Convergent validity of measures of daily functioning in older adults.
Abstract
INTRODUCTION Although information on daily functioning is needed for accurate diagnoses of neurocognitive disorders, little has been published on the psychometric properties of measures of daily functioning. Convergent validity is one such psychometric property that examines the relationship between scales that purport to assess the same construct. Whereas the convergent validity of measures of daily functioning has typically been examined relative to cognitive tests, the current objective was to analyze the relationships between different measures of daily functioning in older adults.
Methods
Two hundred thirty-seven older adults (age: M = 74.8, SD = 5.6; education: M = 16.9, SD = 2.4) completed a battery of functional assessments, including two subjective scales (Alzheimer's Disease Cooperative Studies - Activities of Daily Living for Mild Cognitive Impairment 23-item version and Quick Dementia Rating System) and three performance-based measures (Independent Living Scales, Night-Out Task, and Virtual Kitchen Challenge). Scales were initially compared with Spearman's correlations and then factor analysis.
Results
Small to medium but statistically significant correlations (p < 0.01) were found among the scores of all functional measures, with better functioning on one scale being seen on other scales (rs = 0.21 to 0.61). Three factors were identified, and scores loaded together based on type of scoring (i.e. correct/incorrect vs. timed) and by assessment method (i.e. subjective vs. performance-based).
Conclusions
Despite differences between these measures that assess daily functioning, their associations suggest that they tap a similar construct. Factor loadings were based on type of scoring within the performance-based measures (i.e. correct/incorrect vs. timed) and type of assessment method (i.e. subjective or performance-based). While further exploration is needed in larger, more diverse samples and using additional functional scales, these findings have implications for both clinical practice and research.