfatal violent assaults among specific populations of interest if there is reason to believe that both types of injuries are linked to the same risk factors.
Field investigations also can be short-term based on specific public health actions or services (10). For example, some investigations can be descriptive explorations that describe the magnitude, rate, and/or trends of a problem to determine whether an epidemic truly exists. Others might focus on other public health functions, such as determining the etiology of an outbreak, assessing service needs of a specific at-risk population affected by a traumatic event (e.g., a natural disaster) as part of a response effort or evaluating a public health program or policy, strategy, or surveillance system.
This chapter provides a brief overview on determining injury-related outbreaks, key questions to address, common exposures, and case examples of investigations by intent and by public health action. It also provides a brief overview of challenges often faced with injury-related epidemic investigations and examples of short-and long-term strategies used to help contain epidemics.
An early step in an epidemiologic field investigation is to assess whether an epidemic exists (i.e., whether the occurrence of incidents in a community or region of a group of injuries of similar nature is clearly in excess of normal expectancy and derived from a common propagated source [a]). “Outbreaks” are sometimes referred to as epidemics that have “sudden or violent increases” in injuries at specific locations. [b] Field epidemiologists also can view epidemics, or outbreaks, as “clustering” of cases, another term that implies cases are grouped together, beyond an expected number, in a specific space and time period (11). For injury-related field investigations, common data sources for estimating rates and determining epidemics include census data, death certificates, law enforcement reports, emergency medical service records, and coroner/medical examiner reports. To identify epidemics, field epidemiologists first determine a case definition that includes person-, place-, and time-related characteristics. More advanced analytic techniques, such as density-based modeling (12), also can be used to identify local epidemic clusters. However, the application of such techniques in suicide, interpersonal violence, and other injury-related field epidemiology is still in its infancy because such techniques require large databases and a large number of cases (12,13); the number of cases often identified in injury-related epidemic clusters is too small to detect with such methods.
Once field epidemiologists establish an increased trend in rates and/or a clustering pattern, they need to focus the investigation to help address key epidemiologic questions. The questions include the following:
- What large-scale exposure(s) was (were) involved in propagating the outbreak or cluster?
- What individual-level or personal factors increase one’s risk for the outcome of interest in the target population?
- Which risk factors in the target population might be more unique than those of other populations?
- Who is at increased risk?
- What factors protect against the outcome of interest, and are these factors unique in the target population?
- What are the possible points of intervention?
Epidemics of Suicidal Behavior
Two basic types of epidemics, or clusters, are known to be related to suicidal behavior: (1) point clusters and (2) mass clusters (14). “Point clusters” are a series of suicides or suicide attempts that occur closely in space and time. These clusters are most commonly investigated and often involve the interaction between at-risk persons and an immediate onset of a localized exposure (e.g., layoffs from a plant closing). “Mass clusters” are widespread geographically but grouped closely in time and are believed to be associated with widespread exposures (e.g., media stories, stock market crash).
During January 2004– December 2007, the suicide rate in a rural community increased from 12 to 141 deaths per 100,000 population, and suicide attempts increased from 273 to 967 per 100,000.
Public Health Response
CDC’s field investigation found that (1) risk factors for suicidal behavior included male sex (deaths only), adolescent/young adult age, gang activity, physical/sexual abuse, mental health problems, substance abuse, and domestic violence; (2) suicide by hanging was overrepresented, and first attempts by persons who recently had a friend die of suicide increased exponentially during this period (both findings suggested that “social contagion” partially precipitated the epidemic); and (3) residents believed the onset of behavior coincided with activities that “glamorized” the suicide deaths (e.g., spray painting verses in memory of victims on water towers).
Suicide can cluster and spread throughout a community through social contagion, and persons with multiple risk factors are particularly vulnerable to this social effect. Field investigators identified the mechanism/venues driving the social contagion and used that information to help stop the spread of suicide. CDC worked with the community to change the messaging about suicide toward how to seek help as a preventive intervention.
Source: Logan J, Halpin J, Diekman S, Vawter L, Crosby, unpublished data, 2009.
