[Skip to Navigation]
Sign In
Sep 2012

Effects of Childhood Maltreatment on Violent Injuries and Premature Death During Young Adulthood Among Urban High-Risk Men

Author Affiliations

Author Affiliations: Center of Alcohol Studies and Department of Sociology, Rutgers University, Piscataway, New Jersey.

Arch Pediatr Adolesc Med. 2012;166(9):814-820. doi:10.1001/archpediatrics.2012.244

Objectives To assess childhood maltreatment as a risk factor for violent injuries and premature death in young adulthood and whether these associations are mediated by adolescent heavy drinking, hard drug use, hard drug selling, and violent offending.

Design Prospective longitudinal study of boys followed from childhood into young adulthood.

Setting Pittsburgh, Pennsylvania.

Participants A total of 1009 men from the Pittsburgh Youth Study.

Main Exposure Childhood maltreatment.

Main Outcome Measures Premature deaths between ages 18 and 38 years from the Social Security Death Index and self-reports of violent injuries inflicted by gunshot or knife between ages 18 and 28 years.

Results Young men who experienced childhood maltreatment, compared with their counterparts who did not experience it, had a greater risk of violent injuries (relative risk = 1.61; 95% CI, 1.10-2.35) and death (hazard ratio = 2.85; 95% CI, 1.37-5.93) during young adulthood. Adolescent violent offending and hard drug selling explained the association between childhood maltreatment and violent injuries, and violent offending partially accounted for the association between childhood maltreatment and premature death. Although adolescent violent offending predicted both outcomes, maltreated boys still had an increased risk of premature death (hazard ratio = 2.54; 95% CI, 1.21-5.34) after accounting for their adolescent violence.

Conclusions Childhood maltreatment significantly predicts premature death and violent injuries during young adulthood. These associations are partially explained by adolescent involvement in violence and drug dealing. Targeted interventions for maltreated boys to reduce their involvement in adolescent deviant behaviors may help decrease their risks for later serious injuries and premature death.

Childhood maltreatment is a significant public health concern in the United States. Data from the National Child Abuse and Neglect Data System indicate that 9.3 children per 1000 were victims of childhood maltreatment in 2009,1 and data from the 2005-2006 National Incidence Study indicate rates of 39.5 children per 1000.2

Individuals maltreated in childhood, compared with nonmaltreated individuals, are at higher risk for morbidity3-8 and mortality9-12 during the life span. Nevertheless, research focusing on the effects of childhood maltreatment on injuries and mortality has been largely limited to studies in early developmental periods5,10 or later adulthood.3,11 Most studies examining later outcomes have relied on retrospective self-reports of childhood maltreatment.3,11 To our knowledge, no studies have clearly identified the potential mechanisms that might account for premature death and nonfatal injuries beyond early childhood. This study attempts to fill these gaps.

Findings from longitudinal studies of maltreated youths in adolescence and adulthood have been inconsistent with regard to whether these individuals are at higher risk for premature death. One study followed youths through age 18 years and found that abused children had an elevated risk of death compared with nonabused children.10 Similarly, another found that adverse childhood experiences, including childhood abuse, were associated with an elevated risk of premature death in midlife.11 In contrast, another study did not find that mortality rates differed between maltreated individuals and matched control subjects in young adulthood after controlling for several demographic and family characteristics.12 These inconsistencies may be due to differences in measures, observation periods, age cohorts, geographic areas, and/or analytic methods.

Because death rates of young adults are more than twice those of adolescents,13 the mortality gap between maltreated and nonmaltreated individuals might further increase in young adulthood. Accordingly, it is imperative to investigate the association between childhood maltreatment and premature death through young adulthood. Moreover, death might not be the only severe ramification of childhood maltreatment in adolescence and young adulthood; adolescents are likely to engage in risk behaviors,14 which might lead to injuries that are not fatal but are disabling. To date, no studies to our knowledge have concurrently investigated the association of childhood maltreatment with both violent injuries and premature death.

