Combined information from multiple sources increases the validity of ADHD diagnosis in children; as indicated by results of a psychometric modelling study.
This study attempted to standardise how symptom reports from multiple informants were collated to diagnose ADHD. A community sample of 725 children, their primary caregivers (675 mothers, 476 fathers) and 629 teachers participated in this study. Of these children, 330 (45.5%) exhibited clinically significant levels of ADHD symptoms.
Parents/teachers of eligible children with ADHD completed a thorough diagnostic assessment,* which was passed to a clinical diagnostic team† who provided a ‘gold-standard’ best-estimate diagnosis. This ‘gold-standard’ diagnosis was then compared with three algorithms for collating parent and teacher symptom reports: (i) symptom present if reported by parents or teachers; (ii) symptom present if reported by both parents and teachers; and (iii) average ratings of symptom severity considered, even if parents/teachers disagree.
Structural equation modelling yielded a theoretical model of ADHD diagnosis, which comprised two factors — hyperactivity-impulsivity and inattention; each with nine associated symptom items — that relied on ratings from mothers, fathers and teachers. Applying constraints to test the model indicated that whilst both parents and teachers were equally valid raters of hyperactivity-impulsivity symptoms, teachers were better raters of inattention symptoms, which may reflect their greater opportunity to observe children in situations that demand attention.
To compare the performance of the different scoring algorithms, scores of hyperactivity-impulsivity and inattention were generated for mother, father and teacher ratings of ADHD symptoms using the ‘average’, ‘or’ or ‘and’ algorithms, and regressed onto all nine symptom items of the model. Correlation analyses indicated that scores of the ‘average’ algorithm were most strongly correlated with the model. This correlation increased in strength when average scores from just four key symptom items for hyperactivity-impulsivity and inattention were included; and this ‘short average’ approach performed as well as or better than the ‘or’ approach in terms of diagnostic sensitivity (83% vs 68%, respectively) and specificity (91% for both), when both were compared with the ‘gold-standard’ diagnosis made by the clinical diagnostic team.
Researchers concluded that information from multiple raters increases the validity of ADHD diagnosis, with averaging an optimal way to combine information from multiple raters. Averaging also makes it possible to reduce the number of symptom items from 18 to 8 without a significant decrease in validity.
*The diagnostic assessment comprised diagnostic interviews, rating scales and cognitive testing
†The clinical diagnostic team comprised a child clinical psychologist and a board-certified child psychiatrist, who utilised a structured diagnostic interview, clinician observations and parent/teacher ratings to arrive independently at a best estimate option
Martel MM, Schimmack U, Nikolas M, Nigg JT. Integration of symptom ratings from multiple informants in ADHD diagnosis: a psychometric model with clinical utility. Psychol Assess 2015; 27: 1060-1071.