Clusters of suicidal behavior occur for a variety of reasons, but “social contagion” is often perceived to play a role. Social contagion refers to a social effect of suicide, whereby one person attempts or dies of suicide and then others who felt connected or drawn to the victim, or the act, follow suit (15,16). The essential element of social contagion is that exposure to the suicidal behavior enhances risk for subsequent suicide attempts. Understanding the mechanism of social contagion (or the path of suicide exposure) can help field epidemiologists and other investigators take action (Box 25.1). Persons who attempt suicide after exposure to suicidal behavior usually are those already at risk or “primed” (i.e., have many risk factors) for suicidal behavior (14). The social contagious path between victims might be strong if the two victims had a tight bond (e.g., suicide of one person resulted from him or her being depressed about another person’s suicide). Sometimes the path is less direct: at-risk persons might simply become attracted to a community’s reaction to a suicide, especially if local residents “glamorized” the victim or the act (Logan J, Halpin J, Diekman S, Vawter L, unpublished data; 15,16); this positive public attention can make suicide appear to be an appealing and honorable option to solving problems. In some clusters related to social contagion, victims even copy specific mechanisms of suicide (Logan et al., unpublished data). For example, cases may replicate a specific method of hanging.
Epidemics of Homicide and Nonfatal Interpersonal Violence
Similar to investigations into suicidal behavior, homicides and nonfatal forms of interpersonal violence can cluster and be linked to exposures that lend to intervention. When field epidemiologists work with law enforcement officials, the role of field epidemiology is to focus on preventing future injuries and deaths by identifying at-risk persons and relevant exposures in need of immediate public health action. For example, identifying and arresting a homicide perpetrator on a killing spree in a community can sometimes be a long process and not a feasible short-term strategy for intervening on the violent outbreak; however, field epidemiologists can use public health methods to identify ways to protect community members from being victimized by the perpetrator (Box 25.2).
During July 1, 1979– March 15, 1981, 22 unsolved child homicides and two child disappearances occurred in Atlanta.
Public Health Response
CDC field investigators conducted a case-comparison analysis of the victims and a case– control study comparing victims with other children in the community. Victims shared similar characteristics (e.g., young [children], black, male, overrepresentation of death by asphyxiation) that were distinct from other child homicide victims; these findings suggested the deaths were a discrete cluster. Case-children were more likely than control children to run errands for money and spend time alone on the streets or in shopping centers, which suggested a single perpetrator who was local and knew the children were approachable.
Field epidemiologists helped law enforcement by determining that case-children were part of a unique cluster, the deaths were most likely linked to the same exposure (i.e., same perpetrator), and specific factors placed children at risk for abduction (i.e., being alone, running errands for money). The public health approach also informed local residents about measures for protecting children (e.g., early curfews, parental supervision) until an arrest could be made.
Source: Reference 17.
Homicide and assault outbreaks also can result from gang activity. Gang-related homicides and assaults can spike when rival gangs engage in conflict and retaliation. Although field epidemiologists might not be able to immediately rid a community of gangs, they may be able to identify locales with escalated gang conflict and provide information to public health programs that specialize in stopping retaliatory gang activity in a community. Such programs include those that use the Cure Violence model (initially called CeaseFire) ). This model recruits former gang members and other community leaders who are respected among gang members to resolve gang conflicts peacefully, thereby interrupting the sequelae of retaliatory acts of violence. The Cure Violence program members who conduct these actions are called “the interrupters.” A future growth area for this program involves leveraging epidemiologic analyses to better inform and target the efforts of Cure Violence interrupters and other staff members who interact with the communities they serve.
Epidemics or clusters of nonfatal and fatal unintentional injuries also occur. Unintentional injuries are most often investigated in accordance with the mechanism of injury, such as drowning, motor vehicle crash, drug overdose, or natural disaster (e.g., hurricanes, tornados, earthquakes, tsunamis). The subsequent field investigations of such injuries usually focus on understanding the mechanism, intervening on the mechanism, or determining what places individuals at risk of injury by the mechanism. Investigations initially focused on unintentional injury and/or a specific mechanism may eventually broaden the scope of cases to investigate by including multiple manners of death if manner classification is difficult and unreliable. For example, medical authorities sometimes find it difficult to classify poisoning cases as “unintentional” or “suicide” and therefore sometimes misclassify such poisoning cases as one or the other or as having an “undetermined” manner of death (19).
