research help

From the So Constans); Louisiana S Affairs Men Constans); D Orleans, Lo University o ment of P Pennsylvani

Address Missouri-Sa Boulevard, nanneyj@um

0749-379 http://dx

Published

Criminal Behavior and Repeat Violent Trauma

A Case–Control Study

John T. Nanney, PhD, Erich J. Conrad, MD, Michael McCloskey, PhD, Joseph I. Constans, PhD

u D tat tal e ui f M sy a co int St sl 7/ .do

by

Introduction: Repeat violent injury is common among young urban men and is increasingly a focus of trauma center–based injury prevention efforts. Though understanding risk factors for repeat violent injury may be critical in designing such interventions, this knowledge is limited. This study aims to determine which criminal behaviors, both before and after the initial trauma, predict repeat violent trauma. Gun, violent, and drug crimes are expected to increase risk of subsequent violent injury among victims of violence.

Methods: A case–control design examined trauma registry and publicly available criminal data for all male patients aged o40 years presenting for violent trauma between April 2006 and December 2011 (N¼1,142) to the sole Level 1 trauma center in a city with high rates of violence. Logistic regression was used to determine criminal behaviors predictive of repeat violent injury. Data were obtained and analyzed between January 2013 and June 2014.

Results: Regarding crimes committed before the first injury, only drug crime (OR¼5.32) predicted repeat violent trauma. With respect to crimes committed after the initial injury, illegal gun possession (OR¼2.70) predicted repeat victimization. Initiating gun (OR¼3.53) or drug crime (OR¼5.12) was associated with increased risk. Conclusions: Prior drug involvement may identify young male victims of violence as at high risk of repeat violent injury. Gun carrying and initiating drug involvement after the initial injury may increase risk of repeat injury and may be important targets for interventions aimed at preventing repeat violent trauma. (Am J Prev Med 2015;49(3):395–401) Published by Elsevier Inc. on behalf of American Journal of Preventive Medicine

Introduction

V iolent trauma plagues young men in many urban, typically African American, commun- ities.1–4 Violence is the leading cause of death for

African American men aged 18–35 years and remains a

theastern Louisiana Veterans Healthcare System (Nanney, epartment of Psychiatry (Nanney, Conrad, Constans), e University School of Medicine; South Central Veterans Illness Research, Education, and Clinical Center (Nanney,

partment of Psychology (Constans), Tulane University, New siana; Department of Psychological Sciences (Nanney), issouri-Saint Louis, Saint Louis, Missouri; and the Depart-

chology (McCloskey), Temple University, Philadelphia,

rrespondence to: John T. Nanney, PhD, University of Louis, Department of Psychological Sciences, 1 University

adler Hall Room 236, Saint Louis MO 63121. E-mail: .edu. $36.00 i.org/10.1016/j.amepre.2015.02.021

Elsevier Inc. on behalf of American Journal of Preventiv

leading cause of death through age 40 years.3,4 For victims of violence, repeat injury is common,5–9 and trauma center–based interventions to reduce repeat violent trauma have recently emerged.10–15 Such inter- ventions have yielded only mixed results, possibly because most interventions focus on enrolling patients in general outpatient case management services rather than changing specific risk behaviors.16 Interventions targeted at specific behaviors known to increase risk of later violence/violence injury may have greater chances of success.17

Certain criminal behaviors—specifically violent, gun, and drug offenses—may be strong candidate risk factors for repeat violent trauma.2,5,7 Violent behavior invites violent retaliation. Assaults are more likely to involve more-severe gunshot injuries if assailants expect the target to be similarly armed.18 Violence also permeates illicit drug economies, as disputes cannot be settled

