Mini-Mental State Examination (MMSE)

Purpose of the measure

The Mini-Mental State Examination (MMSE) was originally developed as a brief screening tool to provide a quantitative evaluation of cognitive impairment and to record cognitive changes over time (Folstein, Folstein, & McHugh, 1975). Since that time it has become recognized that repeated use of the MMSE with the same client reduces its validity, so it is advised that this screening tool not be used repeatedly with the same individual if the time interval between testing is short. Rather than provide a diagnosis, the measure should be used to detect the presence of cognitive impairment (Folstein, Robins, & Helzer, 1983). The MMSE briefly measures orientation to time and place, immediate recall, short-term verbal memory, calculation, language, and construct ability. While the measure was originally used to detect dementia within a psychiatric setting, its use has become widespread. Since 1993, the MMSE has been available with an attached table that enables patient-specific norms to be identified on the basis of age and educational level (Crum, Anthony, Bassett, & Folstein, 1993).

Available versions

The MMSE was published by Folstein et al. in 1975.

Features of the measure

Items:
The MMSE consists of 11 simple questions or tasks that look at various functions including: arithmetic, memory and orientation.

Scoring:
The score is the number of correct items. The measure yields a total score of 30. A score of 23 or less is the generally accepted cutoff point indicating the presence of cognitive impairment (Ruchinskas & Curyto, 2003).

Levels of impairment have also been classified as none (24-30); mild (18-23) and severe (0-17) (Tombaugh & McIntyre 1992).

More recently, Folstein, Folstein, McHugh, and Fanjiang. (2001) recommended the following cutoff scores:

Score Level of impairment
≥ ? 27 None
21-26 Mild
11-20 Moderate
≤ 10 Severe

Crum et al. (1993) reported that cognitive performance as measured by the MMSE varies within the population by age and educational level. There is an inverse relationship between MMSE scores and age, ranging from a median of 29 for those aged 18 to 24 years, to 25 for individuals 80 years of age and older. There is also an inverse relationship between MMSE scores and education. The median MMSE score is 29 for individuals with at least 9 years of schooling, 26 for those with 5 to 8 years of schooling, and 22 for those with 0 to 4 years of schooling.

The following table, created by Crum et al. (1993) can be used to compare your patient’s MMSE score with a reference group based on age and education level.

(Source: Crum et al., 1993)

Age
Education 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 >84
4th grade 22 25 25 23 23 23 23 22 23 22 22 21 20 19
8th grade 27 27 26 26 27 26 27 26 26 26 25 25 25 23
High school 29 29 29 28 28 28 28 28 28 28 27 27 25 26
College 29 29 29 29 29 29 29 29 29 29 28 28 27 27

Subscales:
Orientation (total points = 10), Registration (total points = 3), Attention and calculation (total points = 5), Recall (total points = 3), and Language (total points = 9).

Equipment:
The MMSE requires no specialized equipment.

Training:
Little information has been reported on training for the MMSE, however a standardized version of the MMSE has been developed (Molloy & Standish, 1997).

Time:
Administration by a trained interviewer takes approximately 10 minutes.

Alternative form of the MMSE
  • The modified mini-mental state examination (3MS) (Teng & Chui, 1987).
    An expanded version of the MMSE was developed by Teng and Chui (1987) increasing the content, number and difficulty of items included in the assessment. The score of the 3MS ranges from 0 – 100 with a standardized cut-off point of 79/80 for the presence of cognitive impairment. This expanded assessment takes approximately 5 minutes more to administer than the original MMSE, which takes approximately 10 minutes to complete. Grace et al. (1995) compared the MMSE to the 3MS in geriatric patients with stroke. Test-retest reliability of the 3MS was excellent  (r = 0.80). The 3MS also correlated with a battery of neuropsychological assessments and with some cognitive domains missed by the MMSE. The 3MS was a significantly better predictor of functional outcome (as measured by the Functional Independence Measure) than the MMSE. The 3MS was found to have higher sensitivity than the MMSE (69% vs. 44%) and similar specificity (80% vs. 79%). The area under the curve (AUC) was 0.798 for the 3MS.

    3MS + Clock-drawing (Suhr & Grace, 1999).
    The addition of clock drawing, a simple measure of constructional ability, increased the sensitivity in detecting focal brain damage of the 3MS in patients with right hemisphere stroke (87%). The addition of the Clock Drawing Test requires about two extra minutes in administration time.

    Standardized MMSE (SMMSE) (Molloy & Standish, 1997).
    Molloy and Standish (1997) developed the SMMSE to improve the reliability of the measure. The idea was to develop strict guidelines for administration and scoring. To examine the reliability of the SMMSE in 48 older adults, university students were randomized to administer either the MMSE or the SMMSE, and were trained on that test to give to participants on three different occasions. The SMMSE had significantly better inter-rater and intra-rater reliability compared with the MMSE. The inter-rater variance was reduced by 76% and the intra-rater variance was reduced by 86%. It took less time to administer the SMMSE compared with the MMSE (average 10.5 minutes and 13.4 minutes, respectively. The intraclass correlation (ICC) for the MMSE was adequate (ICC = 0.69), and was excellent  for the SMMSE (ICC = 0.90).

