PROGRESS IN GERIATRICS

The Mini-Mental State Examination:

A Comprehensive Review
Tom N. Tombaugh, PhD, CPsych and Nancy J. Mclntyre, MA

Objective: The purpose of this paper is to provide a compre- of sensitivity for moderate-to-severe cognitive impairment
hensive review of information accumulated over the past 26 and lower levels for mild degrees of impairment. Content
years regarding the psychometric properties and utility of the analyses revealed the MMSE was highly verbal, and not all
Mini-Mental State Examination (MMSE). items were equally sensitive to cognitive impairment. Items
Participants: The reviewed studies assessed a wide variety measuring language were judged to be relatively easy and
of subjects, ranging from cognitively intact community resi- lacked utility for identifying mild language deficits. Overall,
dents to those with severe cognitive impairment associated MMSE scores were affected by age, education, and cultural
with various types of dementing illnesses. background, but not gender.
Main Outcome Measures: The validity of the MMSE was Conclusions: In general, the MMSE fulfilled its original goal
compared against a variety of gold standards, including DSM- of providing a brief screening test that quantitatively assesses
111-R and NINCDS-ADRDA criteria, clinical diagnoses, Ac- the severity of cognitiveimpairmentand documents cognitive
tivities of Daily Living measures, and other tests that puta- changes occurring over time. The MMSE should not, by itself,
tively identify and measure cognitive impairment. be used as a diagnostic tool to identify dementia. Suggestions
Results: Reliability and construct validity were judged to be for the clinical use of the MMSE are made. J Am Geriatr SOC
satisfactory.Measures of criterion validity showed high levels 40922-935,1992

he use of screening tests to provide brief, objective umenting cognitive change.” The MMSE was not, on

T measures of cognitive functioning has increased

dramatically over the last 10 years. Although a
substantial number of screening tests exist, the Mini-
its own, intended to provide a diagnosis for any partic-
ular nosological entity.
The MMSE consists of a variety of questions (see
Mental State Examination (MMSE)’ is the most widely Appendix), has a maximum score of 30 points, and
used. It is a popular clinical measure that is available ordinarily can be administered in 5-10 minutes. The
in many The MMSE is also widely used questions typically have been grouped into seven cat-
in epidemiological studies and community surveys. It egories, each rationally representing a different cogni-
forms part of the Diagnostic Interview Schedule tive domain or function: Orientation to time (5 points);
(DIS),” a structured interview recently used in a large Orientation to place (5 points); Registration of three
five-site Epidemiologic Catchment Area (ECA) stud
sponsored by the National Institute of Mental Health, x words (3 points); Attention and Calculation (5 points);
Recall of three words (3 points); Language (8 points)

‘
and has been incorporated into several standardized
interviews designed to assess co nitive impairment and
to help diagnose dementia.”, 13, Moreover, the MMSE
serves as one of the tests recommended by the National
Institute of Neurological and Communicative Disorders
and Visual Construction (1 point). Originally, however,
all of the orientation questions were combined into a
single orientation category, and the visual construction
task was classified as one of the language items.

and Stroke and the Alzheimer’s Disease and Related Variations

Disorders Association (NINCDS-ADRDA)15 to docu- Variations in the wording and content of some ques-
ment the clinical diagnosis of probable Alzheimer’s tions, as well as in the administration and scoring of
disease. The aim of this paper is to review information the MMSE, commonly occur. Some of these variations
accumulated over the past 26 years regarding the psy- are described below.
chometric properties and utility of the English version Orientation to Place Since the MMSE was devel-
of the MMSE. oped to test hospital patients, the orientation questions
require the respondent to specify the name and floor
DESCRIPTION OF THE MMSE of their hospital. However, alternative items frequently
Folstein, Folstein, and McHugh’ originally created are used when the MMSE is administered outside the
the MMSE to differentiate organic from functional hospital, particularly in community surveys and epi-
psychiatric patients. They stated explicitly that MMSE demiological studies.’6-’8
scores were “useful in quantitatively estimating the Registration and Recall of Three Words The
seventy of cognitive impairment” and “in serially doc- choice of words used to test a person’s ability to learn
and retain three words was left o r i ~ n a l l vto the discre-
Address correspondene to Tom N. Tombaugh, Psychology Department,
tion of the examiner. When the GMSE was incorpo-
Carleton University, Ottawa, Ontario, Canada, K1S 586. rated into the DIS, the words apple, penny, and table