e Medicine Am J Prev Med 2015;49(3):395–401 395

Nanney et al / Am J Prev Med 2015;49(3):395–401396

legally.19 Other forms of crime, like unarmed, non- confrontational property crime (e.g., auto theft) may be less likely to provoke retaliatory violence and may be less associated with repeat injury risk. Empirical studies focused on crime and trauma recidivism are generally consistent with this pattern, but methodologic limita- tions preclude definitive conclusions. One study5 found that violent, gun, and drug crimes were more common among repeat victims of violence than among patients injured accidentally. This study, however, did not com- pare repeat to single episode victims of violence and it relied exclusively on survey methods to assess criminal- ity. A second study7 found that gun, drug, and violent crime, but not property crime or crime in general, were more common among repeat than single-episode trauma patients. Nonetheless, this study7 combined violent and accidental trauma patients, so it is not clear if these findings hold for those who specifically experience violent trauma. Extant literature also has not differentiated between

crimes committed before and those committed after first injury. This issue is important for the development of trauma center–based interventions. Understanding which behaviors occurring prior to the first trauma are associated with rehospitalization is useful in identifying those initial trauma victims most at risk of future violent trauma. However, this information may be less relevant for intervention development because these historic risk factors may be static or unchangeable through intervention. By contrast, understanding the risk behav- iors that occur after the initial hospitalization is critical for development of hospital-based interventions,17 as these are the behaviors such interventions can influence most directly. For example, the experience of trauma may lead to new risk behaviors (e.g., regular gun carrying or drug involvement).2,20,21 Whether such changes increase risk of repeat trauma, making them potential targets for intervention, has not been examined directly. The present case–control study aims to identify differ-

ences in criminal behavior between repeat and single- episode victims of violence. To address limitations of previous studies, administrative hospital and criminal data are examined, and criminal behaviors occurring prior to the initial trauma are coded separately from those occurring after first injury. It is hypothesized that violent crime, gun possession, gun use, and drug crime, both prior to and following initial trauma, would predict repeat violent trauma, but that property crime, both before and after, would not. Finally, as a stronger demonstration that risk behaviors may be useful targets of intervention, it is examined whether those who do not engage in specific criminal behaviors prior to the first

trauma but begin engaging in that crime afterwards have higher chances of repeat violent injury than those who continue to abstain. It is expected that initiating gun possession, gun use, and violent or drug crime following a violent trauma would increase risk of repeat violent injury, but that initiating property crime would not.

Methods Study Population

The study was approved and a waiver of informed consent was granted by the IRB of the Louisiana State University Health Sciences Center, New Orleans. A study population consisting of all adult male patients aged r40 years from Orleans Parish who were admitted to the Spirit of Charity Level 1 Trauma Center (SOCTC) with a violent injury between April 2006 and December 2011 and who survived their initial injury was identified from the trauma center trauma registry (N=1,243). The SOCTC is the only Level 1 trauma center in New Orleans and thus treats all severe violent injuries (e.g., gunshot wounds) that occur in the metro- politan area. From Hurricane Katrina in August 2005 until the Trauma Center’s reopening in April 2006, there is an 8-month gap in the trauma registry. For this reason, only data after April 2006 were examined.

Violent trauma was operationalized as hospital presentations with an ICD-9 e-code of 960–969, specifically indicating inten- tional violent injury. Cases of violent trauma recidivism were identified by linking trauma center presentations according to patient name, Social Security Number (SSN), and birth date. Patients who presented with a violent trauma between April 2006 and December 2011 and then presented with at least one additional violent trauma between April 2006 and December 2012 were classified as violent trauma repeaters (n=93). The control group consisted of patients who presented to the trauma center with a violent trauma between April 2006 and December 2011 but no additional violent traumas from April 2006 to December 2012 (n=1,150). All databases with patient identifiers were destroyed following the linking of hospital and criminal databases.

