Reliability of global impressions for assessing methylphenidate effects in children with attention-deficit hyperactivity disorder.

Agreement between raters using global impressions to assess methylphenidate response was analyzed for children with Attention-Deficit Hyperactivity Disorder (ADHD) undergoing double-blind, placebo-controlled, crossover methylphenidate trials. Caregivers were more likely to disagree than agree when asked to rate the children as "better, same, or worse" during each day of the trial. Over-all agreement was 42.9%, only 9.6% above what would be expected based on chance alone. Further, none of the interrater reliability coefficients (Cohen's kappa) for the individual children were statistically significant.