Attention should be paid to the mechanisms by which childhood maltreatment increases the risk of death and violent injuries. A problem behavior perspective15 has the potential to explain this association. For example, involvement in deviant behaviors may increase the risk of nonfatal and fatal consequences. Specifically, adolescent substance use,16,17 drug selling,18 and violence19,20 are well-known precursors of the risks for mortality and violent injuries among adolescents and young adults. Considering that these behaviors are highly prevalent among individuals with childhood maltreatment histories,21-29 these behaviors may be direct mechanisms linking childhood maltreatment to premature death and violent injuries in young adulthood.

This study focuses on adolescent risk behaviors because of the importance of early risk behaviors on developmental trajectories of deviant behaviors24,29 and the need to maintain the temporal order between mediators and outcomes. The study contributes to the literature by examining whether childhood maltreatment increases the risk of violent injuries and/or premature death during young adulthood through adolescent heavy drinking, hard drug use, hard drug selling, and violent offending after controlling for other childhood adversities and sociodemographic characteristics.


Study design and study population

Subjects were boys from the Pittsburgh Youth Study, a prospective longitudinal study of youth development.30 The Pittsburgh Youth Study began in 1987 and initially consisted of 3 cohorts (N = 1517) who had been randomly selected from the first, fourth, and seventh grades of public schools in Pittsburgh, Pennsylvania (referred to as the youngest, middle, and oldest cohorts, respectively). After an initial screening, the top 30% who screened at highest risk for antisocial behavior were included in the sample for follow-up, together with 30% randomly selected from the remainder. More than half of the sample were black, and the rest were almost all white. More than 90% lived with their biological mother at the time of their screening. Although about one-third received public assistance, the boys were fairly evenly distributed across socioeconomic status levels.30 The current study was limited to the youngest (n = 503; age range at baseline, 5-9 years) and oldest (n = 506; age range at baseline, 11-16 years) cohorts because the middle cohort was surveyed only from 1987 to 1991. Data were collected at approximate ages of 7 to 20, 25, and 28 years for the youngest cohort and approximate ages of 13 to 25 and 34 years for the oldest cohort. The study was approved by the University of Pittsburgh Institutional Review Board.

Definitions of variables

Boys were identified as maltreated if any type of maltreatment (ie, physical or sexual abuse, physical neglect, emotional maltreatment, or moral, legal, or educational maltreatment) was substantiated and they were first referred to Children and Youth Services by age 13 years. Age 13 years was chosen as the cutoff because a history of childhood maltreatment was fully documented for both cohorts through age 13 years. Those who were first referred to Children and Youth Services between ages 14 and 17 years (20 of 298 referrals [6.7%]) were excluded to ensure temporal order between childhood maltreatment and the mediators (ages 14-17 years). In addition, those who were referred to Children and Youth Services but whose cases were unsubstantiated (76 of 298 referrals [25.5%]) were excluded from the study because prior findings have been inconsistent about whether they more closely resemble substantiated cases31 or nonmaltreated cases32 in terms of their risk behaviors. Also, referral to Children and Youth Services could be biased by race and poverty.33-35 The final sample consisted of 913 men (202 maltreated and 711 nonmaltreated).

Violent injuries were defined as any gunshot or knife wound from ages 18 through 28 years (mean age at their last report, 25 years). Death data came from mortality records obtained from a search of the Social Security Death Index through July 27, 2010, when the youngest cohort was aged 27 to 32 years (mean age, 29 years) and the oldest cohort was aged 34 to 38 years (mean age, 35 years). Premature death was any death that occurred between ages 18 and 38 years.

The mediators were measured at ages 14 to 17 years, an intermediate period between childhood maltreatment and the outcomes. Because the frequency of engaging in the risk behaviors was low, all the mediators were coded as dichotomous variables. Heavy drinking was defined as typically having 6 or more drinks (beer, wine, or liquor) per occasion. Hard drug use was defined as using any illegal drug except marijuana (hallucinogens, cocaine, crack, heroin, phencyclidine, tranquilizers, barbiturates, codeine, amphetamines, and other prescription medications for nonmedical reasons). Hard drug selling was defined as selling any hard drugs (eg, heroin, cocaine, or LSD). Violent offending was assessed using official conviction records (county, state, and/or federal) or self-report or caretaker report of the youth committing robbery, attacking to hurt or kill, homicide, manslaughter, rape or forced sex, simple assault, or gang fighting. Violence was defined as any violent offense between ages 14 and 17 years.

