Predictors of comorbid polysubstance use and mental health disorders in young adults-a latent class analysis

The co-occurrence of mental health and substance use disorders adds complexity to already-significant health burdens. This study tests whether mental health disorders group differently across substance use disorder types and compares associations of early factors with the development of differing comorbidities.

Consecutive antenatal clinic attendees were recruited to the longitudinal Mater-University of Queensland Study of Pregnancy (MUSP). Mother/offspring dyads were followed over 21 years.

Mater-Misericordiae Public Hospital, Brisbane, Australia.

MUSP offspring with maternal baseline information (n = 7223), offspring behaviour data at 14 (n = 4815) and psychiatric diagnoses at 21 (n = 2575).

The Composite International Diagnostic Interview yielded life-time diagnoses of mental health (MH) and substance use (SU) disorders for offspring, then latent class modelling predicted membership of polydisorder groups. We fitted the resulting estimates in multinomial logistic regression models, adjusting for maternal smoking, drinking and mental health, adolescent drinking, smoking and behaviour and mother-child closeness.

Fit indices [Bayesian information criterion (BIC) = 12 415; Akaike information criterion (AIC) = 12 234] from LCA supported a four-class solution: low disorder (73.6%), MH/low SU disorder (10.6%), alcohol/cannabis/low MH disorder (12.2%) and poly SU/moderate MH disorder (3.5%). Adolescent drinking predicted poly SU/MH disorders [odds ratio (OR) = 3.34, 95% confidence interval (CI) = 1.42-7.84], while externalizing predicted membership of both SU disorder groups (ORalcohol/cannabis  = 2.04, 95% CI = 1.11-3.75; ORpolysubstance  = 2.65, 95% CI = 1.1-6.08). Maternal smoking during pregnancy predicted MH (OR = 1.53, 95% CI = 1.06-2.23) and alcohol/cannabis-use disorders (OR = 1.73; 95% CI = 1.22-2.45). Low maternal warmth predicted mental health disorders only (OR = 2.21, 95% CI = 1.32-3.71).

Mental health disorders are more likely in young adults with polysubstance use disorders than those with alcohol/cannabis use disorders. Predictors of comorbid mental health/polysubstance use disorders differ from those for alcohol/cannabis use disorders, and are detectable during adolescence.