Data Sources

The trauma registry contains demographic, medical, and patient outcome data on patients for whom the hospital trauma activation protocol is initiated. Demographic data included date of injury; name; birth date; SSN; gender; race (self-report or if necessary as determined by medical staff); age; and ZIP code of residence. Cause of injury data included the ICD-9 e-code identifying the mechanism (e.g., gun or knife) and apparent intent (i.e., inten- tional versus accidental) of injuries. Criminal data were retrieved from the Orleans Parish Criminal District Court docket master. Patients were linked to criminal records by their name and date of birth. Dates and nature of all criminal convictions were recorded. Criminal behavior was classified according to five categories: gun possession, gun use, drug crime, violent crime, and property crime. Gun possession crimes were defined as convictions that only involved the illegal possession or use of a firearm, without any use or threat of use against another person. Gun use crimes required

www.ajpmonline.org

Nanney et al / Am J Prev Med 2015;49(3):395–401 397

firearm use or threat of use against others. Violent crimes were defined as all those involving interpersonal aggression or violence. Drug crimes were defined as those that included the possession or distribution of illicit substances. Property crimes were defined as those that require the unlawful entry or unlawful taking, pos- session, or destruction of another person’s property. Criminal behavior categories, with the exception of gun

possession, were not mutually exclusive, so each criminal act could count toward multiple categories (e.g., shooting a person would count as a violent crime and a gun use crime). To better isolate the impact of mere gun possession, gun possession and gun use were coded to be mutually exclusive. That is, if an individual used a gun in a crime at any point during a given time period (i.e., before or after initial injury) they could not be coded positive for gun possession during that time period. Separate variables were created for crimes committed before and crimes committed after the initial injury. For those with multiple violent injuries, crime after the first trauma included only those crimes that occurred before the last violent injury in order to focus only on crimes that could logically contribute to repeat trauma risk. Trained research assistants conducted the crime ratings. Inter-rater reliability was assessed by having independent raters separately code criminal history for 20% of patients. Inter-rater reliability was excellent (κ=0.90).

Statistical Analysis

Data were analyzed in April–June 2014 using SPSS, version 21.0. A multivariate logistic regression model was used to examine the independent contribution of crimes in predicting repeat trauma. Criminal behaviors occurring prior to the initial injury were entered at Step 1. Criminal behaviors occurring after the initial trauma were then entered at Step 2. To examine how behavior change following initial trauma may

impact repeat injury risk, patients who did not engage in a given

Table 1. Demographics and Injury Characteristics

Total (n¼1,243)

Repea (n¼93

Age (years), M (SD) 26.55 (6.12) 23.70 (5

18–25 622 (50.1) 67 (7

26–32 377 (30.3) 16 (1

33–40 244 (19.6) 10 (1

Race

Black 1089 (82.8) 90 (9

White 79 (6.4) 2 (2

Asian 11 (0.9) 0 (0

Other 124 (10.0) 1 (1

Time to second injury (years), M (SD) 1.68 (1

First injury gunshot 860 (69.2) 77 (8

Second injury gunshot 77 (8

Note: Boldface indicates statistical significance (po0.01) Values are n (%) u

September 2015

form of crime prior to the initial trauma were coded dichoto- mously as to whether they either (1) continued to abstain from that type of crime or (2) began engaging in that type of crime. For each type of crime, a separate logistic regression was conducted with repeat victimization as the dependent variable and this crime initiation variable as the independent variable. History of other types of crime was controlled. A potential study confounder is time following initial injury in

which crime can be observed. For those without re-injury, there was no repeat injury to signal end of observation, which then continued until the end of the study (December 2012). The observation period was thus considerably longer among those without later injury compared with repeat victims (F=135.39, po0.001), potentially biasing results. Observation time after the first injury was thus controlled in all analyses. Race was controlled in all analyses because African Americans may be disproportion- ately likely to receive certain criminal convictions22 and be assaulted with a weapon.18

Results Demographics and injury characteristics appear in Table 1. Repeat victims of violent trauma (n¼93) comprised 7.5% of the overall sample (N¼1,243). Repeat violent trauma victims were significantly younger than those with only a single violent injury and were more likely to have initially presented with a gunshot injury than single-episode victims. Though African Americans predominated in the overall population of violently injured young men (82.8%), repeat victims of trauma were almost exclusively black. Rates of criminal con- viction, both before and after the first injury, are included

t )