To disentangle the effect of childhood maltreatment from other adverse childhood experiences, 4 domains of household dysfunction were measured. Approximately at ages 8 and 15 years for the youngest and oldest cohorts, respectively, the primary caretaker (usually the mother) was asked about female and male caretaker problems. Caretaker's alcohol and drug problems were defined as any caretaker having sought help for a substance use problem. Caretaker's mental illness was defined as a caretaker ever having sought help for any mental health problem (depression, anxiety, mental retardation, or schizophrenia). Caretaker's incarceration was defined as any caretaker ever having been incarcerated. Not living with 2 biological parents was defined as the primary caretaker being ever widowed, never married, or ever divorced or separated or if there was no female caretaker up to age 17 years for the participant. The household dysfunction score was the sum of these 4 domains, ranging from 0 to 4, and is similar to a measure used in previous research.11 Because cohort differences in patterns of alcohol use and violence have been found in this sample,36,37 cohort (oldest vs youngest) was controlled in the analyses. Parental socioeconomic status, based on the Hollingshead Four Factor Index of Social Status38 at the first follow-up assessment, was also controlled. Race/ethnicity was controlled with black individuals compared with individuals of white, other, or mixed races/ethnicities. Age at the initial screening was included to account for within-cohort differences in age and for age differences in outcomes.

Statistical analysis

All analyses were conducted with Stata version 11.0 statistical software (StataCorp LP). We used χ2 analyses and t tests to compare characteristics between maltreated and nonmaltreated young men. Bivariate correlations are reported for all the variables included in the multivariate analyses. A generalized linear model with a log link and binomial error distribution was used to examine the effects of childhood maltreatment experiences on violent injuries through the hypothesized mediators. The Kaplan-Meier method and the log-rank test were conducted to assess the relative risk (RR) of death between the maltreated and nonmaltreated groups.39 The Cox proportional hazards model for survival analysis was used to estimate the hazard of death for nonmaltreated and maltreated young men through the proposed mediators.39 Two respondents who died prior to age 17 years were eliminated owing to a temporal order issue. We report RRs for serious injuries, hazard ratios for premature death, and 95% CIs.

For both outcomes, individual mediators were first added separately, and all significant mediators (P < .05) from the single-mediator models were tested simultaneously in a multiple-mediator model.40,41 Missing data ranged from 4.3% to 6.5%. All missing data except the dependent variables were imputed under the missing-at-random assumption.42


Characteristics of maltreated young men

Table 1 presents results from bivariate analyses comparing characteristics of maltreated (22.1%) and nonmaltreated (77.9%) young men. Maltreated young men compared with nonmaltreated young men were slightly younger at screening (mean, 9.2 vs 9.8 years, respectively; P = .02), more likely to be black (48.5% vs 37.6%, respectively; P = .006), and more likely to be in the youngest cohort (51.9% vs 40.6%, respectively; P = .005). Maltreated young men compared with nonmaltreated men were also significantly more likely to grow up in a household with a lower socioeconomic status (mean Hollingshead Four Factor Index of Social Status score, 29.4 vs 37.6, respectively; P < .001), with a caretaker who was in an unstable marriage (87.6% vs 65.0%, respectively; P < .001), and with at least 1 caretaker who was more likely to have an alcohol or drug problem (18.3% vs 8.9%, respectively; P < .001), a mental illness (28.7% vs 17.9%, respectively; P = .001), and/or a higher rate of incarceration (10.9% vs 2.1%, respectively; P < .001). Maltreated young men compared with nonmaltreated young men reported significantly higher rates of hard drug selling (29.7% vs 17.6%, respectively; P < .001) and violent offending (51.0% vs 30.7%, respectively; P < .001) during adolescence. There was no significant difference in heavy drinking or hard drug use between the 2 groups. During young adulthood, maltreated young men compared with nonmaltreated young men had a greater risk of violent injuries (18.8% vs 11.1%, respectively; P = .005) and mortality (7.4% vs 2.5%, respectively; P = .001).

