Update on The Evidence-Based Assessment of Attention-Deficit/Hyperactivity Disorder in Youth

The goal of this review is to provide an updated evaluation of the reliability and validity of ADHD assessment measures since the last comprehensive review.

Using an empirical rationale for measure selection, we evaluated measures with a revised version of the Rubric for Evaluating Norms, Validity, and Utility.

The ADHD-RS, Conners, SNAP, and Vanderbilt scales had at least adequate descriptive information and excellent sensitivity to treatment, with most also showing adequate internal reliability and construct validity. Few studies examined clinical interviews and information was insufficient on most indices; only the DISC and K-SADS demonstrated adequate construct validity. Neuropsychological tests generally had adequate descriptive information, reliability, and construct validity, but discriminative validity and treatment sensitivity varied, with BRIEF values possibly inflated. Findings on impairment rating scales were mixed, though the AAPC, COSS, and IRS showed excellent sensitivity to treatment, and the IRS demonstrated broader reliability and validity. Strikingly little work addressed interrater or retest reliability, content validity, or generalization of reliability or validity. Reporting of ethnicity/race, SES, and medication status was inconsistent. Although concerns about reporting design constraints were raised a decade ago, we found no improvement in this regard: internal reliability was reported in only 18% of studies and assumptions were rarely tested.

We recommend more rigorous empirical research and further revisions to the rubric to advance reproducible science.