Recall of childhood psychopathology more than 10 years later

Objective: This study investigated recall in a sample of depressed, anxious, and normal children followed up as adults. Strengths of this study were that the length of the retest interval was substantial, follow-up information was collected by blind interviewers, and childhood diagnoses were clearly documented. Method: The sample consisted of 144 subjects with a childhood diagnosis of depression, 48 with a childhood diagnosis of anxiety, and 128 normal controls. Best-estimate diagnoses assigned at follow-up were compared with childhood primary diagnoses. Results: Reliability and sensitivity were fair for major depressive disorder (mean = 0.46 and 50%, respectively) and any depression (mean = 0.57 and 65%, respectively). Reliability and sensitivity were relatively lower for anxiety (mean = 0.32 and 43%, respectively). Sensitivity for any diagnosis was good (mean = 71%). Specificity was good among all diagnostic categories (range = 73%-100%). Results suggest better diagnostic recall for females than for males. Recall was slightly better for subjects who were older than age 12 during their original episode. Age-of-onset reliability was poor (major depressive disorder = 0.22, any depression = 0.22, and any anxiety = -0.13). Conclusions: Recall of any childhood disorder is moderately reliable and accurate. Recall of a specific disorder is less accurate. Depression was more likely to be recalled than anxiety. High specificity suggests that participants were not biased to report disorders not present in childhood.