Emerging adult outcomes of adolescent psychiatric and substance use disorders

Abstract

This study investigates the age 25 outcomes of late adolescent mental health and substance use disorders. A hierarchical cluster analysis of age 19 DSM-III-R mental health and substance diagnoses placed participants into one of 9 clusters: Anxious, Depressed, Antisocial, Drug Abuser, Problem Drinker, Anxious Drinker, Depressed Drug Abuser and Antisocial Drinker, and No Diagnosis. Diagnoses were generated from the University of Michigan Composite International Diagnostic Interview. Repeated measures multivariate analyses of variance revealed distinct trajectories of improvement and decline among the 9 clusters. Clusters with co-occurring substance and mental health disorders improved over adolescent levels, but continued to have higher levels of depression symptoms, poorer global functioning, and higher levels of substance use than the No Diagnosis cluster. Members of the The Problem Drinkers cluster, who tended to have alcohol use disorders only at age 19, did not differ from their peers with no diagnoses. Drug use disorders in adolescence, with or without a co-occurring mental health disorders, were associated with a poor prognosis in emerging adulthood. Clinical interventions should distinguish among these diverse clinical presentations.

Introduction

The substantial comorbidity between psychiatric and substance use disorders has prompted increasing research interest in co-occurring disorders (Cerdá et al., 2008, Gonzales and Insel, 2004). The National Comorbidity Survey (Kessler, 2004) reported that 42.7% of participants with a substance use disorder (SUD) also had a psychiatric disorder (other than a substance use disorder); and of those with at least one psychiatric disorder, 14.7% also had at least one SUD. Other studies have reported similarly high overlap between psychiatric and SUDs among adults (de Graaf et al., 2002, Kessler et al., 2005, Regier et al., 1998) and youth (Armstrong & Constello, 2000).

Cross-sectional studies have reported that co-occurring psychiatric disorders and SUDs are associated with more functional impairment than either type of disorder alone, including greater symptom severity, more frequent disability days, higher rates of substance relapse, and additional treatment challenges and greater unmet need for treatment compared with individual psychiatric disorders or SUDs (Beitchman et al., 2001, Bijl and Ravelli, 2000, Clark et al., 2008, Kessler et al., 2005, Urbanoski et al., 2008). These studies point to the importance of addressing co-occurring disorders, however, longitudinal research, particularly prospective studies that follow participants over a period of several years, is needed to trace the outcomes associated with SUDs in combination with psychiatric impairment (Cerdá et al., 2008, Crome and Bloor, 2005).

In this paper, we follow clusters of youth with various combinations of substance use, psychiatric, and co-occurring disorders from late adolescence (age 18/19) to early adulthood (age 25). This age range has recently been conceptualized as a distinct developmental period — emerging adulthood — because the multiple role transitions, and the social contexts and expectations that characterize this period, including an unprecedented degree of autonomy, differentiate it sharply from adolescence and from subsequent years of adulthood (e.g., Arnett, 2004, Tanner, 2006). Emerging adulthood is likewise an important developmental period with respect to co-occurring disorders because it is at the juncture of conflicting hypotheses about the outcomes of psychiatric and substance use disorders. In contrast to the negative sequelae associated with co-occurring disorders is the phenomenon of adolescents ‘maturing out’ of teenage disorders, particularly SUDs (Cunningham et al., 2005, Maggs and Schulenberg, 2004–2005, O'Malley, 2004–2005).

Outcomes of co-occurring disorders vary depending on the specific combination of SUD and psychiatric disorder (Kessler, 2004, Cerdá et al., 2008). A number of longitudinal studies with adolescent and young adult samples have used variable-centred approaches to track the course, order, and potential causal associations between SUDs and other psychiatric disorders (e.g., Goodwin et al., 2004, Pardini et al., 2007). These studies track the relation between symptoms or diagnoses and specific outcomes, but they do not examine the life outcomes of youth with particular combinations of symptoms.

In contrast, person-centred approaches, which are relatively rare in the comorbidity literature, can be used to identify sub-groups that reflect the complexity of diagnostic patterns and can be followed over time (Beitchman et al., 2001, Beitchman et al., 2005, Kessler et al., 2005, Windle and Davies, 1999). Advanced statistical techniques such as latent class analysis and cluster analysis have the advantage of identifying empirically observed patterns among constellations of diagnoses that can be validated against external criteria. In this paper, we use a cluster analytic approach to trace the life outcomes at age 25 of distinct sub-groups of adolescents identified based on their combinations of substance use and other psychiatric disorders.

Previously (Beitchman et al., 2001), we assessed mental health and psychiatric status in a community sample of 19-year-olds. Diagnoses were entered into a hierarchical clustering algorithm; 8 empirically-derived clinical cluster groups emerged: Depressed, Depressed Drug Abusers, Anxious, Anxious Drinkers, Antisocial, Antisocial Drinkers, Drug Abusers, and Problem Drinkers. A ninth cluster with no psychiatric or SUD diagnoses was added. The clusters were labelled based on the predominant diagnostic groups they represented; however, not every member of each cluster is precisely described by the labels.

Here we report the results of assessments of these participants 6 years later at age 25, focusing on symptom severity, substance use and life-outcome variables. We address the following research questions: 1) Do adolescents with various combinations of SUDs and psychiatric disorders have different life outcomes than their non-disordered peers when they reach adulthood? and 2) Does substance-psychiatric comorbidity influences the trajectory of improvement or decline between adolescence and emerging adulthood?, and 3) Which clinical clusters do well and which do poorly?