Childhood Sexual Abuse and Longitudinal Substance Use Among Individuals with Severe Mental Illnesses

Purpose: Childhood sexual abuse (CSA) research from the past 30 years suggests that CSA survivors can face many psychosocial problems. While studies have found that CSA can contribute to both substance use and mental illness development, few studies specifically examine the longitudinal impact of CSA issues on substance use in clinical samples. This study aimed to identify substance use patterns among individuals with severe mental illnesses (SMIs) who have been sexually abused versus those who have not. It was hypothesized that individuals who experienced CSA would have higher levels of substance use over the course of the study than those who did not experience CSA, and that anger would have a moderating effect on the relationship between CSA and substance use. 

Methods: This study utilized data from the MacArthur Violence Risk Assessment Study. A total of 1136 individuals with SMIs were recruited from three psychiatric inpatient units in different cities. Participant data were collected in the hospital by research clinicians and every 10 weeks after discharge up to one year. CSA and substance use were assessed by participant self-report, and anger was assessed by Part A of the Novaco Anger Scale (NAS). Mixed–effects growth curve modeling was used to examine longitudinal patterns of substance use and related variability. Parameter estimates were obtained using generalized linear mixed effects growth curve models with binary outcomes utilizing penalized-quasi likelihood estimation and continuous outcomes utilizing restricted maximum likelihood estimation. Anger was examined as a moderating variable by entering baseline NAS scores into the final models.

Results: Of the 1136 participants, 1100 answered the sexual abuse screening question, with 46.5% (N=511) reporting at least one incident of sexual abuse. Chi-square tests revealed no significant differences between CSA and non-CSA groups in terms of follow-up interview completion or substance use at any follow-up interview time point (all p > .098). Results indicated that the proportion of individuals using substances over the one year period significantly decreased for alcohol (B = -.026, p < .001) and drug use (B = -.039, p < .001). Conditional GCM revealed that individuals that experienced CSA had a lower likelihood of baseline alcohol (B = -.481, p < .001) and drug (B = -.632, p < .001) usage; but had increased growth in drug use (B = .206, p = .014) and drug use frequency (B = .045, p= .018) over time. Moderator analysis found a significant interaction between sexual abuse, anger, and time, suggesting a moderating effect of anger on the relationship between CSA and the change in drug use frequency over time. 

Implications: These findings suggest that while CSA survivors used less substances at baseline, they also experienced less recovery over time. Anger also plays an important role in the relationship between CSA and substance use. In addition to supporting relapse prevention for CSA survivors, social workers should utilize screening tools that can identify problematic anger. Anger is a common response to CSA, but if addressed, social workers can help CSA survivors heal and avoid substance use.