Detailed analysis of error patterns in the number-transcoding task on the Japanese version of the Rapid Dementia Screening Test

Aim: The number-transcoding task on the Japanese version of the Rapid Dementia Screening Test requires mutual conversion between Arabic and Chinese numerals (e.g. 209 → 二百九, 4054 → 四千五十四, 六百八十一 → 681, 二千二十七 → 2027). During this task, some characteristic errors have been seen among patients with Alzheimer's disease (AD). The objective of this study was to clarify whether the frequency of appearance of error patterns differs between patients with mild and severe AD according to Clinical Dementia Rating (CDR) scores. Methods: A total of 250 patients with AD were recruited and subsequently categorized into two groups based on CDR scores (mild AD: CDR of 0.5 or 1; severe AD: CDR of 2 or 3). We analyzed 19 qualitative error patterns, including 15 that had been reported to date and 4 previously unreported errors, in each subtest. Results: The frequency of appearance of two previously reported and four previously unreported errors in the mild and severe AD groups, respectively, were statistically significant. Conclusions: Characteristic error pattern distributions in number transcoding can be observed in patients with mild and severe AD according to CDR scores and offers useful information for interpreting cognitive screening data.