P2‐322: Item Analyses for the Chinese Mini‐Mental State Examination
Binyin Li,Huidong Tang,Shengdi Chen
DOI: https://doi.org/10.1016/j.jalz.2016.06.1452
2016-01-01
Abstract:Mini-Mental State Examination (MMSE) is the most widely used tool in different clinic settings. Chinese version MMSE (CMMSE) also has its own threshold score for differentiating normal aging people, MCI patients and AD patients. However, these studies were carried out in the 1990s or early 2000s. As the development of education system in China, aging people in the 2010s are better educated. The sensitivity and specificity of items in CMMSE might change. In this study, we explored all 30 items in CMMSE and analyzed their own impact in diagnosis. We collected information of patients, including diagnosis, demographics, CMMSE scores, neuroimaging, and other psychometric assessments. A total of 997 patients were included and classified, with 783 cognitively impaired and 214 cognitively normal. The diagnostic accuracy of each item in the MMSE was assessed by calculating the area under the receiver operating characteristic curve (AUC). The relationship between cognitive impairment and individual items from the MMSE was also further assessed using multiple logistic regression model in which a subset of items from the MMSE was selected. An abbreviated cognitive assessment instrument was created. The paired difference in AUC between the MMSE and new assessment instrument was assessed. The AUC of total MMSE score was 0.808 (95% CI: 0.805-0.854). In the MMSE, AUCs of orientation to time and place, registration and recall, calculation, repetition, reading, writing, comprehension and drawing were significant, while that of naming was not. Among these, the AUCs of orientation to time and place, attention, recall were over 0.700. The logistic regression selected the following items: day, the second and third recall, the first step in comprehension task. As a result, we selected orientation to time, registration, recall, and comprehension as a new instrument. The AUC of new assessment instrument was 0.808 (95% CI: 0.782-0.834).