CERAD practice effects and attrition bias in a dementia prevention trial

The Consortium to Establish a Registry for Alzheimer's Disease (CERAD) set of tests is frequently used for tracking cognition longitudinally in both clinical and research settings. Repeated cognitive assessments are an important component in measuring such changes; however, practice effects and attrition bias may obscure significant clinical change over time. The current study sought to examine the presence and magnitude of practice effects and the role of attrition bias in a sample of cognitively normal older men enrolled in a prevention trial.

Participants were grouped according to whether they completed five years of follow-up (n = 182) or less (n = 126). Practice effects were examined in these participants as a whole (n = 308) and by group.

Findings indicate that moderate practice effects exist in both groups on the CERAD T-score and that attrition bias likely does not play a contributing role in improved scores over time.

The current study provides additional evidence and support for previous findings that repeated cognitive assessment results in rising test scores in longitudinally collected data and demonstrates that these findings are unlikely to be due to attrition.