Field epidemiologists frequently investigate ways to reduce injury and death related to drug overdose. During 2000– 2014, unintentional drug poisoning deaths and death rates per 100,000 persons tripled in the United States (4,5). This increase in drug-related poisoning mortality has been attributed largely to abuse of prescription and illicit opioids (20). During the 1990s, opioid prescribing became more common for managing chronic noncancer pain, consistent with clinical recommendations at the time based on the evidence available, and the widespread misperception about the low risk for addiction when opioids were prescribed long term (21,22). Over time, the numbers of prescriptions increased, at higher dosages, and for longer durations. In 2015, the amount of opioids prescribed was three times higher than in 1999 (23). Recently, the supply of heroin and illicitly manufactured fentanyl increased throughout parts of the United States, which has been associated with an increase of overdoses involving these drugs (24). In response to the opioid epidemic, field epidemiology has made important contributions to help understand the associated exposures (i.e., drugs involved in overdose deaths), the characteristics of overdose victims and risk factors for death, and the points of intervention (Box 25.3).
The aggregate findings of many epidemiologic field investigations and evaluation studies have also shown substantial decreases in deaths in areas with stricter laws governing opioid supply channels (e.g., pill mill laws) (26,27), which further highlights the importance of understanding the epidemiologic burden of overdoses in relation to key large-scale sources of exposure. Epidemiologic studies have informed other strategies to help reduce the opioid overdose epidemic. Such strategies include
- Promoting safe opioid-prescribing practices through guidelines to help reduce future risk for addiction.
- Enhancing clinician use of prescription drug monitoring programs to identify patients who might be seeking prescriptions from multiple providers.
- Implementing naloxone distribution programs (naloxone can inhibit the effects of opioids and immediately prevent death among persons experiencing an overdose).
- Improving access to substance abuse and medication-assisted treatment for persons addicted to opioids (23,28).
During 1999–2004, West Virginia had the greatest increase (550%) in unintentional poisoning–related deaths in the United States.
Public Health Response
Using data sources provided by the state’s Office of the Chief Medical Examiner (i.e., autopsy reports, toxicology reports, death-scene investigation reports, death certificates, and copies of the medical records) and records from the state’s controlled substance monitoring program, CDC field investigators characterized West Virginia residents who died of drug overdoses in 2006 with regard to potential risk factors and the types of drugs that were associated with their deaths. Most overdose deaths were associated with nonmedical use of pharmaceuticals, primarily opioid analgesics. Drug diversion (i.e., use without a prescription) was identified in 63% of the cases, and 21% of the decedents sought drugs from multiple providers (i.e., showed evidence of “doctor shopping”).
This seminal investigation of overdoses helped focus areas for prevention efforts (e.g., safer prescribing of opioids, clinical use of prescription drug monitoring programs that can identify persons seeking multiple prescriptions).
Source: Reference 25.
As mentioned, field investigations can sometimes focus on public health action or service. For example, investigations can be descriptive in nature, use etiologic inquiries, or be response oriented or evaluation oriented (10). Large epidemics sometime require different agencies taking on different actions. For example, in interagency responses, the local health departments might be tasked with describing the magnitude and timing of an epidemic in their respective geographic areas, whereas a federal agency might conduct the etiologic investigation to help understand why it occurred. This division of labor is sometimes done to help gather resources if the epidemic is believed to be too large for one agency to handle. Field investigators must also focus their investigation based on the needs of the authorities requesting assistance and the communities they represent.
Description-based epidemiologic field investigations are conducted to help understand the magnitude of a problem and to establish basic epidemiologic patterns (e.g., person, place, time characteristics and recent trends). This type of investigation particularly helps to determine whether there is an epidemic or outbreak and to generate hypotheses. These investigations often use existing data sources or survey data (Box 25.4) or link multiple data sources.
Etiology-based field investigations examine why an epidemic is occurring, which helps identify appropriate evidence-based practices as prevention strategies. These investigations often assess risk and protective factors in the target population, compare these factors with those of other populations to determine how the target population is unique, determine which factors are associated with the timing and onset of the epidemic, and identify which factors present the best points for interventions. Field investigators might use mixed methods of investigation, including quantitative methods (e.g., case–control, case–cohort or cohort, and cross-sectional study designs) and qualitative methods (e.g., focus group analyses). These investigations sometimes not only identify risk factors but also use advanced analytic models with the etiologic data to refine definition of groups at greatest risk. One example was a field investigation that used predictive analytics with etiologic data to identify youth at greatest risk for firearm violence (Box 25.5).
Concern has increased about sexual violence against children in sub-Saharan Africa.
