Revisiting ADHD age-of-onset in adults : to what extent should we rely on the recall of childhood symptoms?

Background. ADHD diagnosis requires the presence of symptoms before the age of twelve. In clinical assessment of adults, the most frequent strategy to check this criterion is investigating self-report recall of symptoms, despite little evidence on the validity of this approach. We aim to evaluate the recall accuracy and factors associated with its reliability in a large populationbased sample of adults. Methods. Individuals from the 1993 Pelotas Birth Cohort were followed-up from childhood to adulthood. At the age of 22, 3810 individuals were assessed through structured interviews by trained psychologists regarding mental health outcomes, including ADHD diagnosis and ADHD symptoms in childhood. The retrospective recall was compared with available information on ADHD childhood symptoms at the age of eleven. We also assessed factors related to recall accuracy through multiple regression analyses. Results. Self-reported recall of childhood symptoms at 22 years of age had an accuracy of only 55.4%, with sensitivity of 32.8% and positive predictive value of 40.7%. Current inattention symptoms were associated with lower risk and social phobia with higher risk for false-positive endorsement, while higher levels of schooling correlated with lower risk and male gender with higher risk for false-negative endorsement. Conclusions. Clinicians treating male patients with social phobia and ADHD symptoms should assess even more carefully retrospective recall of ADHD childhood symptoms. Moreover, characteristics associated with recall improvement do not impact accuracy robustly. In this context, the recall of childhood ADHD symptoms seems an unreliable method to characterize the neurodevelopmental trajectory in adults with currently-impairing ADHD symptomatology