Public health framing of relationship violence

C. Everett Koop convened a workshop on violence as a public health problem in October 1985. The proceedings were unprecedented: physicians, epidemiologists, criminologists, and victim advocates in the same room, treating violence as a phenomenon with rates, risk factors, and potentially modifiable causes rather than a fixed feature of human nature. The workshop did not invent the public health framing—Koop's predecessors had hinted at it—but it gave the framing federal legitimacy. Within five years, the CDC had created a Division of Injury Control with a violence portfolio. The institutional consequence of the workshop was larger than its intellectual content. It made a budget line possible.

Linda Saltzman led the CDC team that published Intimate Partner Violence Surveillance: Uniform Definitions and Recommended Data Elements in 1999. Before this document, every researcher used different definitions, every state's police data was incommensurable, and the national prevalence number depended on which study you cited. The document was technical and dry. It also did more for the field than any number of campaigns. Once jurisdictions adopted the definitions, trends could be tracked, interventions could be compared, and the question "is this getting better" had a real answer.

Public health borrowed its violence-prevention model from injury prevention, which had reduced motor vehicle deaths by half through systematic application of the same four steps. Define the problem (incidence, prevalence, characteristics). Identify risk and protective factors (who, when, where, why). Develop and test interventions (pilot, evaluate, iterate). Implement at scale (policy, dissemination, monitoring). Applied to relationship violence, the model exposes how much work the field had skipped. Many widely deployed programs had never completed step three. They went from "this should work" to "implement nationwide" without the trials in between.

Public health frames intimate partner violence as the product of nested risk levels: individual (history of childhood abuse, substance use), relationship (conflict patterns, economic dependence), community (concentrated poverty, weak social networks), and societal (gender norms, gun availability). The model is descriptively right and operationally hard. Interventions tend to live at one level; the risk lives across all four. Programs that target only individual behavior change while the community and societal factors are untouched produce small effects. The frame is honest about why.

A persistent tension: public health needs surveillance data, which works best when reporting is mandatory and identifiable. Intervention works best when victims trust the system enough to engage, which is undermined by mandatory identifiable reporting. James Mercy has argued for separating the two streams—anonymous epidemiological surveillance for population data, voluntary service systems for individual intervention—but in practice the streams keep collapsing into each other. Hospitals are asked to count and to refer using the same intake. The collision is built into the architecture.

The CDC's Dating Matters program, piloted from 2012, was the most rigorous attempt to take a public health approach to teen dating violence at scale. It combined school curricula, parent programs, community policies, and youth communications across multiple grades. The trial in Chicago, Baltimore, Fort Lauderdale, and Oakland showed reductions in perpetration and victimization of roughly fifteen percent compared with a standard curriculum. Modest, durable, replicable. The frame had produced something the criminal system could not: a measurable downstream reduction in the behavior, achieved before the violence occurred.

A recurring weakness of public health DV interventions is reliance on attitudinal measures—agreement with statements like "a man is justified in hitting his partner under some circumstances." Attitudes shift more easily than behavior. Programs report successful attitude change while behavior at population level stays flat. Douglas Kirby's reviews of adolescent sexual health programs found the same pattern in adjacent territory: knowledge and attitudes move, behaviors often do not. The frame's preference for measurable proximal outcomes can become a substitute for the harder question of whether anything that matters has changed.

Public health has consistently shown that intimate partner violence, child maltreatment, youth violence, sexual violence, and suicide cluster in the same populations and share risk factors. The CDC's Connecting the Dots report formalized this. The implication for intervention is significant: a program that reduces adverse childhood experiences may reduce all five forms downstream. The implication for politics is awkward: it competes with the framing of each form as distinct, with distinct advocacy communities and funding streams. The frame's strength is its integration; its political weakness is that integration threatens existing turf.

Jonathan Shepherd's Cardiff Violence Prevention Model showed that emergency departments share data on violent injuries—location, time, weapon used—with police while protecting patient identity. Police deploy resources accordingly. Violent injuries in Cardiff fell roughly forty percent over a decade compared with control cities. The model is public health at its best: data flows where it can act, individual privacy is preserved, and the intervention sits upstream of arrests. Adoption in the U.S. has been slow because the institutional cooperation required between health and law enforcement is rare.