Public Health Response
Using a two-stage cluster design, CDC field investigators and UNICEF surveyed female children to ascertain national estimates of sexual violence among this population, discern characteristics of the sexual violence, and identify who was at highest risk. In addition to determining that one in three women aged 13–24 years experienced sexual violence, investigators learned that the sexual violence largely involved perpetrators who were partners (e.g., husbands, boyfriends) of, or someone well known to, the victims.
Using this information, field investigators provided recommendations to in-country authorities for addressing the problem, including developing programs that improved intimate partner relationships and trained persons in communities to recognize signs of and respond to domestic-related sexual violence incidents.
Source: Reference 29.
Epidemiologic field investigations can be conducted after a communitywide traumatic event to identify critical needs and ways to prevent future injuries. Such rapid needs assessments can help local authorities allocate appropriate medical or mental health services. Examples of response-based investigations are those conducted after Hurricane Andrew in 1993; Hurricane Allison in 2001; the 2002 Washington, DC, sniper shootings (Box 25.6); Hurricanes Rita and Katrina in 2005; the Iowa floods in 2008; the Washington, DC, Metrorail crash in 2009; and tornados in Alabama in 2011. These investigations generally use cross-sectional study designs.
During 2011–2013, the number of victims of shootings in Wilmington increased from 95 to 154.
Public Health Response
Focusing on youth, CDC field investigators identified which groups were at greatest risk for firearm-related crime by analyzing administrative databases across multiple sectors (health, child welfare, juvenile services, labor, and education). They conducted a matched case–control study (case-youths were those who engaged in violent firearm crime and controls were other youth) and compared groups with respect to sentinel life events that were key documented events in these databases that could be risk factors for firearm violence (e.g., having been arrested, having been a victim of violence). Using predictive analytic methods, field investigators determined these linked administrative data sources can identify at-risk persons with great accuracy. Youth involved in firearm crime on average had 13 sentinel events, compared with an average of two for controls.
The key result of this investigation was that cities can merge existing data sources, examine risk factors, and prospectively identify who is at-risk for firearm-associated violence with great precision.
Source: Reference 30.
Evaluation-oriented field investigations can provide pilot evaluation findings to inform policy or program efforts, other prevention strategies, public health surveillance systems, or even public health survey instruments. They typically gather pilot data and evaluation results to inform decision-makers on whether they should embark on a larger evaluation effort. One example is a field investigation that evaluated whether children or parents can validly report on a child’s swim skill (Box 25.7). This report greatly enhanced surveillance efforts intended to monitor child swimming ability, a known protective factor against drowning, across different populations.
In October 2002, 10 people were shot and killed in public during a 3-week shooting spree in the Washington, DC, metropolitan area. Victims were shot while pumping gas, mowing lawns, walking outdoors, sitting on public benches, and conducting other activities of everyday life. The shootings evoked widespread fear among residents of the affected communities.
Public Health Response
CDC launched an epidemiologic field investigation to assess the psychological and behavioral responses to shootings among the local residents. Investigators examined measures of traumatic stress symptoms, perceptions of safety, behavioral responses, and exposures to the shootings. They discovered 45% of residents reported going to public places less frequently, and women who resided within 5 miles of any shooting were more likely than women living farther away to report symptoms consistent with posttraumatic stress disorder.
The findings showed severe violent incidents, such as the sniper shootings, can profoundly affect the psychological well-being of communities. The aftermath of such incidents requires the partnering of clinical and community leaders to address fear and to ensure that residents in close proximity to the violent exposure can access mental health services.
Source: Reference 31.
Epidemiologic field investigations of injuries have encountered challenges in common with and different from responses to investigations of infectious disease and other acute outbreaks. Here are three common challenges and strategies for addressing these challenges drawn from experiences in previous investigations:
- Local authorities and stakeholders can exert pressure to influence the investigation and its findings. Sometimes local authorities or stakeholders are motivated to be part of an investigation because of political or even personal reasons (e.g., a family member died by suicide). These stakeholders can set high expectations, demand exploration into personal theories, or simply attempt to derail an investigation to avoid unwelcome consequences. Lead field investigators who briefly explain to requesting parties what an investigation can realistically yield, including that not all questions of interest can be answered because of time and resource constraints and that the findings might not be popular with all stakeholders, will have better outcomes with the community. It would also be beneficial for field investigators to indicate clearly that the investigation will base recommendations on what the findings reveal, even if they are unpopular.