JAGS 40:922-935, 1992

0 1992 by the American Geriatrics Societ!! 0002-8614/92/$3.50
JAGS-SEPTEMBER 1992-VOL. 40, NO.9 MINI-MENTAL STATE EXAM 923

were employed, a convention adopted by most subse- ability exists, a prediction substantiated by the .96
quent studies. Exceptions to this practice, however, alpha obtained with a mixed sample of hospital pa-
have included words such as shirt, brown, honesty, t i e n t ~Second,
. ~ ~ since the MMSE attempted to measure
flag, ball, tree, rose, ring, elephant, and d ~ g . ’ ~ - ’ ~ a variety of cognitive processes, item heterogeneity was
Attention and Calculation Folstein et al’ routinely intentionally created. Thus, in this case, lower alpha
administered the serial 7s task on every test. However, levels may be viewed as desirable.
patients were permitted to spell the word WORLD Test-Retest Reliability In order to reduce the in-
backward if they could not or would not perform the fluence that illness-induced changes might exert on
serial 7s task. While using WORLD as an alternative reliability estimates, reliability coefficients only are re-
task has been followed in many studies, several other ported from studies where the test-retest interval was
procedures have been adopted. Some applications, 2 months or less. The results from these studies (Table
including CERAD,13 use only WORLD23-26 while 1) show that reliability coefficients for both cognitively
others, including CAMDEX,” use only the serial 7s intact and impaired subjects generally fell between .80
task.”, 2op27, Others routinely include both tasks, and .95. These reliability estimates are consistent with
which are scored in one of the following ways: (1)the those reported by Lesher and WhelihanS3 for other
higher of the two scores is used 2,16,19,21,29-33; (2) the brief cognitive screening tests. The unusually low coef-
two scores are combined21,30r34r35 or (3) each task is ficient of .56 for delirium patients45probably reflects
analyzed separately.16,22, 36, 37 Variations also exist in
how WORLD is scored.’,13, 23, 36, 38 coefficient for the control subjects in the Moms et al. z
the fluctuating course of this illness. The .38 reliabilit

study probably reflects the truncated distribution of

General Scoring Procedures and Cut-off Scores scores (ie, mean MMSE = 28.9) that statistically restricts
Ordinarily, the MMSE score is the total number of correlation coefficient^.'^ Several studies reported in-
correct answers. Although, this may not be appropriate creased scores on retest, presumably due to practice
when individuals refuse to answer many questions, effects.l. 29,32,46.48 Moreover, evidence exists suggesting
Fillenbaum et al.39concluded that, at least in epidemi- that some patients ”study” for mental status tests by
ological surveys, a refusal most likely represents an rehearsing answers given on a previous administration
inability to correctly answer the question. of the MMSE.” The site of testing can also influence
A score of 23 or less generally has been accepted as retest scores.37For example, two studies reported that
indicating the presence of cognitive impairment. This patients tested at home achieved significantly higher
cut-off score evolved from research findings rather scores than when they were tested in a clinic.56.57
than being recommended in the original article. Its The results with short test-retest intervals are corn‘-
high degree of acceptance is illustrated by the fact that pared to those obtained in six studies that used ex-
several community surveys have employed the cut-off tended test-retest intervals (eg, 1 to 2 years) with
score, even though modified forms of the MMSE were subjects judged to be cognitively inta~t.33.50~58-61 These
used.18, 31, 32, 34 More recently, largely due to the ECA results show that the amount of change was relatively
study,40the trend has been to classify the severity of small (usually within two points) and not statistically
cognitive impairment into three, rather than two, lev- significant. However, in the only two studies reporting
els: 24-30 = no cognitive impairment; 18-23 = mild reliability estimates,33r 6o correlation coefficients were
cognitive impairment; and 0-17 = severe cognitive less than .50, a value substantially less than those
impairment. MMSE scores are also frequently used to reported previously for studies using short test-retest
classify dementia patients as mild, moderate, or severe, intervals. Olin and Z e l i n ~ k iattribute
~~ the decreased
a practice supported by evidence from longitudinal reliability to several psychometric problems, including
studies (reviewed in the current paper) and dementia regression to the mean, method for assessing attention/
rating scale^.^'-^^ concentration, and lack of explicit scoring criteria for
the pentagon. The finding reported by O’Connor et
RELIABILITY a13’ that the second lowest kappa value for interrater
Internal Consistency Four studies that provide reliability occurred for the pentagon item further sug-
data on internal consistency (ie, item homogeneity) are gests that the lack of scoring criteria may affect the
shown in Table 1. The highest alpha level (.96) was stability scores. Regardless of their cause, the relatively
obtained with a mixed group of medical patients, while low reliability coefficients that occurred for “normal”
more modest levels (.68 and .77) were reported with subjects have important implications for using the
community samples. Jorm et a13’ reported an associa- MMSE with longitudinal assessment, suggesting that
tion between alpha levels and years of education: the small changes in scores should be interpreted with
alpha level for a community sample with only a pri- caution.
mary education (.65) was higher than for a sample
with 8 or more years of education (.54). Holzer et all6 VALIDITY
discuss two factors that may influence alpha levels. Sensitivity and Specificity One way to assess the
First, in community surveys, individuals correctly an- validity of the MMSE is to determine how well it
swered most of the questions, thereby reducing the correctly identifies normal and impaired individuals.
range of scores and decreasing the likelihood of obtain- The sensitivity of the MMSE refers to its ability to
ing high alpha coefficients. Higher alphas should be correctly identify those individuals who have been
obtained in clinical populations in which greater vari- classified as cognitively impaired according to some
924 TOMBAUGH AND MCINTYRE IAGS-SEPTEMBER 1992-VOL. 40, NO.9