Public health treats the perpetrator as a person whose behavior is shaped by risk factors and is therefore potentially modifiable. Batterer intervention programs grounded in this premise have effects ranging from small to moderate depending on design and population. The Duluth model, the cognitive-behavioral models, the trauma-informed models—each has evidence, none has dramatic effects. The frame's limitation is honest: changing established violent behavior in adults is hard, slower than primary prevention with adolescents, and unreliable enough that it cannot be the sole strategy.

Jacquelyn Campbell's danger assessment, now embedded in Lethality Assessment Programs across many states, allows first responders to triage cases by risk of homicide. Officers ask eleven questions at the scene; high scores trigger immediate connection to a DV advocate. Evaluations in Maryland and Oklahoma show increased service uptake by high-risk victims and reductions in near-fatal violence. This is public health at the case level: stratify by risk, allocate intensive intervention where the consequences of failure are highest. It does not require abandoning the criminal frame; it adds a layer.

The public health frame is allergic to questions of meaning, narrative, and moral structure. It can tell you that a woman in this zip code has a thirty percent higher lifetime risk of intimate partner violence. It cannot tell her, or anyone, how to make sense of what happened to her, or how to love again, or how to forgive a parent who let it happen. The frame does not pretend otherwise. The risk is that institutions built on the frame come to believe that what they measure is what matters, and that the unmeasurable is unreal. Law 0 humility requires the frame to know its own limit. The work it cannot do is still work that has to be done.

1. Mercy, James A., Mark L. Rosenberg, Kenneth E. Powell, Claire V. Broome, and William L. Roper. "Public Health Policy for Preventing Violence." Health Affairs 12, no. 4 (1993): 7–29.

2. Saltzman, Linda E., Janet L. Fanslow, Pamela M. McMahon, and Gene A. Shelley. Intimate Partner Violence Surveillance: Uniform Definitions and Recommended Data Elements. Atlanta: Centers for Disease Control and Prevention, 1999.

3. Campbell, Jacquelyn C. "Danger Assessment: Validation of a Lethality Risk Assessment Instrument for Intimate Partner Femicide." Journal of Interpersonal Violence 24, no. 4 (2009): 653–74.

4. Krug, Etienne G., Linda L. Dahlberg, James A. Mercy, Anthony B. Zwi, and Rafael Lozano, eds. World Report on Violence and Health. Geneva: World Health Organization, 2002.

5. Shepherd, Jonathan. "Criminal Deterrence as a Public Health Strategy." The Lancet 358, no. 9294 (2001): 1717–22.

6. Niolon, Phyllis Holditch, Megan C. Kearns, Jenny Dills, Kirsten Rambo, Shalon Irving, Tracy Armstead, and Leah Gilbert. Preventing Intimate Partner Violence Across the Lifespan: A Technical Package of Programs, Policies, and Practices. Atlanta: Centers for Disease Control and Prevention, 2017.

7. Kirby, Douglas. Emerging Answers 2007: Research Findings on Programs to Reduce Teen Pregnancy and Sexually Transmitted Diseases. Washington, DC: National Campaign to Prevent Teen and Unplanned Pregnancy, 2007.

8. Warshaw, Carole. Mental Health Treatment for Survivors of Intimate Partner Violence. Chicago: National Center on Domestic Violence, Trauma and Mental Health, 2014.

9. Wilt, Susan, and Sarah Olson. "Prevalence of Domestic Violence in the United States." Journal of the American Medical Women's Association 51, no. 3 (1996): 77–82.

10. Mercy, James A., Susan D. Hillis, Alexander Butchart, Mark A. Bellis, Catherine L. Ward, Xiangming Fang, and Michael L. Rosenberg. "Interpersonal Violence: Global Impact and Paths to Prevention." In Disease Control Priorities, 3rd ed., vol. 7, edited by Charles N. Mock et al. Washington, DC: World Bank, 2017.

11. Sumner, Steven A., James A. Mercy, Linda L. Dahlberg, Susan D. Hillis, Joanne Klevens, and Debra Houry. "Violence in the United States: Status, Challenges, and Opportunities." JAMA 314, no. 5 (2015): 478–88.

12. Silverman, Jay G., Anita Raj, Lorelei A. Mucci, and Jeanne E. Hathaway. "Dating Violence Against Adolescent Girls and Associated Substance Use, Unhealthy Weight Control, Sexual Risk Behavior, Pregnancy, and Suicidality." JAMA 286, no. 5 (2001): 572–79.