ABSTRACT
This study examines the sensitivity, specificity, predictive values, and likelihood ratios of laboratory measures of attention and impulsivity (the Gordon Diagnostic System; GDS) in 99 school-aged boys with a history of suspected language disorders. Classification analyses comparing scores from these tests with parent and teacher ratings of attention deficit hyperactivity disorder (ADHD) symptoms revealed low positive predictive values (20.0% to 36.8%) and high negative predictive values (71.9% to 87.9%). Likelihood ratios for abnormal test scores were low to moderate (0.74 to 1.73), suggesting that these tests may not accurately identify children with ADHD. Likelihood ratios for normal scores were also low to moderate (0.41 to 1.16). These findings suggest that GDS scores have clinical utility in ruling out a diagnosis of ADHD, but not in confirming the diagnosis in a clinic population of boys with communicative disorders who are at risk for developing ADHD.
