Proximity Versus Early Emergence: Longitudinal Effects of Mental Health Problems On Adolescent Substance Use in a National Sample

Background: The identification of developmental windows at which key predictors of adolescent substance use are most influential is a crucial task for informing the design of appropriately targeted substance use prevention and intervention programs. The current study focused on the effects of two broad-based mental health indicators - conduct problems and depressive symptoms - and their interaction on changes in alcohol, cigarette, and marijuana use during the period from 8^th-12^th grade. Two competing hypotheses were tested: 1) that earlier-emerging mental health problems would be more associated with substance use, because earlier-emerging problems tend to be more severe and associated with more developmental difficulties, or 2) that mental health problems occurring closer in time to substance use would be more strongly associated with substance use than those occurring at a more distant time, because proximal predictors tend to be stronger than distal predictors due to both measurement and true effects.

Method: In a national, longitudinal sample from the Monitoring the Future study (N = 3,014), structural equation modeling was used to examine the effects of 8^th and 10^th grade conduct problems, depressive symptoms, and their interaction on changes in alcohol, cigarette, and marijuana use from 8^th through 12^th grades.

Results: Earlier-emerging (8^th grade) mental health problems were stronger predictors of increases in alcohol, cigarette, and marijuana use during this period than were later, more proximally measured (10^th grade) mental health problems. Mental health problems in 8^th grade were consistently positively associated with increases in substance use at 10^th and 12^th grades, whereas mental health problems at 10^th grade were not significantly associated with changes in substance use from 10^th to 12^th grade. 8^th grade conduct problems were most strongly associated with increases in alcohol and marijuana use, and 8^th grade depressive symptoms were most strongly associated with increases in cigarette use, with some evidence of interactive effects of conduct problems and depressive symptoms on marijuana and cigarette use.

Conclusions: The effects of early mental health problems on substance use were stronger than those of more proximally measured mental health problems, despite factors such as measurement artifact that generally lead to larger effects for more proximally measured variables. These results therefore make a strong case that interventions targeting mental health problems in order to reduce substance use should be implemented relatively early in adolescence, by 8^th grade or earlier.