Related Experiment Video
Updated: May 9, 2026
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
[Effectiveness of the Mini-Mental State for detection of cognitive impairment in primary care]
Cristóbal Carnero Pardo1, Isabel Cruz Orduña, Beatriz Espejo Martínez
1Unidad de Neurología Cognitiva y Conductual, Servicio de Neurología, Hospital Universitario Virgen de las Nieves, Granada, España; FIDYAN Neurocenter, Granada, España.
Objective:
To evaluate the diagnostic accuracy (DA) of the Mini-Mental State (MMS) for the detection of cognitive impairment (CI) in Primary Care (PC) and to determine the best conditions of use for that purpose.
Design:
Pooled analysis of two prospective, double blind, studies on the evaluation of diagnostic tools with complete verification that were conducted in Madrid and Granada (Spain).
Setting:
The MMS was administered in PC and the final cognitive diagnosis (gold standard) was made in Specialized Care.
Participants:
Subjects with cognitive complaints or suspected of having CI were consecutively recruited in the PC clinic.
Principal Measures:
The DA of the MMS was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC). The best cut-off point was selected according to the ratio of cases correctly classified (RCC) and to the kappa index. Direct (MMSd) and age- and education-adjusted (MMSa) total scores were analyzed separately.
Results:
In the total sample of 360 subjects (214 CI), the DA of the MMSd was significantly superior to that of the MMSa (0.84±0.02 vs 0.82±0.02, p≤.001). The yield obtained by the best cut-off point of the MMSd (22/23) was modest (RCC 0.77, kappa 0.52±0.05) and was not improved by any MMSa cut-off point.
Conclusion:
The DA of the MMS for detection of CI in PC was modest and did not improve with adjustment of the score by age and education. The best cut-off point was 22/23, inferior to the usually recommended cut-off.
