Practice effect of repeated cognitive tests among older adults: Associations with brain amyloid status and other influencing factors

Abstract

AbstractBackgroundPractice effects (PE), after repeated cognitive measurements, may mask cognitive decline and remain a major issue in clinical and research settings. However, an attenuated practice effect may indicate cognitive decline or the presence of brain pathologies. Exploring factors that influence PE can inform strategies to distinguish true signal from bias in future research. This study aimed to evaluate practice effects on the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) scale, and their associations with brain amyloid status in a cohort of older adults enrolled in the Cognitive Health in Ageing Register: Investigational, Observational, and Trial Studies: Prospective Readiness cOhort Study (CHARIOT‐PRO) Sub‐study.Method502 cognitively healthy participants aged 60‐85 years were assessed with RBANS in both screening and baseline clinic visits (median time gap of 3.5 months). We assessed the reliability of the RBANS scale and tested PE based on differences between test and retest scores in total scale and domain‐specific indices. Multiple linear regressions were used to examine factors influencing PE, after adjusting for age, sex, educational attainment, and APOE‐ε4 carriage. PE and the initial RBANS score were also examined as predictors for amyloid positivity status based on defined thresholds, using logistic regression.ResultThe internal consistency (Cronbach's α=0.64) and test‐retest reliability (r=0.79) of RBANS were good. Participants’ Total Scale, Immediate Memory and Delayed Memory indices were significantly higher in the second test than in the initial test (d=3.8, 7.6 and 3.2, P<0.001). On the Immediate Memory index, the PE was significantly lower in the amyloid positive group than the amyloid negative group (d=6.5 vs 8.6, P=0.028). Older participants (≥70 years), women, non‐APOE‐ε4 carriers, and those with worse initial RBANS test performance had larger PE. In addition, worse performance and lower practice effects in immediate or delayed memory index were independent predictors for amyloid positivity (OR=0.97, P<0.05).ConclusionSignificant practice effects on RBANS total scale and memory indices were influenced by age, sex, APOE‐ε4, initial RBANS scores and amyloid status. The association with amyloid suggests that practice effects are not simply a source of measurement error but may be informative with regard to underlying neuropathology.