Table 1. Characteristics of Participants by Childhood Maltreatment
Table 1. Characteristics of Participants by Childhood Maltreatment
Table 1. Characteristics of Participants by Childhood Maltreatment

Table 2 shows the correlations among all the variables included in the multivariate analyses. All the problem behaviors were significantly related to each other, but multicollinearity was not a problem in the multivariate model.

Table 2. Bivariate Correlations for All Variables in Multivariate Analysisa
Table 2. Bivariate Correlations for All Variables in Multivariate Analysisa
Table 2. Bivariate Correlations for All Variables in Multivariate Analysisa

Direct and mediated effects of childhood maltreatment on violent injuries

Table 3 shows that, after adjusting for covariates, maltreated young men compared with nonmaltreated young men were more likely to be injured (RR = 1.61; 95% CI, 1.10-2.35). The effect of childhood maltreatment was slightly reduced by individually adding heavy drinking, hard drug selling, and violent offending to the model (panels 2, 4, and 5). After adding these 3 significant mediators to the same model (panel 6), the effect of childhood maltreatment was no longer significant (RR = 1.32; 95% CI, 0.93-1.89). Hard drug selling (RR = 1.89; 95% CI, 1.30-2.75) and violent offending (RR = 2.27; 95% CI, 1.51-3.40) significantly predicted violent injuries.

Table 3. Direct and Mediated Effects of Childhood Maltreatment on Violent Injuriesa
Table 3. Direct and Mediated Effects of Childhood Maltreatment on Violent Injuriesa
Table 3. Direct and Mediated Effects of Childhood Maltreatment on Violent Injuriesa

Direct and mediated effects of childhood maltreatment on premature death

There were a total of 35 deaths during the 913 person-year observation period, yielding a mortality rate of 3834 deaths per 100 000 person-years (7426 for maltreated individuals and 2813 for nonmaltreated individuals). The major cause of death was homicide (n = 26), followed by accident (n = 3), unknown (n = 3), natural (n = 2), and suicide (n = 1). Survival curves between nonmaltreated and maltreated young men based on the Kaplan-Meier estimator are shown in the Figure. Compared with nonmaltreated young men, maltreated young men had a significantly diminished survival time through young adulthood (log-rank statistic = 9.72; P = .002).

Figure. Kaplan-Meier estimate of survival probabilities among maltreated and nonmaltreated young men. The survival curves between the 2 groups differed significantly (log-rank statistic = 9.72; P = .002).

Figure. Kaplan-Meier estimate of survival probabilities among maltreated and nonmaltreated young men. The survival curves between the 2 groups differed significantly (log-rank statistic = 9.72; P = .002).

Table 4 shows that, after controlling for the covariates, death rates were higher for maltreated young men than nonmaltreated young men (hazard ratio = 2.85; 95% CI, 1.37-5.93). The risk of childhood maltreatment on premature death was reduced by adolescent violent offending, although the effect of childhood maltreatment on premature death remained significant (hazard ratio = 2.54; 95% CI, 1.21-5.34), indicating only a partial mediation effect. Adolescent heavy drinking, hard drug use, and hard drug selling were not significant mediators.

Table 4. Direct and Mediated Effects of Childhood Maltreatment on Premature Deatha
Table 4. Direct and Mediated Effects of Childhood Maltreatment on Premature Deatha
Table 4. Direct and Mediated Effects of Childhood Maltreatment on Premature Deatha


This study sought to extend our understanding of the association between childhood maltreatment and premature death and violent injuries. The results indicated that young men who experienced childhood maltreatment and were referred to Children and Youth Services by age 13 years, compared with their counterparts who did not experience it, had a greater risk of violent injuries and death during young adulthood as well as greater risks of engaging in risk behaviors during adolescence. Hard drug selling and violent offending during adolescence (ages 14-17 years) were underlying mechanisms linking childhood maltreatment to violent injuries. Adolescent violence, in particular, remained a strong and significant predictor for both violent injuries and death after accounting for sociodemographic and familial factors as well as childhood maltreatment.

