Validation of the Efficient Online MCI Screening System (EOmciSS) for Screening of Mild Cognitive Impairment Among Community-dwelling Older Adults (Preprint)

BACKGROUND Early detection of mild cognitive impairment (MCI) symptoms is an important step to the subsequent medical diagnostic and intervention. We developed a new screening test called the Efficient Online MCI Screening System (EOmciSS) for use in community-dwelling older adults. It is a self-paced rapid cognitive test to be completed within 10 minutes on tablets or mobile phones in homes or elderly centers. OBJECTIVE This study was aimed to test the validity of the EOmciSS for identifying community-dwelling older adults with MCI risks. METHODS The participants (N=827) completed the EOmciSS and other criterion instruments. The evidence of the psychometric properties included subscale item difficulty, discriminative index, internal consistency, and construct validity. We further tested the between-group discrimination using cross-validation method on the labelled MCI and normal cognitive groups. RESULTS Three accuracy and one reaction time factors explained the structure of the 20 item factors. Difficulty levels of the item factors in the accuracy factors, namely Trail Making, Clock Drawing and Cube Copying, and Delayed Recall were 0.63 to 0.99; while those of the reaction time factor were 0.77 to 0.95. Discriminative indices of the medium to high difficulty item factors were ranged from 0.39 to 0.97. The internal consistency indices (Cronbach’s Alpha) ranged from 0.41 (for few item factor) to 0.96. Training dataset indicated nine item factors as the significant predictors for the MCI versus NC classification. Presence of depressive symptoms was identified as a significant factor influencing participants’ performance, which was an integral part of the test system. Age, education level and proficiency level of using electronic device of the participants did not significantly influence their scores and classification. Application of the MCI/NC cut off score (7.90 out of 9,67) to the validation dataset yielded an AUC value of 0.912 (P < .001) (95% C.I. 0.868-0.955). The sensitivity and specificity were 84.9% and 85.1% respectively, and the Youden index was 0.70. CONCLUSIONS The EOmciSS shows satisfactory evidence on its validity and reliability for identifying older adults with significant risks of MCI. The user-interface, online operation and self-paced format allow the test system to be operated by older adult users or their caregivers in different community settings such as home and elderly centers. Depressive symptoms, found influencing the test performances and hence the MCI risks, should be an integral part in future MCI screening system. CLINICALTRIAL Chinese Clinical Trial Registry ChiCTR2000039411