TABLE 1. INTERNAL RELIABILITY AND TEST-RETEST RELIABILITY OF THE MMSE

Testmetest
Measure Sample n Age* Interval Correlations**
Internal Consistency
Holzer et all6 Community survey 4917 18-85+ .77
Kay et aP4*** Community survey 274 70-80+ .68
Foreman44 Medical patients (normal, 66 76 .96
dementia, & delirium)
Jorm et a13" Community survey 269 70+ .65 (grades 0-8)
.54 (> grade 8)
Test-Retest
Folstein et all Medical patients
1. Depression 22 41 1 day (same tester) .89
2. Depression 19 46 1 day (different .83
tester)
3. Dementia, depression 23 74 28 days (unspeci- .99
& schizophrenia fied)
Anthony et a145 Medical patients
1. Cognitively intact 58 20-80+ 1 day .85
2. Dementia 12 .90
3. Delirium 7 .56
Pfeffer et a14'j Mixed sample of demen- 23 58-86 4 days median in- .94
tia/delirium & cogni- terval
tively intact
Dick et a147 Neurological patients 15 50 1 day (different .95
tester)
44 1-70 days (same .92
tester) (M = 31)
Thal et aI4* Probable AD 40 50-90 1 week 34
3 weeks .79
6 weeks .80
Bird et a1' Community survey 189 44 same day .90
Fillenbaum et a139 Probable AD 24 54-75 1 month .89
O'Connor et a13' Cognitively intact 285 75+ 2 months .64
Dementia 196 75+ 33
Kafonek et a149 Dementia, delirium, & 29 65+ 1 week (different .84
depression tester)
Morris et a15" Control 278 68 1 month .38
Mild AD 200 72 1 month .74
Moderate AD 132 72 1 month .79
Teng et a15' Cognitively intact 27 64 days .79
van Belle et a15' AD 8 60+ 1-2 weeks .94
AD 30 60+ 3-4 weeks .85
Zaudig et all4 Cognitively intact 66 77 24-72 hours (M = .97
26 hours)
Jorm et alZ9 Mixed sample of demen- 57 80 1-14 days (M = .79
tia, depression & cogni- 2.8 days)
tivelv intact
* Single age score represents mean age. Pair of age scores represents range of ages.
**All correlations are significant, P < 0.05.
*** The Kay et a14 and Jorm et a13 analyses are based on data from the same study.

