The effectiveness of an ED-based violence prevention program

Original Contributions

The effectiveness of an ED-based Violence prevention programB

Leslie S. Zun MD, MBAa,b,*, LaVonne Downey PhDa, Jodi Rosen MPHb

aDepartment of Emergency Medicine, Roosevelt University/Chicago Medical School, Chicago, IL, USA

bDepartment of Emergency Medicine, Mount Sinai Hospital, Chicago, IL 60608, USA

Accepted 9 May 2005


Background: Youth violence continues to be a problem in the United States, most prominent in the inner-city minority youth population. The recurrence rate for repeat violence has been reported from 6% to 44% with a 5-year mortality of 20%. This study describes the results of a program to reduce violence recurrence based in the ED.

Methods: Patients aged 10 to 24 years who were victims of interpersonal violence (excluding child abuse, sexual assault, and intimate partner violence) were randomly enrolled in the study in level 1 trauma center. The control group was given a written list of services, and the treatment group received an assessment and Case management for 6 months. Both groups were evaluated 6 and 12 months after enrollment in the study. The primary indicators of the success of the intervention were reduction of self- reported revictimization or arrest and state-reported incarceration and reinjury. The study was approved by the institutional review board. The results compared the change in treatment and control groups over the time using a combination of v2 and analysis of covariance.

Results: One hundred eighty-eight victims of interpersonal violence met the criteria and had the initial evaluation completed. By v2 analysis, the treatment group (96 subjects) and the control group (92 subjects) were similar in age, sex, and racial composition. The average age was 18.6 years (range, 11- 24), and 82.5% were boys. Most youth were African Americans (65.4%), followed by Hispanic (31.4%), whites (1.6%), or others (1.5%). A reduction in the self-reported reinjury rate was significantly reduced over time in the treatment group (v2 3.87, P = .05). There were no differences between the groups in the number of self-reported arrests, state-reported reinjuries via the trauma registry, or state- reported incarcerations ( P b .05).

Conclusions: The results of this study demonstrated a reduction in self-reported reinjury rate in the intervention group. Further research is needed to confirm if ED-based Violence prevention programs are effective in reducing other determinants for revictimization.

D 2006

B Supported by the Joyce Foundation, Woods Fund of Chicago, Michael Reese Health Trust, The Open Society Institute-The Center on Crime, Communities, and Culture, and Baxter International.

T Corresponding author. Department of Emergency Medicine, Mount Sinai Hospital Medical Center, Chicago, IL 60608, USA.

0735-6757/$ – see front matter D 2006 doi:10.1016/j.ajem.2005.05.009


The rate of violence-relatED recidivism is from 6% to 44% for victims of interpersonal violence [1-3]. It is postulated that a means to reduce this high reoccurrence rate is to address the unmet psychosocial needs of the youth. A common first point of contact, the ED, has done an inadequate job of addressing these needs. A study performed by Melzer-Lange et al [4] demonstrated that referral rate of victims of violence for follows-up is lower than patients seen for other problems in the ED.

The American Academy of Pediatrics Task Force on Adolescent Assault Victim Needs developed an ED-based model for victims of violence that included psychosocial, emotional, and spiritual support for the victims, family, and friends [5]. Anderson and Taliaferro [6] discuss the importance of integrating violence prevention into the delivery of health care. They recommend the development of case management, referral, and treatment programs for Trauma victims with high-risk behavior for violence. The Society for Academic Emergency Medicine Public Health

and Education Committee recommends that the role of the emergency physician is to provide information for the injured youth about available community resources [7].

This study was based on the current theory and practice of health education on changing health-related behavior. This theory is based on 3 concepts: (1) youth will alter their behavior if they believe that they are susceptible to violence,

(2) the behavior has appropriate severity, and (3) the young person understands the consequences as well as the benefits of the intervention [8]. ED-based intervention performed in this study capitalizes on the later constructs of this theory. We hypothesized that the intervention would have a significant effect on reducing the youth’s subsequent revictimization.


