Comparing the associations between performance-based, participant-reported, and study partner-reported instrumental activities of daily living tasks and cognition in cognitively unimpaired adults

Background Assessing instrumental activities of daily living (IADLs) is an integral diagnostic process for cognitive decline due to Alzheimer's disease (AD). While there are questionnaire-based and performance-based IADL measures, there is no clear consensus on the best methodology for a preclinical population. Objective We aimed to compare performance-based and questionnaire-based IADL measures and investigate whether IADL measures differently predict cognition. Methods One hundred and sixty-one cognitively unimpaired participants (mean age=70.69 ± 6.33 years; 62% female) completed performance-based IADL measures (i.e., Harvard Automated Phone Task [APT]; Czaja Functional Assessment Battery [CFAB]), questionnaire-based IADL measures (i.e., Cognitive Function Index [CFI]; The Alzheimer's Disease Cooperative Study Activities of Daily Living-Prevention Instrument [ADCS ADL-PI], also completed by study partners), and cognitive assessments. Separate cross-sectional linear regressions were used to determine (1) whether IADL questionnaires predicted performance-based measures, (2) whether IADL assessments predicted cognition. Results Greater impairment on participant self-reported questionnaires (CFI and ADCS ADL-PI) were associated with worse performance on the CFAB automated teller machine task (R 2 = 0.36, p = 0.010; R 2 = 0.44, p = 0.003). Worse performance on the APT primary care physician task and greater impairment on participant self-reported CFI were associated with worse executive functioning (R 2 = 0.13, p = 0.003; R 2 = 0.06, p = 0.035) and slower processing speed (R 2 = 0.15, p = 0.009; R 2 = 0.22, p < 0.001). Greater impairment on the participant self-reported CFI was the only assessment associated with worse memory scores (R 2 = 0.14, p = 0.005). Conclusions Among cognitively unimpaired older adults, few questionnaire-based IADL questionnaires were associated with performance-based measures. There was variability within and across IADL assessments’ associations with cognition, suggesting that not all IADL assessments are interchangeable.