One limitation of the present analyses includes the varying definitions of the Substance Use Disorders for those diagnosed in childhood and adolescence (DSM-III) and those receiving a DSM-IV diagnosis from the age-appropriate CIDI interview that was administered for those seen after age 19. Another limitation was that some predictors and outcomes both occurred in close proximity during childhood and adolescence during the 8–18 year old portion of the study while for others the childhood predictor (Conduct Disorder or ADHD) occurred several years before the young adult SUD outcome This would tend to introduce variation in the predictor/outcome relationship because conditions more proximal might be expected to have greater influence than those separated in time. This may be of more limited concern because there is good reason to believe Conduct Disorder and ADHD are persistent disorders that once diagnosed rarely remit.