Assessment and evaluation tools were given prospec- tively to consecutive youth with parental consent (if younger than 18 years) in the ED of an inner-city, teaching, level I pediatrics and adult trauma center with approximate-

Fig. 1 Study enrollment and completion.

ly 43000 visits by an almost equal number of African Americans and Hispanic patients per year. The study took place from July 1998 to October 1999. The inclusion criteria were individuals aged 10 to 24 years, presentation to the ED as victims of interpersonal violence with criteria for life- threatening or limb-threatening emergencies (established by the principal investigator, LZ), residence within the hospi- tal’s service area, and the ability to communicate. The exclusion criteria included those victims of violence resulting from self-declared child abuse, sexual assault, and intimate partner violence (programs already existed to address these problems), severe disability, or those who were unable to be followed up. Patients who went to the operating room would be interviewed on the floor after their surgery. The inclusion age was chosen to mirror the age group identified as byouthQ according to the Youth Risk Behavior Surveillance study performed by the Center for Disease Control [9]. The study was approved by the hospital’s institutional review board.

An evaluation instrument for use in the ED and an assessment tool were developed before the initiation of the linkage of an ED with a health care system and social service agency and case management process for the youth. We used content validity for these tools. The health system used in the study consists of a 450-bed acute care hospital, rehabilitation hospital, medical group, community institute, and a 50-clinic Primary care clinic. The services used included primary care, gang-related tattoo removal, and dental care. The social service agency used in the study, The Boys & Girls Clubs of Chicago, had 20 centers in the city with programs in personal development and education, citizenship and leadership development, cultural enrichment, outdoor and environmen- tal education, social recreation, and health and physical education. Some services, not available at the health care system or the social service agency, were obtained outside this linkage, including legal assistance, alternative education, spiritual counseling, and financial assistance.

ED personnel or the case managers would identify youth who were victims of violence and met the entrance criteria. The youth and their guardian (if younger than 18 years) were approached while in the ED to ascertain their interest in enrolling in the study. We proposed that the young person’s mere presentation to the ED met the criteria of susceptibility to violence and the youth’s possession of the entrance criteria for the study as having the appropriate severity. Consenting youth and their guardian were explained the outline of the program and were randomly assigned by envelope selection with either treatment or control group noted in the envelope. The case managers began the assessment and evaluation in the ED but may have completed these tools in the hospital or at home (if the patient was discharged). Patients who presented when a case manager was not present were contacted by phone or visited in the hospital to determine their interest in study enrollment.

The treatment group received an assessment tool developed by a research team consisting of an emergency

physician, social worker, and public health expert (Appen- dix 1). The constructs of the assessment tool included questions about school, gang activity, peers, family, violence, delinquency, and gun accessibility. The assess- ment tool was used to identify and address the psychosocial needs of the youth; all constructs led to referrals except for gun accessibility because appropriate referral for this problem could not be located. Patient referrals were standardized by weekly discussions of new cases by the principal investigators. The evaluation tool was developed by Patrick Tolan, PhD, and consisted of the following scales: Parenting Practices Questionnaire [10], Family Relations Scale [10,11], The Child Self-Report of Delin- quency [10,12], CYDS Stress Measure [13], Parents and Peers [14], Peer Delinquency [15], Future Expectations (PH Tolan, oral communication, 2000), Attitudes and Beliefs about Achievement [15], Aspirations/Values [15], and Social Competence [16] (according to PH Tolan; L Aber, oral communication, 1988). This evaluation tool has been used for longitudinal studies of inner-city youth [17].

The treatment group was case managed for 6 months, and the control group only received a brochure describing the same available services. Case managers provided access to primary and preventive health care and social services, individualized case management plans, and anger manage- ment and conflict resolution training. The case managers would contact the referral sources and arrange for enroll- ment into various programs and clinics. Case management demonstrated to the youth not only the potential conse- quences, but also the benefits of the intervention. The case managers met with the youth weekly for the first 2 months, every other week for the second 2 months, and monthly thereafter. Follow-up evaluations of both groups by the case managers in person occurred at 6 months (wave 1) and at 12 months (wave 2) after entry into the study. Each case manager was limited to 20 patients at a time, and the study was capped when they reached this limit.

Table 1 Characteristics of the groups

ns indicates not significant.



