ABSTRACT
Objective:To study on performance comparison between clinic digital memory assessment and previously used assessments in dementia risk screening.To compare the performance comparison between Beijing Aging Brain Rejuvenation Initiative(BABRI)Brain Health System's Clinic Digital Memory Detection as study and Alzheimer Disease-8(AD8)and the Brief Community Screening Instrument for Dementia(BCSID)as controls, We calculated and evaluated the accuracy, sensitivity and specificity of screening Mild cognitive impairment(MCI)among these tests.Furthermore, BABRI Brain Health System was used to conduct a large sample brain health examination and early dementia screening to test the validity, adaptability and stability of the evaluation results by BABRI Brain Health System'Clinic Digital Memory Detection.Methods:Dataset 1 contained 669 elderly subjects from five communities in Beijing were recruited according to inclusion and exclusion standard.The diagnosis of MCI was based on the full set of neuropsychological scale and Petersen standard.Dataset 1 was used to compare the discriminant effect of BABRI Brain Health System'Clinic Digital Memory Detection as study versus AD8 and BCSID as controls.The sensitivity, specificity, positive predictive value, negative predictive value and Youden index of each measurement tool were calculated.Then, the receiver operator characteristic(ROC)curve was prepared to compare the discrimination ability of MCI between each measurement tool.While the area under the curve(AUC)of different tools was compared by Wald χ2 test.Dataset 2 contained 284 103 subjects from 16 communities in Beijing, which were used to test the applicability of large sample screening in BABRI Brain Health System. Results:77 patients with MCI were found among 666 people, and incidence rate was 11.56% using the full set of neuropsychological scales in dataset 1.Compared with the results of other tests, the sensitivity of BABRI Brain Health System to correctly distinguish MCI was 0.753, which was close to BCSID, and better than AD8.In addition, BABRI Brain Health System's Youden's index was 0.741 and AUC was 0.905, which suggested that the specificity, positive predictive value, negative predictive value and cognitive domain coverage of MCI screening were generally better in BABRI Brain Health System than in AD8 and BCSID.Finally, the Brain Health Examination results of 284, 000 people in dataset 2 showed that the high-risk detection rate of MCI(8.65%)of the tool for people over 50 years old under a large sample was quite close to the results of dataset 1(8.67%), indicating that the BABRI Brain Health System had high stability.Conclusions:BABRI Brain Health System has not only high sensitivity and specificity, but also wide cognitive field coverage and high stability.BABRI Brain Health System is suitable for large-scale brain health examination and dementia risk screening in grass-roots communities, and is worthy of popularization.
ABSTRACT
Objective:To develop rapid screening tools for assessing the risk of mild cognitive impairment(MCI)based on neuropsychological scales and cognitive paradigms.Methods:Two baseline datasets from the Beijing Ageing Brain Rejuvenation Initiative(BABRI)cohort were studied: dataset 1 contained 5 593 subjects, with 1 500 cases with MCI and 4 093 cases with normal cognitive function(the control group); dataset 2 consisted of 588 subjects, with 92 cases with MCI and 496 cases with normal cognitive function(the control group). Dataset 1 was used to simplify the Mini-Mental State Examination(MMSE), and the sub-item combination with the strongest MCI discriminative ability was selected to integrate into the cognitive rapid assessment(BABRI-mini MMSE). Dataset 2 with scores of encoding-recognition episodic memory task was used for further MCI discriminant analysis and was adapted into an episodic memory test(BABRI-EMT). We applied the receiver operating characteristic curve(ROC)for those analyses.Results:The control group and the MCI group showed significant differences in multi-domain cognitive ability and episodic memory task performance( P<0.01). Among sub-items of MMSE measured using dataset 1, MMSE12 and MMSE19 had the highest discriminative accuracy for MCI, and the area under the ROC(AUC)was 0.699 and 0.631, respectively.Dataset 2 was used to investigate the discriminative ability of the episodic memory score in combination with the above two MMSE sub-items for MCI, and the AUC value was 0.732, the sensitivity was 0.731, and the specificity was 0.656. Conclusions:The BABRI-mini MMSE and BABRI-EMT are suitable for the large-scale universal screening of MCI risk.