The findings are consistent with a problem behavior perspective15 and prior research, which has shown that childhood maltreatment plays a major role in the development of deviant behaviors21-29 and that these behaviors appear to increase the risk of violent injuries and mortality.16-20 However, to our knowledge, previous studies have not formally tested this potential mediation effect.

Researchers have suggested that childhood maltreatment might be only 1 type of childhood adversity that contributes to the development of negative health outcomes. That is, the adverse association between childhood maltreatment and health outcomes might be accounted for by the potentially confounding effects of other childhood adversities such as parental substance use and poverty43,44 that may ultimately increase the likelihood of a false-positive relationship. Except 1 study,12 prior studies focusing on childhood maltreatment and mortality could not fully disentangle confounding associations. This study intended to clarify this issue. After adjusting for demographic characteristics and household dysfunction scores, the risk of death was almost twice as high and the risk of violent injuries was approximately 60% higher for maltreated young men than nonmaltreated young men. The fact that these findings are different from those of White and Widom,12 who found no significant effects of childhood maltreatment on premature death, may be due to several differences between the 2 studies. There may be a period effect because the youths in this study grew up in an inner city in the 1980s and 1990s, when American society experienced a dramatic increase in rates of violence and homicide during the crack-cocaine epidemic.18,45 The fact that 74.3% of the deaths were homicide might in part explain the period effect. The sample in the study by White and Widom was born during an earlier period (before 1971) and also included both men and women. Further, racial and geographic differences between the 2 samples might account for the differences in findings.

We focused on deviant behaviors only during adolescence as mediators owing to the issue of temporal order. Yet, persistent risk behaviors throughout adulthood might lead to fatal or near-fatal consequences. Future research should examine adulthood risk behaviors as potential mediators. In addition, there might be other mechanisms linking childhood maltreatment to violent injuries and premature death, such as social support,46,47 community violence,48 peer delinquency,49 accessibility to emergency care, or attitudes toward seeking help from “the system” (ie, police),50 which should be tested in future studies.

Several features of this study warrant attention when interpreting the results. First, the sample represents young men who grew up in an inner city, and approximately half were prone to antisocial behavior at screening. Thus, the results may not generalize to lower-risk men, women, or youths who live in different environments. Moreover, any individuals who experienced childhood maltreatment after age 13 years were excluded owing to a lack of comparable information for the youngest cohort and also to establish the hypothesized temporal order between the predictor and mediators. Thornberry et al21 found that maltreatment during adolescence had significant effects on substance use, violence, and suicidal thoughts. Therefore, the effects of childhood maltreatment on the outcomes in this study may have been underestimated. In addition, although the measures improved on those of previous studies because they included both official and reported data from diverse sources, measures of heavy drinking, drug use, drug selling, and violent injuries were based only on youth self-reports. Nonetheless, reliability and validity of self-report data have been established in many studies.51 Owing to a relatively small sample, this study did not examine whether different types, severity, and age of childhood maltreatment as well as receiving foster care might affect the risk of premature death and violent injuries among maltreated individuals. Future research with larger samples should examine these potential moderators.

Despite these limitations, this study contributes compelling evidence supporting childhood maltreatment as a key risk factor for both violent injuries and death during young adulthood in a contemporary sample. In addition, violent offending during adolescence appears to be an important mediator. Violent offending and being a victim of violence, however, are intertwined and mutually influence each other.20 Adolescent violent offending is strongly correlated with violent crime in adulthood, especially for maltreated individuals,24 which in turn increases the risk of violent injuries and mortality. Accordingly, it is imperative to provide targeted prevention for maltreated youths to help them deal with their maltreatment experience and reduce their involvement in adolescent risk behaviors and later violent offending. In addition, social policies should focus on alleviating fundamental structural problems (eg, poverty, single parenting, community violence, difficulties in accessing health care) that contribute to childhood maltreatment as well as to the risk of death and injuries for maltreated individuals.