Statistical significance


Range (y)























African American
















0.03, ns



0.002, ns



0.06, ns



3.87, .05



v2, P

% Yes

Outcome variable

Victim of violence at posttest Treatment (n = 59)

Control (n = 62) Return visit to the ED

Treatment (n = 93) Control (n = 95)

Arrested after enrollment Treatment (n = 93) Control (n = 95)

Incarcerated after enrollment Treatment (n = 93) Control (n = 95)

Table 2 Results of v2 analyses for intervention impact on targeted behaviors

Evaluation of the program’s effectiveness used the following primary indicators: reduction of self-reported reinjury and arrests and state-reported incarceration and reinjury. The state’s department of corrections list of inmates and the state’s trauma registry were used as source material. Additional evaluation was performed by comparing the change in treatment and control groups over time. Second- ary effectiveness measures included violent delinquency, nonviolent delinquency, drug use, and violence victimiza- tion as measured by the evaluation tool. Accordingly, a combination of v2 and analysis of covariance (ANCOVA) were the primary analytic methods for the results.


Six hundred twenty-six victims of interpersonal violence were seen in the ED from July 1998 to October 1999, and 222 met the study entry criteria (Fig. 1). Most of the ineligible patients were out of the service area (44%) or were due to the enrollment cap (25%). Of 238 patients meeting the entrance

criteria, 16 youth and/or their parents refused; 27 young persons of these were lost before evaluation completion, and 7 had significant incomplete entries. One hundred eighty- eight victims of interpersonal violence, 96 in the treatment group and 92 in the control group, met the criteria and had the 6-month evaluation completed (wave 1). The average age was 18.6 years (range, 11-24), and 82.5% were boys (Table 1). Most youth were African Americans (65.4%), followed by Hispanic (31.4%), whites (1.6%), or others (1.5%). There were 59 patients in the treatment group, and 62 patients in the control group completed wave 2 (12 months). Seventy-one patients were lost between time zero and wave 1, and 34 were lost in both groups between waves 1 and 2.

The primary analysis was a combination of v2 for dichotomous outcomes and ANCOVA for continuous out- comes. The v2 analysis was conducted to compare the treatment and control groups on proportion of sample group in that condition. The only significant difference between treatment and control groups was for self-reported reinjury (Table 2). There was no difference found between the control and treatment groups for the other targeted behaviors: arrest (after enrollment), return visits to the ED, and incarceration after enrollment.

Table 3 presents the results of ANCOVA on the following targeted behaviors: violent delinquency, nonviolent delin- quency, drug use, and violence victimization. Provided in this table are the mean values, SDs, and ANCOVA results for the time, condition, and time x condition effects of the intervention for 2 waves of the program on the control and intervention groups. Although not statistically significant, there was a drop in violent delinquency from waves 1 to 2 for both the intervention and control groups. There is also a slightly lower level of drug use for the control group in both waves 1 and 2. However, there was no evidence that the intervention led to greater reduction in drug use. By ANCOVA, no significant differences were found between the other targeted behaviors of violent delinquency, nonvi-

Table 3 Results of ANCOVA analyses for intervention impact on targeted behaviors (2 waves of data)

Outcome variable

Assessment 1



Assessment 2



ANCOVA results

Violent delinquency

T: F2,114 = 2.96, P b .10

Treatment (n = 62)





C: F1,114 = 0.20, ns

Control (n = 58)





TxC: F2,114 = 0.83, ns

Nonviolent delinquency

T: F2,114 = 0.31, ns

Treatment (n = 62)





C: F1,114 = 0.51, ns

Control (n = 58)





TxC: F2,114 = 0.89, ns

Drug use

T: F2,114 = 0.56, ns

Treatment (n = 62)





C: F1,114 = 1.67, P = .20

Control (n = 58)





TxC: F2,114 = 0.01, ns

Violence victimization

T: F2,114 = 0.41, ns

Treatment (n = 62)





C: F1,114 = 0.02, ns

Control (n = 58)





TxC: F2,114 = 0.00, ns

T indicates the time (wave 1 vs wave 2 vs wave 3) main effect; C, the condition (treatment vs control) main effect; TxC, the time x condition interaction effect. TxC is the effect of primary interest.

olent delinquency, drug use, and violence victimization. All other comparisons of self-reported delinquency and violence were also not found to be significant.

Although not statistically significant, there was a substantial drop in violent delinquency from pretest to posttest, whether in the intervention or control group (Table 3) ( P b .05). Similarly, for the precomparison to postcomparison of drug use, the treatment group had slightly lower levels of drug use than did the control group at both assessments. However, there was no evidence that the intervention led to greater reduction in drug use.


