Converting MMSE to MoCA and MoCA 5‐minute protocol in an educationally heterogeneous sample with stroke or transient ischemic attack

• Published in: International journal of geriatric psychiatry 2018-05, Vol.33 (5), p.729-734
• Main Authors: Wong, Adrian, Black, Sandra E., Yiu, Stanley Y.P., Au, Lisa W.C., Lau, Alexander Y.L., Soo, Yannie O.Y., Chan, Anne Y.Y., Leung, Thomas W.H., Wong, Lawrence K.S., Kwok, Timothy C.Y., Cheung, Theodore C.K., Leung, Kam‐Tat, Lam, Bonnie Y.K., Kwan, Joseph S.K., Mok, Vincent C.T.
• Format: Article
• Language: English
• Subjects: Age; cerebrovascular disorders; Cognitive ability; cognitive impairment; Demography; early diagnosis; Geriatric psychiatry; Geriatric psychology; Ischemia; neurocognitive disorders; neuropsychology; Stroke; Transient ischemic attack
• ISSN: 0885-6230; 1099-1166
• DOI: 10.1002/gps.4846

Background The Montreal Cognitive Assessment (MoCA) is psychometrically superior over the Mini‐mental State Examination (MMSE) for cognitive screening in stroke or transient ischemic attack (TIA). It is free for clinical and research use. The objective of this study is to convert scores from the MMSE to MoCA and MoCA‐5‐minute protocol (MoCA‐5 min) and to examine the ability of the converted scores in detecting cognitive impairment after stroke or TIA. Methods A total of 904 patients were randomly divided into training (n = 623) and validation (n = 281) samples matched for demography and cognition. MMSE scores were converted to MoCA and MoCA‐5 min using (1) equipercentile method with log‐linear smoothing and (2) Poisson regression adjusting for age and education. Receiver operating characteristics curve analysis was used to examine the ability of the converted scores in differentiating patients with cognitive impairment. Results The mean education was 5.8 (SD = 4.6; ranged 0–20) years. The entire spectrum of MMSE scores was converted to MoCA and MoCA‐5 min using equipercentile method. Relationship between MMSE and MoCA scores was confounded by age and education, and a conversion equation with adjustment for age and education was derived. In the validation sample, the converted scores differentiated cognitively impaired patients with area under receiver operating characteristics curve 0.826 to 0.859. Conclusion We provided 2 methods to convert scores from the MMSE to MoCA and MoCA‐5 min based on a large sample of patients with stroke or TIA having a wide range of education and cognitive levels. The converted scores differentiated patients with cognitive impairment after stroke or TIA with high accuracy.