Back to top
Article Information

Correspondence: Helene R. White, PhD, Center of Alcohol Studies, Rutgers University, 607 Allison Rd, Piscataway, NJ 08854-8001 (hewhite@rci.rutgers.edu).

Accepted for Publication: February 23, 2012.

Published Online: May 7, 2012. doi:10.1001/archpediatrics.2012.244

Author Contributions: Dr White had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Lee and White. Acquisition of data: White. Analysis and interpretation of data: Lee and White. Drafting of the manuscript: Lee and White. Critical revision of the manuscript for important intellectual content: Lee and White. Statistical analysis: Lee. Obtained funding: White. Study supervision: White.

Financial Disclosure: None reported.

Funding/Support: This work was supported in part by grants ARRA R01 AA 016798 and R01 AA 019511 from the National Institute on Alcohol Abuse and Alcoholism, R01 DA411018 from the National Institute on Drug Abuse, P30 MH079920 and R01 MH73941 from the National Institute of Mental Health, and 96-MU-FX-0012 and OJJDP 2005-JK-FX-0001 from the Office of Juvenile Justice and Delinquency Prevention and by the Department of Health of the Commonwealth of Pennsylvania and the Centers for Disease Control and Prevention.

Disclaimer: The points of view in this article are those of the authors and do not necessarily represent the official position or policies of the US Department of Justice.

Additional Contributions: Magda Stouthamer-Loeber, PhD, Rolf Loeber, PhD, Eun-Young Mun, PhD, Dustin Pardini, PhD, Kristen Springer, PhD, and Allan Horwitz, PhD, provided comments and suggestions, Kristen McCormick, BA, helped with the data set, and Courtney Cronley, PhD, suggested references.

