Evaluating the modified common liability hypothesis of psychiatric comorbidity

We sought to evaluate whether the timing of co-occurrence of psychiatric disorder groups provides evidence for one disorder group predisposing individuals to subsequent onset of other disorder groups above and beyond a common liability model. Data were drawn from the 2012-2013 National Epidemiologic Survey on Alcohol and Related Conditions-III (n = 36,309). We identified statistically significant sequences of the age of onset of mood, anxiety, and substance use disorders (SUD) using a novel method distinguishing significant sequences from patterns expected only due to correlations induced by common liability among disorders. At every age, the lifetime probability of the three disorder groups was positively correlated consistent with a common liability model, but using a sequence analysis that accounts for different population level age of onset patterns and age specific correlations due to common liability, only SUD was more likely to precede mood disorders than expected (4.0% vs 3.3%, p < .001). Onsets within the same year of mood and anxiety disorders (21.5%) occurred over 6 times more commonly than expected by chance. SUD and mood disorders onsets occurred during the same year (9.0%) twice as commonly as expected, whereas SUD and anxiety disorder onsets did not occur in the same year significantly more commonly than expected. These results suggest that above and beyond common liabilities to comorbid mood, anxiety, and SUD disorders across the lifespan, SUDs predispose individuals for future onset of mood disorders, but not the reverse. Frequent simultaneous onset of psychiatric disorders further supports shared underlying vulnerabilities across disorder groups.