Single episode (n¼1,150) χ2 F p-value

.29) 26.78 (6.13) 22.20 o0.001 2.0) 556 (48.3)

7.2) 360 (31.3)

0.8) 234 (20.3)

6.8) 939 (81.7) 13.81 o0.001

.2) 77 (6.7) 2.99 0.08

) 11 (1.0) 0.90 0.34

.1) 123 (10.7) 8.87 0.003

.37)

2.8) 783 (68.1) 8.73 0.003

2.8)

nless otherwise noted.

Table 3. Multivariate Logistic Regression of Crime and Trauma Recidivism

Crime Wald OR (95% CI)

Before only

Race (black/not black) 7.77** 5.32 (1.64, 17.25)

Gun possession before 0.04 1.11 (0.54, 2.21)

Gun use before 0.13 0.75 (0.16, 3.54)

Violence before 0.71 1.36 (0.70, 2.46)

Drug crime before 5.43** 1.71 (1.09, 2.69)

Property crime before 0.17 0.88 (0.47, 1.63)

Table 2. Frequency of Crime

Number (%) with conviction before

first injury

Number (%) with conviction after

first injury

Gun possession

105 (8.4) 72 (5.8)

Gun use 25 (2.0) 19 (1.5)

Violence 148 (11.9) 121 (9.8)

Drug 478 (38.5) 193 (15.5)

Property 172 (13.8) 77 (6.2)

Nanney et al / Am J Prev Med 2015;49(3):395–401398

in Table 2. Drug crime was most common and gun crimes the least common, both before and after initial injury. Results of the multivariate logistic regression (Table 3)

indicate that when considering only crimes prior to the initial trauma (Step 1), only pre-injury drug crime significantly predicted repeat trauma (p¼0.01). When crimes committed after the initial trauma were consid- ered (Step 2), illegal gun possession after the initial injury significantly predicted repeat violent trauma (p¼0.03) and drug crime committed after the first injury trended toward significance (p¼0.054). Drug crime before the initial injury remained a significant predictor (p¼0.02). A series of logistic regressions examining how crime

initiation following the first trauma related to risk of repeat trauma (Table 4) found that initiating illegal gun possession (p¼0.01) and drug crime (p¼0.01) signifi- cantly predicted repeat trauma. Initiating gun use, violence, and property crime did not.

Before and aftera,b

Race (black/not black) 4.45** 3.56 (1.10, 12.18)

Time after first injury 82.26** 0.91 (0.90, 0.93)

Gun possession before 0.02 1.06 (0.50, 2.24)

Gun use before 0.08 0.77 (0.13, 4.53)

Violence before 0.61 0.76 (0.39, 1.51)

Drug crime before 5.19** 1.80 (1.09, 2.97)

Property crime before 0.01 0.97 (0.49, 1.89)

Gun possession after 4.95* 2.70 (1.13, 6.48)

Gun use after 0.68 0.37 (0.03, 3.97)

Violence after 0.17 1.33 (0.52, 3.40)

Drug crime after 3.70 1.93 (0.99, 3.76)

Property crime after 0.26 0.76 (0.27, 2.14)

Note: Boldface indicates statistical significance (*po0.05; **po0.01). aORs are adjusted by including in Step 2 the span of time covered when evaluating criminal behavior after the first injury.

bHosmer-Lemeshow model goodness of fit χ2(8)¼9.78, p¼0.28.