Children's Bureau, Administration on Children, Youth and Families, Administration for Children and Families, US Department of Health and Human Services.  Child maltreatment 2009. http://www.acf.hhs.gov/programs/cb/pubs/cm09/cm09.pdf. Accessed November 11, 2011
Sedlak AJ, Mettenburg J, Basena M,  et al.  Fourth National Incidence Study of Child Abuse and Neglect (NIS-4): Analysis Report. Washington, DC: US Dept of Health & Human Services, Administration for Children and Families; 2010
Felitti VJ, Anda RF, Nordenberg D,  et al.  Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: the Adverse Childhood Experiences (ACE) Study.  Am J Prev Med. 1998;14(4):245-2589635069PubMedGoogle ScholarCrossref
Goodwin RD, Stein MB. Association between childhood trauma and physical disorders among adults in the United States.  Psychol Med. 2004;34(3):509-52015259836PubMedGoogle Scholar
Billmire ME, Myers PA. Serious head injury in infants: accident or abuse?  Pediatrics. 1985;75(2):340-3423969337PubMedGoogle Scholar
Horwitz AV, Widom CS, McLaughlin J, White HR. The impact of childhood abuse and neglect on adult mental health: a prospective study.  J Health Soc Behav. 2001;42(2):184-20111467252PubMedGoogle Scholar
Brown J, Cohen P, Johnson JG, Smailes EM. Childhood abuse and neglect: specificity of effects on adolescent and young adult depression and suicidality.  J Am Acad Child Adolesc Psychiatry. 1999;38(12):1490-149610596248PubMedGoogle Scholar
Springer KW. Childhood physical abuse and midlife physical health: testing a multi-pathway life course model.  Soc Sci Med. 2009;69(1):138-14619446943PubMedGoogle Scholar
Herman-Giddens ME, Brown G, Verbiest S,  et al.  Underascertainment of child abuse mortality in the United States.  JAMA. 1999;282(5):463-46710442662PubMedGoogle Scholar
Sabotta EE, Davis RL. Fatality after report to a child abuse registry in Washington State, 1973-1986.  Child Abuse Negl. 1992;16(5):627-6351393723PubMedGoogle Scholar
Brown DW, Anda RF, Tiemeier H,  et al.  Adverse childhood experiences and the risk of premature mortality.  Am J Prev Med. 2009;37(5):389-39619840693PubMedGoogle Scholar
White HR, Widom CS. Does childhood victimization increase the risk of early death? a 25-year prospective study.  Child Abuse Negl. 2003;27(7):841-85314627082PubMedGoogle Scholar
National Center for Injury Prevention and Control.  WISQARS: leading causes of death, fatal and non-fatal injury reports. http://www.cdc.gov/injury/wisqars/fatal.html. Accessed November 14, 2011
Mulye TP, Park MJ, Nelson CD, Adams SH, Irwin CE Jr, Brindis CD. Trends in adolescent and young adult health in the United States.  J Adolesc Health. 2009;45(1):8-2419541245PubMedGoogle Scholar
Jessor R, Donovan JE, Costa FM. Beyond Adolescence: Problem Behavior and Young Adult Development. Cambridge, England: Cambridge University Press; 1994
Clark DB, Martin CS, Cornelius JR. Adolescent-onset substance use disorders predict young adult mortality.  J Adolesc Health. 2008;42(6):637-63918486875PubMedGoogle Scholar
Hingson RW, Heeren T, Zakocs RC, Kopstein A, Wechsler H. Magnitude of alcohol-related mortality and morbidity among US college students ages 18-24.  J Stud Alcohol. 2002;63(2):136-14412033690PubMedGoogle Scholar
Goldstein PJ, Brownstein HH, Ryan PJ, Bellucci PA. Crack and homicide in New York City, 1988: a conceptually based event analysis.  Contemp Drug Probl. 1989;16(4):651-687Google Scholar
Loeber R, DeLamatre M, Tita G, Cohen J, Stouthamer-Loeber M, Farrington DP. Gun injury and mortality: the delinquent backgrounds of juvenile victims.  Violence Vict. 1999;14(4):339-35210751043PubMedGoogle Scholar
Loeber R, Kalb L, Huizinga D. Juvenile delinquency and serious injury victimization.  Juv Justice Bull. 2001;8:1-8Google Scholar
Thornberry TP, Henry KL, Ireland TO, Smith CA. The causal impact of childhood-limited maltreatment and adolescent maltreatment on early adult adjustment.  J Adolesc Health. 2010;46(4):359-36520307825PubMedGoogle Scholar
Ireland TO, Smith CA, Thornberry TP. Developmental issues in the impact of child maltreatment on later delinquency and drug use.  Criminology. 2002;40:359-399Google Scholar
Molnar BE, Buka SL, Kessler RC. Child sexual abuse and subsequent psychopathology: results from the National Comorbidity Survey.  Am J Public Health. 2001;91(5):753-76011344883PubMedGoogle Scholar
Widom CS, Maxfield MG. An Update on the “Cycle of Violence.” Washington, DC: National Institute of Justice; 2001
Widom CS, Marmorstein NR, White HR. Childhood victimization and illicit drug use in middle adulthood.  Psychol Addict Behav. 2006;20(4):394-40317176174PubMedGoogle Scholar
