Cross-sectional assessment of telephone- and computer-administered instruments in detection of cognitive impairment

We investigated the associations of telephone- and computer-administered cognitive screening instruments for Alzheimer's disease with in-person measurement in a population-based sample of individuals without a prior diagnosis of dementia-causing neurodegenerative disease.

We studied 202 TWINGEN participants (126 female) who had in-person administered Consortium to Establish a Registry for Alzheimer's Disease neuropsychological battery (CERAD-nb) data. The participants were aged 65-85 years and of European ancestry. Telephone-administered instruments were telephone assessment for dementia (TELE) and modified version of Telephone Tnterview for Cognitive Status, and computer-administered instrument was web-based cCOG. We utilized correlation analyses, areas under the receiver operating characteristic curves (AUC) and hierarchical linear mixed-effects models in main analyses.

Screening instruments exhibited moderate to high correlations with CERAD-nb, with the strongest correlation between cCOG and CERAD-nb (r = .62). Despite exclusion criteria, 20 participants had cognitive impairment according to the CERAD-nb total score. When evaluating the instruments' ability to distinguish between participants with and without cognitive impairment, AUCs ranged from .65 (95% CI [.53, .78]) for TELE to .76 (95% CI [.67, .86]) for cCOG. All instruments predicted in-person CERAD-nb score in linear mixed-effects models, when controlling for demographic factors. Including both telephone- and computer-administered measures as predictors resulted in better accuracy compared to including only one modality as a predictor.

Performance in telephone- and computer-administered cognitive screening instruments was associated with performance in in-person neuropsychological battery designed for early detection of Alzheimer's disease. The results indicate validity of the computerized and telephone-administered tests at the population level.