Study limitations include use of a convenience sample comprised of affected sibling pairs, which is not generalizable to all cocaine dependent adults. The reliability of SSADDA psychiatric diagnoses is less than that for substance dependence diagnoses and the validity of these diagnoses has not been formally demonstrated. Because individuals with a clinical diagnosis of schizophrenia were excluded from participation, there are limitations to the applicability of these findings to a more severely mentally ill patient population. Several variables were used to represent cocaine dependence severity, which may have led to false positive findings, although the use of a conservative multivariate analysis strategy increases the likelihood that relationships between psychiatric comorbidity and cocaine dependence severity are meaningful. Study variables were self-report, which may not accurately describe dependence severity or treatment and self-help program use. Replication is needed with broader samples of cocaine users, confirmation of substance dependence (e.g., toxicology), and external measures of dependence severity (e.g., legal or work records) and service utilization (e.g., medical records).