    Telephone version (ALFI-MMSE) (Roccaforte, Burke, Bayer, & Wengel, 1992).
    This version includes 22/30 of the original MMSE items, the majority of which were removed from the last section (language and motor skills). Roccaforte et al. (1992) examined the validity of the ALFI-MMSE in 100 geriatric outpatients. Correlations between phone and face-to-face versions of the MMSE were excellent  (Pearson’s r = 0.85). Patients tended to score slightly higher on in-person testing than on the telephone. Sensitivity (using a brief neurological screening test as the criterion) of 67% and specificity of 100% were reported in a population of elderly, community-dwelling individuals. This was similar to the sensitivity (68%) and specificity (100%) reported for screening with the traditional MMSE.

    26-item version of the ALFI-MMSE (T-MMSE) (Roccaforte et al. cited in Newkirk, Kim, Thompson, Tinklenberg, Yesavage, & Taylor, 2004).
    The T-MMSE was developed from the ALFI-MMSE. It is a 26-point adaptation, containing a 3-step command: “Say hello, tap the mouthpiece of the phone 3 times, then say I’m back”. It also contains a new question that requests that the patient give the interviewer a phone number where they can usually be reached. The T-MMSE had an excellent  correlation with the MMSE (r = 0.88). Neither hearing impairment nor years of education were associated with T-MMSE scores. On the 22 points in common between the 2 scales, scores had an excellent  correlation (r = 0.88), however, telephone scores tended to be lower than in-face scores (Newkirk et al., 2004). The authors provide tables for the conversion of T-MMSE scores to MMSE scores.

Client suitability

Can be used with:

Patients with stroke (Agrell & Dehlin, 2000; Ozdemir, Birtane, Tabatabaei, Ekuklu, Kokino, & Siranus, 2001; Grace et al., 1995; Suhr & Grace, 1999).

Should not be used with:

    • The MMSE was ineffective in detecting cognitive impairment in patients with right-sided stroke (Grace et al., 1995).
    • The MMSE is not suitable for use with a proxy respondent as it is administered via direct observation of task completion.
    • Because the MMSE is heavily language dependent, it is likely to misclassify patients with aphasia.
    • The MMSE has a limited ability to diagnose dementia in general practice and should therefore be used as only one aspect of a patient’s overall cognitive profile (Wind, Schellevis, van Staveren, Scholten, Jonker, & van Eijk, 1997).
    • The MMSE has been criticized for attempting to assess too many functions in one brief test. An individual’s performance on individual items or within a single domain may be more useful than interpretation of a single, overall score (Tombaugh & McIntyre 1992). However, when used to screen for visual or verbal memory problems, or for problems in orientation or attention, it is not possible to identify acceptable cut-off scores (Blake, McKinney, Treece, Lee, & Lincoln, 2002).
    • MMSE scores have been shown to be affected by age, level of education, ethnicity, and sociocultural background (Tombaugh & McIntyre, 1992; Bleeker et al., 1988; Lorentz et al., 2002; Shadlen, Larson, Gibbons, McCormick, & Teri, 1999). These variables may introduce bias leading to the misclassification of individuals. For example, highly educated individuals who have mild dementia may well score within normal range on the MMSE because they find the questions easy. Further, poorly educated individuals may have low scores on the MMSE simply because they find the questions difficult. Thus, their scoring on the MMSE may indicate a diagnosis of dementia when none is present. Although these biases are not always present, Agrell and Dehlin (2000) found that age and education did not influence scores in their study, attention to these factors is warranted when interpreting MMSE results.
    • The MMSE has been found to lack sensitivity in patients with stroke (Blake et al., 2002; Suhr & Grace, 1999; Nys et al., 2005). Other studies have reported low levels of sensitivity among individuals with mild cognitive impairment (Tombaugh & McIntyre, 1992; de Koning et al., 1998) and in patients with right-hemisphere lesions (Dick et al., 1984). One potential solution to increase the sensitivity of the MMSE is the addition of a Clock Drawing Test (Suhr & Grace, 1999). Another solution that has been offered is to administer the Neurobehavioral Cognitive Status Examination (NCSE) in lieu of the MMSE. The NCSE is a highly sensitive measure to detect cognitive impairment in patients with brain lesions (Schwamm, Van Dyke, Kiernan, Merrin, & Mueller, 1997).
    • Da Costa et al. (2010) investigated the cognitive evolution and clinical severity of illiterate and schooled patients with stroke during a 6-month follow-up, using the MMSE and National Institutes of Health Stroke Scale (NIHSS) respectively. Significant improvement in clinical severity as measured by NIHSS was observed in both groups (P<0.001); however, only schooled individuals showed a significant improvement in MMSE scores, indicating an improvement in their overall cognitive function (P=0.008). Schooling was found to significantly influence MMSE scores.
    • Folstein, Folstein, and McHugh (1998) reported that the MMSE demonstrates marked ceiling effects in younger intact individuals and marked floor effects in moderately to severely impaired individuals.
In what languages is the measure available?

The following authorized translations of the MMSE are currently available:

Afrikaans Dutch Israeli English Romanian
Arabic Estonian Italian Russian
Argentinean Spanish Filipino Japanese Russian for Estonia
Belgian Dutch Finnish Kannada Serbian
Belgian French French Korean Slovakian
Bosnian Austrian German Latvian South African English
Brazilian Portuguese German Lithuanian Spanish
Bulgarian Greek Macedonian Swedish
Chilean Spanish Gujarati Malayalam Telugu
Chinese Hebrew Marathi Turkish
Croatian Hindi Norwegian UK English
Czech Hungarian Polish Ukranian
Danish Indian English Portuguese Urdu

Authorized translations of the MMSE can be obtained by contacting Custsupp@parinc.com or call 1.800.331.8378