Which two questions of Mini-Mental State Examination (MMSE) should we start from?

Abstract The MMSE is a simple and commonly used instrument to evaluate cognitive impairment. With the aim of enabling the examiner to skip a large portion of the MMSE when screening subjects with lower possibility of cognitive impairment, we examined the internal distribution of the MMSE scores among 792 older adults visiting a teaching hospital, a long-term care hospital, nursing homes, and a geriatric clinic. The correlation coefficients between the summed scores of any two items of MMSE and the total score were compared. A receiver operating characteristic (ROC) curve was drawn to show the sensitivity and the specificity of predicting cognitive impairment, which was defined by the total MMSE score being less than 24. The mean MMSE score was 20.5 ± 6.9 (±S.D.). A good predictor for cognitive impairment was the summed scores of the time orientation and serial sevens with a sensitivity of 98.2% and a specificity of 69.2% if cut-off was set at 7/7+. This finding appears to help streamline the screening process for cognitive impairment in general elderly population.