- Sample size constraints can limit the amount of information gleaned from an investigation. Clusters of suicide and homicide in particular can involve a small number of cases. Nevertheless, as the examples in this chapter show, investigators manage to conduct case–control, cross-sectional, and/or other descriptive studies. It is best to be aware of the sample constraints, attempt to focus on the most salient factors, and be cautious about how the findings are generalized to the underlying population when sharing these findings with local authorities. Investigators also should be straightforward with authorities and community stakeholders on the limitations inherent in examining small samples and the possibility that such samples might not provide the statistical power to identify anything beyond what is already known.
- Restrictions on primary data collection can limit access to information. For example, when CDC is sponsoring new data collection in response to a state request for urgent epidemiologic assistance, approval must be sought for the data collection from the US Office of Management and Budget (33). This approval process can last up to a year. Mechanisms exist for rapid approvals, but these mechanisms do not always apply to injury-and violence-related epidemic investigations. Therefore, CDC field investigators often must rely on secondary data, which can greatly restrict the kinds of questions that can be answered and the kinds of analyses that can be undertaken. One way to minimize this potential barrier is to link and analyze existing data sources that capture complementary details on the cases of interest. For example, with regard to homicides and suicides, investigators can link multiple data sources, such as law enforcement, coroner/medical examiner, and toxicology reports, by case incident to review details on each case gathered from multiple perspectives (34).
Drowning is the second leading cause of unintentional injury death among children. Although drowning risk is decreased for those who have swimming skills, surveillance data on child swimming ability is normally collected through self-or proxy reports rather than observed in-water performance. The use of validated survey measures of swim skills is essential for public health officials to know the true prevalence and accuracy of this drowning protective factor among youth.
Public Health Response
CDC and Seattle Children’s Hospital responded to a field epidemiologic aid (“Epi-Aid”) request to evaluate the validity of self-reports and parental reports of a child’s swim skill. This pilot evaluation also explored which swim skill survey measure(s) best correlated with children’s in-water swim performance. A total of 482 parent–child dyads were recruited at three outdoor public pools in Washington state. Parent and child reports of three swim survey measures (i.e., “ever taken swim lessons,” “perceived good swim skills,” and “comfort in water over head”) were compared with the child’s in-water swim performance. Only parental reports of “perceived good swim skills” were validated with actual child swim ability.
This field investigation was evaluation-oriented and helped determine how well a public health surveillance system can accurately assess and report on an important protective factor against children’s drowning. Parent report of perceived “good swim skills” was a strong survey measure to assess child’s swim skill; history of swim lessons was not a useful measure. The findings from this investigation informed a statewide youth health survey on how to accurately measure child swimming skills in a community.
Source: Reference 32.
The essence of epidemiologic field investigations ultimately is to help identify short-and long-term prevention strategies. Short-term interventions are intended to stop the immediate spread of the problem; long-term recommendations are directed at making sustainable communitywide changes. A number of resources on injury-related prevention strategies are available to help take action and save lives. However, a good field investigation is always useful to direct public health officials to the most appropriate ones. Furthermore, field epidemiology will remain a cornerstone for innovation and development of new prevention strategies. Some of the most impactful ideas on preventing injury today were once small ideas generated from field investigations. The frontline work of “shoe-leather epidemiology,” a common term for field epidemiology, can change the thinking on injury prevention especially as new problems emerge, technology advances, and the mechanisms for how people interact and communicate evolve. For more details on existing resources useful to field epidemiologists with regard to preventing injury-and injury-related outbreaks, review the following:
- CDC’s Striving to Reduce Youth Violence Everywhere (35).
- University of Colorado, Boulder’s Blueprints for Healthy Youth Development (36).
- Cure Violence (18).
- Technical packages on the best available evidence for preventing child abuse and neglect (37), sexual violence (38), youth violence (39), suicide (40), and intimate partner violence (41).
- CDC Guideline for Prescribing Opioids for Chronic Pain (42).
- World Health Organization. Global Report on Drowning: Preventing a Leading Killer (43).
- World Health Organization. Save LIVES—A Road Safety Technical Package (44).
- “Epidemic,” in Gordis L., ed. Epidemiology. 2nd ed. Philadelphia: W. B. Saunders Company; 2000. CDC, the World Health Organizations, and Merriam-Webster all.