Widom CS, White HR, Czaja SJ, Marmorstein NR. Long-term effects of child abuse and neglect on alcohol use and excessive drinking in middle adulthood.  J Stud Alcohol Drugs. 2007;68(3):317-32617446970PubMedGoogle Scholar
Black MM, Ricardo IB. Drug use, drug trafficking, and weapon carrying among low-income, African-American, early adolescent boys.  Pediatrics. 1994;93(6, pt 2):1065-10728197009PubMedGoogle Scholar
Widom CS. The cycle of violence.  Science. 1989;244(4901):160-1662704995PubMedGoogle Scholar
Stouthamer-Loeber M, Loeber R, Homish DL, Wei E. Maltreatment of boys and the development of disruptive and delinquent behavior.  Dev Psychopathol. 2001;13(4):941-95511771915PubMedGoogle Scholar
Loeber R, Farrington DP, Stouthamer-Loeber M, White HR. Violence and Serious Theft: Development and Prediction From Childhood to Adulthood. New York, NY: Routledge; 2008
Cross TP, Casanueva C. Caseworker judgments and substantiation.  Child Maltreat. 2009;14(1):38-5218794302PubMedGoogle Scholar
Lee C, Cronley C, White HR,  et al.  Racial differences in the consequences of childhood maltreatment for adolescent and young adult depression, heavy drinking, and violence.  J Adolesc HealthIn pressGoogle Scholar
Lane WG, Rubin DM, Monteith R, Christian CW. Racial differences in the evaluation of pediatric fractures for physical abuse.  JAMA. 2002;288(13):1603-160912350191PubMedGoogle Scholar
Laskey AL, Stump TE, Perkins SM, Zimet GD, Sherman SJ, Downs SM. Influence of race and socioeconomic status on the diagnosis of child abuse: a randomized study [published online January 3, 2012].  J PediatrGoogle Scholar
Fluke JD, Yuan YYT, Hedderson J, Curtis PA. Disproportionate representation of race and ethnicity in child maltreatment: investigation and victimization.  Child Youth Serv Rev. 2003;25(5-6):359-373Google Scholar
Lee C, Mun EY, White HR, Simon P. Substance use trajectories of black and white young men from adolescence to emerging adulthood: a two-part growth curve analysis.  J Ethn Subst Abuse. 2010;9(4):301-31921161811PubMedGoogle Scholar
Fabio A, Loeber R, Balasubramani GK, Roth J, Fu W, Farrington DP. Why some generations are more violent than others: assessment of age, period, and cohort effects.  Am J Epidemiol. 2006;164(2):151-16016775045PubMedGoogle Scholar
Hollingshead AB. Four Factor Index of Social Status. New Haven, CT: Dept of Sociology, Yale University; 1975
Blossfeld HP, Golsch K, Rohwer G. Event History Analysis With Stata. Mahwah, NJ: Lawrence Erlbaum Associates; 2007
Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations.  J Pers Soc Psychol. 1986;51(6):1173-11823806354PubMedGoogle Scholar
MacKinnon DP. Introduction to Statistical Mediation Analysis. Mahwah, NJ: Lawrence Erlbaum Associates; 2008
Allison PD. Missing Data. Thousand Oaks, CA: Sage Publications; 2001
Huang S, Trapido E, Fleming L,  et al.  The long-term effects of childhood maltreatment experiences on subsequent illicit drug use and drug-related problems in young adulthood.  Addict Behav. 2011;36(1-2):95-10220947260PubMedGoogle Scholar
Dunn MG, Tarter RE, Mezzich AC, Vanyukov M, Kirisci L, Kirillova G. Origins and consequences of child neglect in substance abuse families.  Clin Psychol Rev. 2002;22(7):1063-109012238246PubMedGoogle Scholar
Blumstein A, Rivara FP, Rosenfeld R. The rise and decline of homicide—and why.  Annu Rev Public Health. 2000;21:505-54110884963PubMedGoogle Scholar
Lynch M, Cicchetti D. An ecological-transactional analysis of children and contexts: the longitudinal interplay among child maltreatment, community violence, and children's symptomatology.  Dev Psychopathol. 1998;10(2):235-2579635223PubMedGoogle Scholar
Borowsky IW, Ireland M. Predictors of future fight-related injury among adolescents.  Pediatrics. 2004;113(3, pt 1):530-53614993545PubMedGoogle Scholar
Teicher MH, Andersen SL, Polcari A, Anderson CM, Navalta CP, Kim DM. The neurobiological consequences of early stress and childhood maltreatment.  Neurosci Biobehav Rev. 2003;27(1-2):33-4412732221PubMedGoogle Scholar
Seagull EA. Social support and child maltreatment: a review of the evidence.  Child Abuse Negl. 1987;11(1):41-523548917PubMedGoogle Scholar
Rich JA, Grey CM. Pathways to recurrent trauma among young Black men: traumatic stress, substance use, and the “code of the street.”  Am J Public Health. 2005;95(5):816-82415855457PubMedGoogle Scholar
Graham JW, Roberts MM, Tatterson JW, Johnston SE. Data quality in evaluation of an alcohol-related harm prevention program.  Eval Rev. 2002;26(2):147-18911949537PubMedGoogle Scholar