As mentioned above, the sample is neither from a clinical or epidemiological population but rather a family study. Comparison of the study’s results to inpatient samples is therefore limited. Further, many but not all potential comorbidities (e.g., full range of DSM-IV personality disorders) were assessed in this family study. However, the structured clinical interview employed (SSAGA) covered many relevant characteristics of comorbid mental and SUDs in bipolar I and II individuals. In previous studies, different assessment methods were employed, including self-reports, global measures of improvement and clinical questionnaires in samples with bipolar disorders and cannabis use. In this sample, comorbid CUD and bipolar disorder subjects were assessed which may in part explain different results across samples. However, the results from the COGA sample may be representative for a non-clinical but severely affected comorbid CUD and bipolar population where a thorough clinical assessment for additional alcohol- and SUDs is indicated.