- ”Outbreak,” Merriam-Webster.com. 2017. https://www.merriam-webster.com/dictionary/outbreak
- Dahlberg LL, Mercy JA. The history of violence as a public health issue. AMA Virtual Mentor. 2009;11:167–72.
- Alcohol, Drug Abuse, and Mental Health Administration. Report on the Secretary’s Task Force on Youth Suicide. Volume 1: overview and recommendations. https://www.hsdl.org/?view&did=743317
- CDC. Homicides among 15–19-year-old males—United States, 1963–1991. MMWR. 1994;43:725–7.
- CDC. Web-based Injury Statistics Query and Reporting System (WISQARS). https://www.cdc.gov/injury/wisqars/
- Rudd RA, Seth P, David F, Scholl L. Increases in drug and opioid-involved overdose deaths—United States, 2010–2015. MMWR. 2016;65:1445–52. Erratum in: MMWR, 2017;66:35.
- US Department of Health, Education, and Welfare. Healthy people: the Surgeon General’s report on health promotion and disease prevention. Washington, DC: US Department of Health, Education, and Welfare, Public Health Service; 1979. https://profiles.nlm.nih.gov/ps/access/NNBBGK.pdf
- Institute of Medicine and National Research Council. Injury in America: A Continuing Public Health Problem. Washington, DC: National Academies Press; 1985.
- US Department of Health and Human Services. US Department of Justice. Surgeon General’s Workshop on Violence and Public Health Report. Washington, DC: Health Resources and Services Administration; 1986. https://www.nlm.nih.gov/exhibition/confrontingviolence/materials/OB10998.pdf
- CDC. National Center for Health Statistics. Medical examiners’ and coroners’ handbook on death registration and fetal death reporting. 2003 revision. Hyattsville, MD: US Department of Health and Human Services; 2003. https://www.cdc.gov/nchs/data/misc/hb_me.pdf
- Halperin WE. Field investigations of occupational disease and injury. In: Gregg MB, ed. Field Epidemiology. 2nd ed. New York: Oxford University Press; 2002:306–23.
- CDC. Lesson 1: Introduction to epidemiology. Section II: Epidemic disease occurrence. Principles of epidemiology in public health practice, third edition: an introduction to applied epidemiology and biostatistics. https://www.cdc.gov/ophss/csels/dsepd/ss1978/lesson1/section11.html
- Ester M, Kriegel H, Sander J, Xu X. A density-based algorithm for discovering clusters in large spatial databases with noise. KDD’96 Proceedings of the Second International Conference on Knowledge Discovery and Data Mining. 1996 Aug 2–4; Portland, OR. Palo Alto, CA: AAAI Press; 1996:226–31. http://www.aaai.org/Papers/KDD/1996/KDD96-037.pdf
- Corcoran JJ, Wilson ID, Ware A. Predicting the geo-temporal variations of crime and disorder. Int J Forecasting. 2003;19:623–34.
- Joiner TE Jr. The clustering and contagion of suicide. Curr Dir Psychol Sci. 1999;8:89– 92.
- Ali MM, Dwyer DS, Rizzo JA. The social contagion effect of suicidal behavior in adolescents: does it really exist? J Ment Health Policy Econ. 2011;14:3–12.
- de Leo D, Heller T. Social modeling in the transmission of suicidality. Crisis. 2008;29:11–9.
- Blaser MJ, Jason JM, Weniger BG, et al. Epidemiologic analysis of a cluster of homicides of children in Atlanta. JAMA. 1984;251:3255–8.
- Cure Violence. http://cureviolence.org/
- Rockett IR, Hobbs G, De Leo D, et al. Suicide and unintentional poisoning mortality trends in the United States, 1987–2006: two unrelated phenomena? BMC Public Health. 2010;10:705.
- Paulozzi LJ, Budnitz DS, Xi Y. Increasing deaths from opioid analgesics in the United States. Pharmacoepidemiol Drug Saf. 2006;15:618–27.
- Leung PTM, Macdonald EM, Stanbrook MB, Dhalia IA, Juurlink DN. A 1980 letter on the risk of opioid addiction. N Engl J Med. 2017;376:2194–5.
- American Academy of Pain Medicine and American Pain Society. The use of opioids for the treatment of chronic pain: a consensus statement from the American Academy of Pain Medicine and the American Pain Society. Clin J Pain. 1997;13:6–8.
