Substance Use Comorbidity Among Seriously Violent Offender Youth in the United States

Background and Purpose. The evidence is clear that substance use problems often accompany other antisocial behavior generally and violent behavior specifically; however, the specificity of the drugs/violence nexus is less clear. Although prior studies have delved into subsamples of violent offenders to examine their substance dependence status, the prevalence and possible heterogeneity of substance use among violent offenders is less well understood. It is in this vein that the current study seeks to fill this void using a rigorous quantitative analysis of a national-level data source. 

Methods. The current study findings are based on data from the National Survey on Drug Use and Health (NSDUH) between 2008 and 2012. Respondents were classified as “seriously violent offenders” if they reported having been arrested and booked for aggravated assault, forcible rape, murder, homicide, or nonnegligent manslaughter in the previous 12 months. We estimated latent class models on the basis of substance use disorders (step one), assigned cases to latent subgroups on the basis of posterior class membership probabilities (step two), and then examined the association between class membership and criminal justice system involvement using multinomial regression (step three). For the latent class analysis, we utilized measures of past 12-month alcohol, analgesic, cocaine/crack, hallucinogen, heroin, inhalant, marijuana, sedative, stimulant and tranquilizer use disorder (abuse or dependence) as indicator variables.  The NSDUH measures are based on the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) criteria (American Psychiatric Association, 2000). The current study restricted analyses to serious violent offenders between the ages of 12 and 25 (n = 1,115). 

Results. Three latent classes were discovered: Class 1 (61.2% of sample) who displayed limited substance use morbidity, Class 2 (28.2% of sample) who displayed comorbid alcohol and marijuana use disorders, and Class 3 (10.6% of sample) who displayed polydrug abuse/dependence and severe criminal careers. A sensitivity analysis in which we conducted LCAs with the adolescent (aged 12-17 years) and young adult (aged 18-25 years) subsamples suggests that—despite minor differences—the latent modeling is generally invariant across the two developmental time periods. Members of Class 2 (Comorbid Alcohol and Marijuana Use Disorders) were far more likely than members of Class 1 (Limited Substance Use Morbidity) to have also been arrested and booked  for drunkenness/liquor law violations (RR = 5.55, 95% CI = 2.12-14.73) or possession/manufacture/sale of drugs (RR = 6.95, 95% CI = 3.22-14.88). Members of Class 3 (Polydrug Abuse/Dependence) were between roughly 5 and 17 times more likely to have been arrested and booked for all criminal justice system variables (i.e., vehicle theft, larceny, burglary, robbery, arson, and alcohol/drug offenses) examined. 

Conclusions and Implications. The present study makes plain the fact that substantial heterogeneity exists among serious violent offenders in the United States. Notably, most offenders displayed limited usage of alcohol and marijuana, and virtually no usage of other serious drugs. However, a minority of serious violent offenders displayed substantial substance use comorbidity and faced markedly increased risk for involvement in an array of criminal behaviors.