Discussion Gun and drug crimes, as expected, predict repeat violent trauma. The relationship of these criminal behaviors to repeat trauma appears to be more complex than pre- viously recognized, however. At the time of the initial injury, only a history of drug crime predicts repeat victimization. This risk continues even when controlling for subsequent criminal behavior. Once one becomes involved and identified with the illicit drug market, it may be difficult to extricate oneself from the violent social milieu and intergroup conflict that surround it.19 Initia- tion of drug crime following first injury is associated with increased risk. Victims of trauma may turn to substances in order to self-medicate,2,23 increasing their vulnerability due to exposure to this violent market. Surprisingly, overall drug crime following first trauma is only margin- ally significant. Perhaps, for those already involved in the drug market, additional drug crime confers only small

incremental risk. History of gun crime at first injury is not associated with increased risk. Only after initial injury does gun crime emerge as a predictor of later victim- ization, and this is only for gun possession, not gun use. For the emblematic patient at highest risk for repeat violent trauma, an initial injury may enhance recognition of the violence associated with the illicit drug market leading to an increase in weapon carrying as a means of protection.2,20,21 The present results underscore the grave risks that may be associated with this, as it appears that initiating gun possession following the initial injury is an important determinant of an individual’s risk for repeat violent injury. It is unclear why weapon carrying is a risk factor after, but not before, the initial injury. Perhaps after an initial injury, individuals may display weapons more openly to deter possible assailants. Being known to carry a weapon increases the likelihood that, when conflict does occur, the other party will arm themselves similarly,

www.ajpmonline.org

Table 4. ORs for Initiation of Crime After Initial Trauma Predicting Trauma Recidivism

Crime initiateda n Unadjusted OR (95% CI) Wald AORa (95% CI)

Gun possession 1,116 2.15 (0.98, 4.69) 7.45 3.53 (1.43, 8.73)

Gun use 1,218 0.73 (0.10, 5.51) 0.36 0.59 (0.06, 4.62)

Violence 1,092 0.70 (0.27, 1.77) 0.08 0.87 (0.32, 2.38)

Drug crime 764 2.08 (0.89, 4.89) 9.03 5.12 (1.77, 14.84)

Property crime 1,070 1.38 (0.53, 3.57) 0.24 1.30 (0.45, 3.74)

Note: Boldface indicates statistical significance (po0.01). aORs are adjusted for criminal history prior to the initial trauma, race, and individual differences in the span of time covered when evaluating criminal behavior after the first injury.

Nanney et al / Am J Prev Med 2015;49(3):395–401 399

leading to higher chances of severe injury.18 Surprisingly, the actual use of a gun in a crime and violence more generally do not predict repeat violent injury. Such crimes carry higher chances of lengthy prison terms, such that some individuals engaging in them may be protected from repeat violent trauma because of their incarceration. Some convicted of gun possession may also have been preparing to use the weapon absent legal intervention. Finally, it is noteworthy that African American race is associated with repeat injury even controlling for other factors. Although it is possible that SES could in part explain this relationship, this finding is consistent with literature suggesting that a bias in perceiving African American men as dangerous makes them more likely to be targets of more severe, armed assaults when conflicts emerge.18

Differentiating between risk behaviors that occur before and those that occur after the initial trauma allows us to provide more nuanced clinical guidance than prior research. A history of drug crime at first injury marks violent trauma patients as higher risk for violent re- injury. Trauma centers may thus benefit from routinely screening for drug involvement and targeting interven- tions to this higher-risk group. Primary substance abuse prevention models24 may help reduce repeat violence by preventing involvement in the violent drug market in the first place. The methods used in the present study allow us to conclude with greater confidence that certain behaviors occurring after the initial injury, particularly gun carrying and drug involvement, may be appropriate targets for trauma center–based interventions. Fortu- nately, such interventions for both are developing. A trauma center–based intervention directly targeting gun- carrying adolescents was recently found to reduce chances of continued firearm carrying at 1 year of follow-up.25 Though replications with adults are required, this finding—in combination with the present findings—suggests that direct behavioral intervention regarding gun carrying may have the potential to reduce repeat violent trauma. Such interventions likely should

