Childhood Maltreatment and Smoking Behavior in Adolescence and Adulthood

Purpose: This study examines the relationship between childhood maltreatment and adolescent and adult smoking, and is guided by two questions: 1. Is there a direct effect of physical and sexual abuse on adolescent and adult smoking behavior accounting for gender and SES status in childhood?;  2. Does adolescent smoking mediate the effect of abuse experiences on later adult smoking?  Reviews of the existing literature show a link between childhood maltreatment and substance use. However, when substance use is measured, smoking is often omitted, overlooked, or averaged into general measures of “substance use” leaving our understanding of this relationship unclear and poorly understood.

Analysis used data from a 32-year prospective, longitudinal study to investigate the link between physical and sexual abuse to outcomes of smoking use and frequency during adolescence and adulthood. A hypothesized mediator of adult smoking was adolescent smoking behavior. Analyses also account for participant’s gender and SES in childhood, as covariates of abuse and stand-alone predictors of later smoking.

Methods and Results:  A censored-inflated modeling technique was used in a path analysis to assess the impact of child sexual and physical abuse on adolescent and adult smoking behavior. Censored-inflated modeling simultaneously conducts a two-part regression model with the first part predicting the probability that the individual would be a smoker or not while the second regression predicts the value of the frequency of smoking for those individuals who did smoke during the assessment period. Results found no significant relationship between childhood maltreatment and the probability of being a smoker or not in adolescence and adulthood. Physical abuse (beta=5.2, p<.01) and sexual abuse (beta=32.2, p<.01) significantly predicted the frequency of use among adolescent smokers, as did gender (beta=-23.7, p<.01) and SES (beta= -2.7, p<.01); Adolescent smoking predicted adult smoking (beta=.5, p<.001), and only SES in childhood predicted adult smoking (beta=-2.2, p<.05) over and above adolescent smoking frequency.  Finally, the indirect pathways from childhood sexual and physical abuse to adult smoking as mediated through adolescent smoking behavior were significant.

Conclusion and Implications: This two-part censored-inflated modeling technique provided important insight into smoking behavior across developmental periods in individuals with histories of sexual and physical abuse. Findings from this study combined indicate a need to better understand the mechanisms that link child maltreatment and adolescent and adult smoking. Furthermore, findings suggest that clinicians may consider the need for smoking cessation efforts in working with maltreated youth, particularly in adolescence.