Full-scale scores of the Mini Mental State Examination can be generated from an abbreviated version

Abstract

Objective To determine whether the full Mini Mental State Examination (MMSE) scale range can be derived from an abbreviated 11-item version that was designed for testing general cognitive function in a cohort where only a small proportion were expected to be severely impaired or demented. Study Design and Setting Using simple computation and multiple imputation, the properties of the abbreviated MMSE were compared with the full MMSE score using data from the Medical Research Council Cognitive Function and Ageing Study. Results Full MMSE scores can be generated for the abbreviated version by assuming high functioning on excluded items. Of the imputed scores, 88.8% were within 1 point of their true value. When the sample was restricted to individuals with normal cognitive functioning (MMSE total score ≥ 24/30), 96.7% of individuals were classified within 1 point of their true total score. The model worked best at predicting cognitive level when cutoff scores were used to classify individuals into impaired vs. not nonimpaired. Conclusion Full-scale MMSE scores can be reasonably accurately derived from an 11-item abbreviated version. This reduced version can be applied within other frameworks that require reduced test length but need results that are comparable to studies where the full version has been administered.