Evaluating working memory outcome measures for children with Down syndrome

Background There is a critical need for the psychometric evaluation of outcome measures to be used in clinical trials targeting cognition in Down syndrome (DS). This study examines a specific cognitive skill that is of particular importance in DS, working memory, and the psychometric properties of a set of standardised measurements to assess working memory in individuals with DS. Methods Ninety children and adolescents ages 6 to 18 years old with DS were assessed on a selection of verbal and visuospatial working memory subtests of standardised clinical assessments at two time points to examine feasibility, distributional qualities, test-retest reliability and convergent validity against a priori criteria. Caregivers also completed an adaptive behaviour questionnaire to address working memory subtests' associations with broader developmental functioning. Results The Stanford Binet-5 Verbal Working Memory, Differential Ability Scales-2 Recognition of Pictures, Stanford Binet-5 Nonverbal Working Memory and Wechsler Intelligence Scale for Children-5 Picture Span measures met the most psychometric criteria overall across the full age and IQ range of the study. Although Differential Ability Scales-2 Recall of Sequential Order and Differential Ability Scales-2 Recall of Digits Backward met the fewest a priori criteria, follow-up analyses suggested greater feasibility in specific age and IQ ranges. Conclusions Several working memory measures appear to be psychometrically sound and appropriate for use in clinical trials for children with DS, especially when focusing on raw scores. However, floor effects on standard scores and feasibility of some measures were problematic. Guidelines for use of the working memory subtests with this population are provided.

Automated summary

The study assessed the psychometric properties of working memory measures in children and adolescents with Down syndrome. Some measures showed good psychometric properties overall, but there were issues such as floor effects and feasibility with certain measures. Guidelines for use of working memory measures in this population are recommended.