Validity of a reduced model of items of DSM-IV criteria according to the parents' and teachers' responses in the diagnostic of the Attention Deficit Hyperactivity Disorder (Combined Type)

Main Objective: To look for a reduced model of symptoms of the attention deficit hyperactivity disorder combined type (ADHD-C) that shows suitable levels of criterion validity for the diagnostic of the disease. Methods: Epidemiological study. Sample of 1095 children between 6 and 16 years. First stage: psicometric study using ADHD RS-IV answered by parents (P) and teachers (I). ADHD is suspected when both questionnaires are over 90th percentile. Second stage: Clinical interview DISC-IV (DSM-IV) only in those selected cases to confirm ADHD-C. Logistic regression is implemented to find the most parsimonious model to predict ADHD-C. Results: The model that predicts clinical ADHD-C consists of 8 of the 36 items of the ADHD RS-IV answered by P and T. Considering the Odds Ratio obtained by regression, the selected items are as follows: 15T, 1P, 16T, 12P, 17P, 10P, 14P and 4T. The model has criterion validity for symptomatic ADHD-C (sensitivity: 97.9%. Specificity: 93.8%. Likelihood ratio: 16.02). Conclusions: It is possible to reduce the list of symptoms of ADHD-C with good criterion validity, removing redundant items and keeping those that provide greater discrimination between ADHD-C and the general population.