This study demonstrated a reduced revictimization rate in the intervention group with little change in the other study determinants. This result was a sentinel success for a violence prevention program that attempts to modify the health behavior of youth, using a hospital-based interven- tion. This paper provides a description of an ED-based violence prevention program and its effectiveness on the recurrence rate of intentionally injured youth.

A review of the emergency medicine and pediatric literature demonstrated few prior studies on violence prevention services initiated or based in the ED. This study linked the injured youth seen in an ED to a social service agency and health care institution to provide services to the youth with case management for 6 months. Few other hospital-based studies have been performed, and so far, none have shown promise to reduce revictimization. DeVos et al [17] performed a hospital-based violence prevention program based on the Boston Violence Prevention Program 6-step intervention [18]. However, these investigators were unable to demonstrate any positive results.

We were disappointed that the results were not as robust for the other primary and secondary study determinants. The reasons for these limited results are uncertain but are thought to be related to funding limitations, qualities of the adolescent population, and problems related to the provision of psychosocial services. The study had to limit the number of participants who were taken into the program and the length of case managed, because of funding constraints. The study dealt with a high-risk sample population at a time in their development during which they are highly mobile. Many of the youth are involved heavily in street life with only minimal connection to conventional institutions and family. Thus, they were hard to find, hard to get to sit for interviews, and often ambivalent about using programmatic resources. These factors not only limited the number of youth enrolled and their retention rates, but also speak to issues in undertaking such an approach to prevent violence. It will require vigilance, persistence, and considerable resources that can be daunting to address these factors. Some of the services the youth were referred to were not of the caliber we had expected.

The estimated costs of this program would preclude its broad-based acceptance. It cost approximately US$45 000 per year plus benefits for a case manager to enroll and follow an estimated 20 youth for 6 months in the program. This estimated cost does not include transportation, youth incentives, and fees or tuition for the services (if there was one). It is doubtful that public funding would be used for such an intervention even if it were shown to reduce societal costs associated with these injuries and death. On the other hand, it is unlikely that the case management activities could be absorbed by ED personnel in this era of reduced reimbursement and limited resources. Modifications of the program to provide referrals with limited, if any, case management is untested but worth considering.

Many limitations are found in this study. Although both the treatment and control were asked the same questions, reporting and recall biases were significant concerns noted in the study. The study design would not facilitate a blinding process that may provide more reliable results. The control group received a modified intervention and was therefore considered benhancedQ control rather than a pure control group based on request of the institutional review board. The determination of the referral needs of the youth was performed based on the best estimate of the needs using an unvalidated assessment tool in the intervention group, not based on violence reduction techniques. All of the compo- nents of the evaluation tool, except for Future Expectations and Social Competence, were validated, but the combina- tion has not been. The study was limited by the availability, effectiveness, and reliability of services at the time of the study. The study was limited by those youth who were excluded, lost to follow-up, or had incomplete documents. Unfortunately, the study population has significant mobility and was commonly unavailable when the case managers attempted to interview them. This study and the results noted were limited by the duration of case management and follow-up to 6 months. Perhaps, a longer period of at least 1 year may produce improved results. We were unable to determine whether the youth had been injured from violence perpetration or victimization. The study was limited by the turnover of case managers.

Little controversy surrounds the need to reduce reoccur- rence of violence in inner-city youth population. The means and interventions to accomplish these goals in the health care setting are elusive. Although difficult to fund and costly to maintain, further studies using similar interventions beginning in the ED with longer case management and follow-up would be valuable in determining if violence prevention activities can produce sustained reduction in the incidence of recidivism for this disease process.


This study presents a novel violence prevention program based in the ED. Although the success of the program was

limited, further work in ED-based violence prevention studies is needed to determine the further effectiveness of such efforts.

What this study adds

Unfortunately, adolescent violence is prevalent in most inner-city communities in the United States. Violence Prevention efforts to stop recidivism have been onerous with limited success. Further studies are needed to determine the optimal means to reduce the level of violence recidivism. This study demonstrated a reduction the amount of repeat violence episodes in youth who were randomized to the intervention group. This intervention began in the ED, determined the needs of the youth, and linked the youth to psychosocial services in the hospital and the community.


The authors thank Andrea Rossi from the University of Illinois Institute for Juvenile Justice for her assistance in the data collection and analysis. They also thank the Policy and Advisory Board for the bWithin Our ReachQ program for their advice in program development and Leslie Stein- Spencer from the Illinois Department of Public Health and Cook County juvenile justice system.