September 2015

not be limited to those who already have a history of carrying guns. Indeed, many of those at risk may not yet have begun engaging in significant gun carrying at the time of the injury.2,20,21 Clinicians may anticipate that patients without histories of gun use may be considering arming themselves in the wake of their trauma. Develop- ment of preventive interventions that directly inquire about plans for gun use after the injury and attempt to guide this decision-making process may be needed. With respect to drug involvement, general substance abuse prevention strategies are widely available20,26 and trauma center–based interventions for illicit drug use are developing.27

Limitations Criminal data include only illegal behaviors that warrant criminal prosecution and conviction. Such data cannot determine if purely legal gun possession confers risk of violent injury, though a relationship between legal gun possession and violent injury has been suggested.28–31

Results also may not extend to less severe illegal behaviors that do not lead to arrest, prosecution, and conviction (e.g., occasional recreational drug use). Whether risk of repeat injury from drug crime is a function of drug use or drug distribution cannot be determined given that both often may yield similar convictions for drug possession.32 The trauma registry contains only the most critical injuries that require treatment at a Level I trauma center. Results thus may not extend to the majority of violent injuries that do not require this level of care. Our study is limited to a single metropolitan area, New

Orleans, at a unique period in its history (i.e., the aftermath of Hurricane Katrina). Owing to the high rates of migration during this period, some participants in this study may have been injured or may have engaged in crime in another location. Even participants who resided in New Orleans throughout the duration of the study may have been injured or committed crimes while outside of the area. Additional research using a broader

Nanney et al / Am J Prev Med 2015;49(3):395–401400

geographic region would thus be needed to confirm the present findings. Also, given the gap in trauma registry data owing to Hurricane Katrina, this study is com- pressed into a relatively narrow time frame. The time between first and second injury in our study suggests that about 90% of patients who will return with repeat violent injuries from the 2006–2011 population are captured in these data, but that about 10% of the repeat victims of violence in this population are likely “incorrectly” classified as single-episode victims of violence because they have not yet returned with their second violent injury. Likewise, it is possible that an unknown number of patients experienced violent injuries prior to the beginning of the study data in 2006 such that they too are incorrectly classified as single-episode victims. Anal- ysis of a narrower cohort (2006–2008) for which 99% of repeat trauma victims are likely captured produced results substantively similar to the larger population, increasing the confidence that the present findings would persist if the database were extended in both the past and future.

Conclusions Though repeat violent injury is a public health priority for young urban men and increasingly a focus of intervention, research examining risk factors for repeat trauma remains sparse. The present results confirm previous scholarship suggesting that gun and drug crimes predict repeat violent injury. Analysis of timing of these crimes in relation to the initial injury indicates that drug crime before and gun possession after predict repeat trauma. Initiating gun or drug crime after injury also predicts later violent injury. Trauma center–based interventions targeting gun carrying and drug involvement may thus have prom- ise in reducing the violence that continues to plague many urban neighborhoods.

The authors would like to acknowledge the assistance of Erin Reuther, PhD for supervision of research assistants and for her comments regarding the manuscript. The authors also acknowledge the assistance of Samia Lalani, Christie Andolena, BS, April Hartman, BS, Ian Comnick, BA, Elena Pueraro, and Catherine Rochefort, BS, in collecting, entering, and managing data. Without their diligent work, this project would not have been possible.

The contents of this report do not represent the views of the Department of Veterans Affairs or the U.S. Government.

JTN, EJC, and JIC contributed to the conception and design of the study; the acquisition, analysis, and interpre- tation of data; the drafting and revision of the manuscript; and statistical analysis. JTN contributed to the supervision

of research assistants. EJC contributed to administrative support. And MM contributed to the interpretation of the data and the drafting and revision of the manuscript. JTN had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. No financial disclosures were reported by the authors of

this paper.

References 1. Anderson E. Code of the Street: Decency, Violence, and the Moral Life of

the Inner City. New York: Norton; 1999. 2. Rich JA, Grey CM. Pathways to recurrent trauma among young black

men: traumatic stress, substance abuse, and the “code of the street.” Am J Public Health. 2005;95(5):816–824. http://dx.doi.org/10. 2105/AJPH.2004.044560.

