Standardisation of the Test Your Memory and evaluation of their concordance with the outcome of the psychometric examination

Abstract

ObjectiveTo explore the relationship between scores on the Test Your Memory (TYM) battery and findings from a more exhaustive neurocognitive assessment. MethodsThe TYM and fourteen psychometric tests were administered to 84 subjects aged 50 or older who attended an outpatient neurology clinic due to cognitive symptoms. Each patient's cognitive state was determined independently from his/her score on the TYM (CDR 0, n=25; CDR 0.5, n=45; CDR 1, n=14). We analysed concurrent validity of TYM scores and results from the psychometric tests, as well as the degree of concordance between the two types of measurement, by contrasting normalised data from each instrument. ResultsAlthough the intraclass correlation coefficient was 0.67 (confidence interval 95%, 0.53-0.77), analysis of the Bland–Altman plot and the curve on the survival-agreement plot (Luiz et al. method) demonstrates that the individual distances between the two methods exhibit excessive dispersion from a clinical viewpoint. TYM-based predictions of the mean z-score on psychometric tests differed substantially from real results in 30% of the subjects. Concordance of 95% can only be achieved by accepting absolute inter-instrument differences of up to 0.87 as identical values. Furthermore, the TYM underestimates cognitive performance for low values and overestimates it for high values. ConclusionsThe TYM is a cognitive screening test which should not be used to predict results on psychometric tests or to detect cognitive changes in clinical trials.