P3‐127: Identification of cognitive decline using a longitudinal telephone screening versus brief, one‐day computerized testing
Year: 2008
Type: article
Abstract: Early identification of cognitive decline is important in order to provide appropriate planning and treatment to lessen morbidity of Alzheimer's disease. Ideally, brief measures that detect likely decliners can be identified. This study examines the relationship between two such measures: a longitudinal telephone-based cognitive screen and brief repeated computerized testing. Participants were 83 volunteers from a larger study of aging and cognition (mean age=78.8). Annual evaluation for 7 years used the AHEAD, a 42-point telephone cognitive screen. Following this, seven computerized CogState cognitive tasks were administered assessing processing speed (Detection), attention (Identification), working memory (One Back), executive function (Strategy Learning) and learning and memory (One Card Learning, Associate Learning, Continuous Paired Associate Learning). As suggested by Darby and colleagues (2002), participants completed CogState four times in one day to elicit learning trajectories. A regression model was fitted for each participant's sets of AHEAD and CogState scores, yielding estimated intercept and slope values. To examine the relationship between the tests, linear modeling for the CogState variables was performed with the AHEAD variables as covariates. AHEAD performance never fell below the clinical cutoff for impairment. There was a positive relationship between increasing serial AHEAD scores and increasing One Back speed and accuracy over the four trials (p<0.05). Those with better initial or ongoing AHEAD performance performed better on their first trial of Strategy Learning (p=0.02) and Continuous Paired Associate Learning (p=0.005). Participants with lower initial AHEAD scores also had a rapid decline in accuracy on One Card Learning over the four administrations, whereas higher initial AHEAD score was correlated with stable One Card Learning performance (p=.005). For all analyses, there were no significant differences based on age, gender, or education level. Overall, improving longitudinal performance on the AHEAD was related to better present performance on CogState measures of attention, memory, and executive function, but not processing speed. There were mixed results regarding the usefulness of multiple CogState assessments in one day. Our findings suggest that this brief computerized assessment may have utility in identifying patients with subtle cognitive decline.
Source: Alzheimer s & Dementia
Institutions: University of Michigan, The University of Melbourne
Language: English
Subfield: Psychiatry and Mental health
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2053864183