3. Hennekens CH, Drowos J, Levine RS. Mortality from homicide among young black men: a new American tragedy. Am J Med. 2013;126(4): 282–283. http://dx.doi.org/10.1016/j.amjmed.2012.07.007.

4. CDC. National Vital Statistics System—mortality tables. www.cdc.gov/ nchs/nvss/mortality_tables.htm.

5. Cooper C, Eslinger D, Nash D, Al Zawahri J, Stolley P. Repeat victims of violence. Arch Surg. 2000;135:837–843. http://dx.doi.org/10.1001/ archsurg.135.7.837.

6. Sims DW, Bivens BA, Obeid FN, Horst HM, Sorenson VJ, Fath JJ. Urban trauma: a chronic recurrent disease. J Trauma. 1989;29(7): 940–947. http://dx.doi.org/10.1097/00005373-198907000-00006.

7. Claassen CA, Larkin GL, Hodges G, Field C. Criminal correlates of injury-related emergency department recidivism. J Emerg Med. 2007; 32(2):141–147. http://dx.doi.org/10.1016/j.jemermed.2006.05.041.

8. Brooke BS, Efron DT, Chang DC, Haut ER, Cornwell III. Patterns and outcomes among penetrating trauma recidivists: it only gets worse. J Trauma. 2006;61(1):16–20. http://dx.doi.org/10.1097/01.ta.0000224143. 15498.bb.

9. Worrell SS, Koepsell TD, Sabath DR, Gentilello LM, Mock CN, Nathens AB. The risk of reinjury in relation to time since first injury: a retrospective population-based study. J Trauma. 2006;60(2): 379–384. http://dx.doi.org/10.1097/01.ta.0000203549.15373.7b.

10. Aboutanos MB, Jordan A, Cohen R, et al. Brief violence interventions with community base management services are effective for high-risk trauma patients. J Trauma. 2011;71(1):228–237. http://dx.doi.org/ 10.1097/TA.0b013e31821e0c86.

11. Becker MG, Hall JS, Ursic CM, et al. Caught in the Crossfire: the effects of a peer-based intervention program for violently injured youth. J Adolesc Health. 2004;34:177–183. http://dx.doi.org/10.1016/j.jadohealth .2003.04.001.

12. Cooper C, Eslinger DM, Stolley PD. Hospital-based violence intervention programs work. J Trauma. 2006;61(3):534–540. http://dx.doi.org/10.1097/ 01.ta.0000236576.81860.8c.

13. Cheng TL, Haynie D, Brenner R, Wright JL, Chung SE, Simons- Morton B. Effectiveness of a mentor-implemented, violence prevention intervention for assault-injured youths presenting to the emergency department: results of a randomized trial. Pediatrics. 2008;122(5): 938–946. http://dx.doi.org/10.1542/peds.2007-2096.

14. Dicker RA, Jaeger S, Knudson MM, et al. Where do we go from here? Interim analysis to forge ahead in violence prevention. J Trauma Acute Care Surg. 2009;67(6):1169–1175. http://dx.doi.org/10.1097/TA. 0b013e3181bdb78a.

15. Zun LS, Downey L, Rosen J. The effectiveness of an ED-based violence prevention program. Am J Emerg Med. 2006;24(1):8–13. http://dx.doi.org/ 10.1016/j.ajem.2005.05.009.

www.ajpmonline.org

Nanney et al / Am J Prev Med 2015;49(3):395–401 401

16. Cunningham R, Knox L, Fein J, et al. Before and after the trauma bay: the prevention of violent injury among youth. Ann Emerg Med. 2009;53(4):490–500. http://dx.doi.org/10.1016/j.annemergmed.2008. 11.014.

17. Douglas KS, Skeem JL. Violence risk assessment: getting specific about being dynamic. Psychol Public Pol L. 2005;11(3):347–383. http://dx.doi.org/ 10.1037/1076-8971.11.3.347.

