Agreement Among Categorical, Dimensional, and Impairment Criteria for ADHD and Common Comorbidities. (August 2016)

• Record Type: Journal Article
• Authors: Sprafkin, Joyce; Steinberg, Elizabeth A.; Gadow, Kenneth D.; Drabick, Deborah A. G.
• Is Part Of: Journal of attention disorders. Volume 20:Number 8(2016:Nov.)
• Page Start: 665
• Page End: 673
• Publication Date: 2016-08
• Subjects: ADHD -- DSM-IV -- rating scales -- functional impairment; Attention-deficit hyperactivity disorder -- Periodicals; Attention -- Periodicals; 616.8589
• DOI: 10.1177/1087054712475083
• Languages: English
• ISSNs: 1087-0547

Abstract

Objective: To compare the results of categorically based versus dimensionally based scoring algorithms for a Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV )–referenced teacher rating scale for assessing ADHD and commonly co-occurring conditions and to determine their relative agreement with ratings of symptom-induced impairment.Method: Teachers completed Child and Adolescent Symptom Inventory–4R (CASI-4R) ratings for 1, 092 youth (ages 6-18 years) referred to a child and adolescent psychiatry outpatient service. Caseness was determined according to DSM-IV symptom count (categorical model) and T -score (dimensional model) criteria.Results: Agreement between symptom count and T -score cutoffs was generally good (kappa ≥ 0.61) for ADHD-Inattentive, ADHD-Hyperactive-Impulsive, ADHD-Combined (except adolescent females), Oppositional Defiant Disorder, and Conduct Disorder, but this was not the case for anxiety and depressive disorders where only 15% of kappas were good. Agreement of impairment cutoff with T- score and symptom count cutoffs ranged from poor to good.Conclusion: In general, although in many cases CASI-4R categorical and dimensional scoring algorithms generated similar results, there was considerable variability across disorders, age groups, scoring method, and in some cases, gender. Moreover, symptom counts and T -scores are not a proxy for assessing impairment suggesting that each scoring strategy likely provides unique information for clinical decision-making.