Appendix A. Appendix 1

Assessment Tool Sample

After interviewing this individual, how would you identify the nature of current stressors in their lives:

5 Problems related to alcohol or drug abuse (specify)

5 Educational problems (specify)

5 Family problems (specify)

5 Occupational problems (specify)

5 Housing problems (specify)

5 Financial problems (specify)

5 Problems with access to health care (specify)

5 Problems related to interaction with the legal system/ crime (specify)

5 Mental health (specify)

5 Other problems (specify)

Of the 10 items listed above, which 3 problems do you feel are the most severe or in need of immediate attention?

Short term:




Long term: 4)


Do you see barriers to this client participating in our program? (eg, attitude, lack of interest, etc)

What is your general impression of this person?

Can you briefly assess them? (eg, age, sex, circumstances) Comments/recommendations:

Date of review Within Our Reach staff:


  1. Weisbeski SD, Bivins B, et al. Urban trauma: a chronic recurrent disease. J Trauma 1989;29:940 – 7.
  2. Morrissey T, Byrd CR, Deitch EA. The incidence of recurrent penetrating trauma in an urban trauma center. J Trauma 1991;127: 668 – 70.
  3. Smith R, Fry WR, Morabito DJ, Organ CH. Recidivism in an urban trauma center. Arch Surg 1992;127:1536 – 8.
  4. Melzer-Lange M, Lye PS, Calhoun AD. Advised follow-up after emergency treatment of adolescents with violence related injuries. Pediatr Emerg Care 1998;14:334 – 7.
  5. Bell C, Jenkins EJ, Kpo W, Rhodes H. Response of emergency rooms to victims of interpersonal violence. Hosp Community Psychiatry 1994;45:142 – 6.
  6. Anderson R, Taliaferro EH. Injury prevention and control. Am J Emerg Med 1997;15:489 – 98.
  7. Christoffel K, Barlow B, Bell C, et al. Task force on adolescent assault victims needs, adolescent assault victims needs: a review of issues and a model protocol. Pediatrics 1996;98:991 – 1001.
  8. Muelleman R, Sanson TG, Gerson L, et al. An emergency medicine approach to violence throughout the life cycle. Acad Emerg Med 1996; 3:708 – 15.
  9. Gorman-Smith D, Tolan P, Zelli A, Huesmann LR. Assessment of family relationship characteristics: a measure to explain risk for antisocial behavior and depression among urban youth. J Fam Psychol 1996;9:212 – 23.
  10. Tolan PH, Guerra NG, Montaini-Klovdahl LR. Staying out of harm’s way: coping and the development of inner-city children. In: Wolchik SA, Sandler IN, editors. Handbook of children’s coping: linking theory and intervention. New York7 Plenum Press; 1997. p. 453 – 80.
  11. Elliott DS, Huizinga D, Ageton SS. Explaining delinquency and drug use. Beverly Hills (Calif)7 Sage; 1985.
  12. Gorman-Smith D, Tolan PH, Henry DB. A developmental-ecological model of the relation of family functioning to patterns of delinquency. Journal of Quantitative Criminology (in press).
  13. Huizinga D, Loeber R, Thornberry TP. Recent findings from the program of research on the causes and correlates of delinquency. Washington (DC)7 Office of Juvenile Justice and Delinquency Prevention; 1995.
  14. Henry D, Tolan PH, Gorman-Smith D. Longitudinal family and peer group effects on violence and non-violent delinquency. J Child Clin Psych (in press).
  15. Gaffney LR, McFall RM. A comparison of social skills in delinquent and nondelinquent adolescent girls using a behavioral role-playing inventory. J Consult Clin Psychol 1981;9:959 – 67.
  16. Gorman-Smith D, Tolan P. The role of exposure to community violence and developmental problems among inner-city youth. Dev Psychopathol 1998;10:101 – 16.
  17. DeVos E, Stone DA, Goetz MA, Dahlberg LL. Evaluation of a hospital-based youth violence intervention. Am J Prev Med 1996;12(Suppl 2):101 – 8.
  18. Kinchen K, Williams BL, Ross JG, et al. Youth risk behavior surveillance– United States 1997. MMWR Morb Mortal Wkly Rep 1998;14:1 – 89.

Leave a Reply

Your email address will not be published. Required fields are marked *