- CDC. Vital Signs: Changes in opioid prescribing in the United States, 2006–2015.
- Gladden RM, Martinez P, Seth P. Fentanyl law enforcement submissions and increases in synthetic opioid– involved overdose deaths—27 states, 2013–2014. MMWR. 2016;65:837–43.
- Hall AJ, Logan JE, Toblin RL, et al. Patterns of abuse among unintentional pharmaceutical overdose fatalities. JAMA. 2008;300:2613–20.
- Rutkow L, Chang HY, Daubresse M, Webster DW, Stuart EA, Alexander GC. Effect of Florida’s prescription drug monitoring program and pill mill laws on opioid prescribing and use. JAMA Intern Med. 2015;175:1642–9.
- Kennedy-Hendricks A, Richey M, McGinty EE, Stuart EA, Barry CL, Webster DW. Opioid overdose deaths and Florida’s crackdown on pill mills. Am J Public Health. 2016;106:291–7.
- CDC. Injury prevention & control: opioid overdose. https://www.cdc.gov/drugoverdose/index.html
- Reza A, Breiding MJ, Gulaid J, et al. Sexual violence and its health consequences for female children in Swaziland: a cluster survey study. Lancet. 2009;373:1966–72.
- Sumner S, Maenner M, Socias C, et al. Sentinel events preceding youth firearm violence: an investigation of administrative data in Delaware. Am J Prev Med. 2016;51:647–55.
- Schulden J, Chen J, Kresnow MJ, et al. Psychological responses to the sniper attacks: Washington DC area, October 2002. Am J Prev Med. 2006;31:324–7.
- Mercado MC, Quan L, Bennett E, et al. Can you really swim? Validation of self and parental reports of swim skill with an inwater swim test among children attending community pools in Washington State. Inj Prev. 2016;22:253–60.
- US Environmental Protection Agency. Laws and Regulations. Summary of the Paper Reduction Act. 44 USC §3501 et seq. (1980). https://www.epa.gov/laws-regulations/
- Paulozzi LJ, Mercy J, Frazier L, Jr., Annest JL. CDC’s National Violent Death Reporting System: background and methodology. Inj Prev. 2004;10:47– 52.
- CDC. Striving to Reduce Youth Violence Everywhere. http://vetoviolence.cdc.gov/stryve/
- Center for the Study and Prevention of Violence, University of Colorado Boulder. Blueprints for healthy youth development. http://www.colorado.edu/cspv/blueprints
- Fortson BL, Klevens J, Merrick MT, Gilbert LK, Alexander SP. Preventing Child Abuse and Neglect: A Technical Package for Policy, Norm, and Programmatic Activities. Atlanta: CDC, National Center for Injury Prevention and Control; 2016.
- Basile KC, DeGue S, Jones K, et al. STOP SV: A Technical Package to Prevent Sexual Violence. Atlanta: CDC, National Center for Injury Prevention and Control; 2016.
- David- Ferdon C, Vivolo- Kantor AM, Dahlberg LL, Marshall KJ., Rainford N, Hall JE. A Comprehensive Technical Package for the Prevention of Youth Violence and Associated Risk Behaviors. Atlanta: CDC, National Center for Injury Prevention and Control; 2016.
- Stone DM, Holland KM, Bartholow B, Crosby AE, Davis S, Wilkins N. Preventing Suicide: A Technical Package of Policies, Programs, and Practices. Atlanta: CDC, National Center for Injury Prevention and Control; 2017.
- Niolon PH, Kearns M, Dills J, et al. A Technical Package to Prevent Teen Dating and Intimate Partner Violence. Atlanta: CDC, National Center for Injury Prevention and Control; 2017.
- Dowell D, Haegerich TM, Chou R. CDC Guideline for prescribing opioids for chronic pain— United States, 2016. MMWR. 2016;65(no. RR- 1):1– 49.
- World Health Organization. Global report on drowning: preventing a leading killer.http://apps.who.int/iris/bitstream/10665/143893/1/9789241564786_ eng.pdf?ua=1&ua=1
- World Health Organization. Save LIVES— a road safety technical package. Geneva: World Health Organization; 2017. http://apps.who.int/iris/bitstream/10665/255199/1/9789241511704-eng.pdf?ua=1.