18. Felson RB, Painter-Davis N. Another cost of being a young black male: race, weaponry, and lethal outcomes in assaults. Soc Sci Res. 2012;41(5): 1241–1253. http://dx.doi.org/10.1016/j.ssresearch.2012.04.006.

19. Weiner MD, Sussman S, Sun P, Dent C. Explaining the link between violence perpetration, victimization and drug use. Addict Behav. 2005;30(6):1261–1266. http://dx.doi.org/10.1016/j.addbeh.2004.12.007.

20. Spano R, Pridemore WA, Bolland J. Specifying the role of exposure to violence and violent behavior on initiation of gun carrying: a longi- tudinal test of three models of youth gun carrying. J Interpers Violence. 2012;27(1):158–176. http://dx.doi.org/10.1177/0886260511416471.

21. Spano R, Rivera C, Bolland J. The impact of timing of exposure to violence on violent behavior in a high poverty sample of inner city African American youth. J Youth Adolesc. 2006;35(5):681–692. http://dx.doi.org/10.1007/s10964-006-9080-3.

22. Wacquant L. Deadly symbiosis when ghetto and prison meet and mesh. Punishment Soc. 2001;3(1):95–133. http://dx.doi.org/10.1177/ 14624740122228276.

23. Garland EL, Pettus-Davis C, Howard MO. Self-medication among traumatized youth: structural equation modeling of pathways between trauma history, substance misuse, and psychological distress. J Behav Med. 2013;36(2):175–185. http://dx.doi.org/10.1007/s10865-012-9413-5.

24. Botvin GJ, Schinke SP, Epstein JA, Diaz T, Botvin EM. Effectiveness of culturally focused and generic skills training approaches to alcohol and drug abuse prevention among minority adolescents: two-year follow-up

September 2015

results. Psychol Addict Behav. 1995;9(3):183–194. http://dx.doi.org/10. 1037/0893-164X.9.3.183.

25. Zatzick D, Russo J, Lord SP, et al. Collaborative care intervention targeting violence risk behaviors, substance use, and posttraumatic stress and depressive symptoms in injured adolescents: a randomized clinical trial. JAMA Pediatr. 2014;168(6):532–539. http://dx.doi.org/ 10.1001/jamapediatrics.2013.4784.

26. Jackson CA, Henderson M, Frank JW, Haw SJ. An overview of prevention of multiple risk behaviour in adolescence and young adult- hood. J Public Health. 2012;34(1):i31–i40. http://dx.doi.org/10.1093/ pubmed/fdr113.

27. Cunningham RM, Bernstein SL, Walton M, et al. Alcohol, tobacco, and other drugs: future directions for screening and intervention in the emergency department. Acad Emerg Med. 2009;16:1078–1088. http://dx.doi.org/10.1111/j.1553-2712.2009.00552.x.

28. Branas CC, Richmond TS, Culhane DP, Ten Have TR, Wiebe DJ. Investigating the link between gun possession and gun assault. Am J Public Health. 2009;99(11):2034–2040. http://dx.doi.org/10.2105/ AJPH.2008.143099.

29. Dahlberg LL, Ikeda RM, Kresnow MJ. Guns in the home and risk of violent death in the home: findings from a national study. Am J Epidemiol. 2004;160(10):929–936. http://dx.doi.org/10.1093/aje/ kwh309.

30. Grassel KM, Wintemute GJ, Wright MA, Romero MP. Association between handgun purchase and mortality from firearm injury. Inj Prev. 2003;9(1):48–52. http://dx.doi.org/10.1136/ip.9.1.48.

31. Kellermann AL, Rivara FP, Rushforth NB, et al. Gun ownership as a risk factor for homicide in the home. N Engl J Med. 1993;329(15): 1084–1091. http://dx.doi.org/10.1056/NEJM199310073291506.

32. Caulkins JP, Sevigny E. How many people does the U.S. incarcerate for drug use, and who are they? Contemp Drug Probl. 2005;32(3):405–428.