Construct validity of the mini mental state examination across time in a sample with low-education levels: 10-year follow-up of the Bambuí Cohort Study of Ageing

• Published in: International journal of geriatric psychiatry Vol. 29; no. 12; pp. 1294 - 1303
• Main Authors: Castro-Costa, Érico, Dewey, Michael E., Uchôa, Elizabeth, Firmo, Josélia O. A., Lima-Costa, Maria Fernanda, Stewart, Robert
• Format: Journal Article
• Language: English
• Published: Hove Blackwell Publishing Ltd 01.12.2014; Psychology Press; Wiley Subscription Services, Inc
• Subjects: Academic achievement; Aged; Aged, 80 and over; Aging; Biological and medical sciences; Brazil; Brief Psychiatric Rating Scale - standards; Cognition Disorders - diagnosis; construct validity; Educational Status; factor structure; Female; Follow-Up Studies; General aspects; Geriatric Assessment; Geriatric psychiatry; Geriatrics; Humans; invariance across time; Longitudinal Studies; Male; Medical sciences; Mental health; Middle Aged; Miscellaneous; MMSE; Older people; Psychology. Psychoanalysis. Psychiatry; Psychometrics; Psychopathology. Psychiatry; Public health. Hygiene; Public health. Hygiene-occupational medicine; Time Factors; Validation studies; Brazil; Human; Mini Mental State Examination Folstein et al; Senescence; Factor analysis; Education level; Ageing; Neuropsychological test; invariance across time; Time; factor structure; Test validation; Construct validity; Gerontology; Follow up study; Cohort study; MMSE; Elderly; Psychiatry; Public health; Geriatrics; construct validity
• ISSN: 0885-6230; 1099-1166; 1099-1166
• DOI: 10.1002/gps.4113

Objective The study aims to investigate whether longitudinal data on the structure of the mini mental state examination (MMSE) collected in an older Brazilian cohort support factorial invariance over time. Design Analysis of 10‐year data from a community‐based cohort study was performed. Setting The study took place in Bambuí, Brazil. Participants The study sample comprised 1558 (89.4%) of all eligible 1742 elderly residents. Measurements A standard Brazilian version of the MMSE was used. Results A five‐factor solution (developed on the baseline of the cohort) either with no constraints or with loadings constrained to equality across time provided a reasonable fit for the MMSE. A comparison between both models suggested that the model with no constraints was superior. However, the five absolute goodness‐of‐fit indices suggest that the fully constrained model was also adequate and did not differ substantively from the model without any restriction. Conclusion The structure of the MMSE remained relatively unchanged across the 10 measurement times, thus providing evidence for the good construct validity of the scale across time. Copyright © 2014 John Wiley & Sons, Ltd.