Study limitations include sample characteristics that may limit the generalizability of findings. The study sample was primarily Caucasian and recruited from one region of the U.S. Furthermore, all subjects in the present study were psychiatrically hospitalized and recruited into a large randomized controlled intervention trial. As such, the recruitment rate, although comparable to that for intervention trials of non-primary treatments, was lower than would be ideal and likely reflects the acuity of the inpatient setting and consequent adolescent and family distress, the lengthy informed consent process, and the fact that families were asked to consent prior to randomization to study condition. Our measure of parents’ histories of mental health problems is not ideal for several reasons. First, it was completed by only one informant, which was usually a mother. A second informant or other validation strategy was not used. Diagnostic criteria and timing of parental mental health problems were not assessed. Consequently, it is unclear when and for what length of time parental psychopathology was present during the youth’s development. It is possible the parent’s difficulties were resolved prior to