generally accepted criteria or gold standard (eg, DSM- It is also important to determine how well a positive
111, NINCDS-ADRDA, clinical judgment) (ie, true pos- or negative test result actually predicts the presence or
itives/total number of impaired cases). Specificity refers absence of impairment. If someone obtains a MMSE
to the MMSE's ability to correctly identify those indi- score of 22, for instance, what is the probability that
viduals who previously have been classified as cogni- cognitive impairment actually exists? The predictive
tively intact (ie, true negatives/total number of cogni- value of a positive test is the ratio of correctly identified
tively intact cases). Sensitivity and specificity data can positive cases to the total number of positive cases (ie,
be used to derive a likelihood ratio [sensitivity/(l- true positives/[true positives + false positives]), while
specificity)]that may be helpful in interpreting a MMSE the negative predictive value refers to the ratio of
score for an individual from a particular population correctly identified negative cases to the total number
(eg, community, hospital), provided the prevalence of of negative cases (true negatives/[true negatives + false
the cognitive impairment is known for the population. negatives]).
(For further information see Refs. 62 and 63). Table 2 shows 25 experiments for which sensitivity
1AGS-SEPTEMBER 1992-VOL.40, NO.9 MINI-MENTAL STATE EXAM 925
TABLE 2. CRITERION VALIDITY FOR MMSE USING 23/24 CUT-OFF SCORES
Mean
Sample Groups n Age* Score Criteria Results**
Dementia Subjects
Folstein et all 1. Cognitively intact 63 74 28 Psychiatric di- Sensitivity = 100%
Exp 1 2. Dementia 29 80 10 agnosis Specificity = 100%
Positive prediction = 100%
Negative prediction = 100%
Exp 2 Dementia 8 76 7 Psychiatric di- Sensitivity = 100%
agnosis
Exp 3 Dementia 9 74 12 Psychiatric di- Sensitivity = 100%
agnosis
Anthony et a145 1. Cognitively intact 74 20-89+ 26 DSM-111 Sensitivity = 87%
2. Dementia or de- 23 20-80+ 15 Specificity = 82%
lirium Positive prediction = 60%
Negative Prediction = 95%
Goldschmidt et 1. Cognitively im- 23 55 20 Clinical assess- Sensitivity = 100%
a164 paired ment
Folstein et all'
Comparison 1 1. Cognitively intact 106 65+ DSM-111 Sensitivity = 100%
(no clinical diag- Specificity = 62%
noses) Positive prediction = 44%
2. Dementia 32 65+ Negative prediction = 100%
Comparison 2 1. Cognitively intact 90 65+ DSM-111 Sensitivity = 100%
(clinical diag- Specificity = 46%
noses Positive prediction = 40%
2. Dementia 32 65+ Negative prediction = 100%
Kay et a134
Comparison 1 1. Cognitively intact 235 70-80+ DSM-111 Sensitivity = 69%
2. Dementia (mild, 39 70-80+ Specificity = 89%
moderate & se-
vere)
Comparison 2 1. Cognitively intact 235 70-80+ DSM-111 Sensitivity = 100%
2. Dementia (mod- 13 70-80+ Specificity = 85%
erate & severe
only)
Davous et aP5 1. Cognitively intact 56 70*** 27 DSM-I11 Sensitivity = 93%
Comparison 1 (functional disor- NINCDS- Specificity = 100%
ders & neurol- ADRDA Positive prediction = 100%
ogy) 44 74 15 Negative prediction = 95%
2. Dementia
Comparison 2 1. Cognitively intact 33 57 26 DSM-111 Sensitivity = 93%
(psychiatric ill- NINCDS- Specificity = 82%
ness) ADRDA Positive prediction = 87%
2. Dementia 44 74 15 Negative prediction = 90%
Fisk and Pannil166 1. AD 113 77 15 DSM-111 Sensitivity = 89%
Foreman4* 1. Cognitively intact 33 66-85 DSM-111 Sensitivity = 82%
2. Dementia or de- 33 66-85 Specificity = 80%
lirium Positive prediction = 80%
Negative prediction = 82%
Huff et a167 1. Cognitively intact 86 63 29 NINCDS- Sensitivity = 44%
2. AD 79 67 22 ADRDA
Pfeffer et a16' 1. AD 162 65+ DSM-111-R Sensitivity = 20%
NINCDS-
ADRDA
Jackson and 1. Cognitively intact 38 27 Clinical assess- Sensitivity = 85%
RamsdelP9 2. Dementia 294 15 ment Specificity = 87%
Positive prediction = 98%
Negative prediction = 43%
Kafonek et a149 1. Cognitively intact 22 77 DSM-111 Sensitivity = 79%
2. Dementiaorde- 47 77 14*** Specificity = 86%
lirium Positive prediction = 92%
Negative prediction = 66%
O'Connor et aP2 1. Cognitively intact 285 75+ 20 CAMDEX Sensitivity = 86%
2. Dementiaorde- 196 75+ 13 Specificity = 92%
lirium Positive prediction = 55
926 TOMBAUGH AND MCINTYRE 1AGS-SEPTEMBER 1992-VOL.40, NO. 9

TABLE 2. CONTINUED ~

Mean
Sample Groups n Age* Score Criteria Results**
Knopman and 1. Cognitivelyintact 55 74 29 NINCDS- Sensitivity = 54%
Ryberg7' 2. AD 28 74 23 ADRDA Specificity = 96%
Positive prediction = 88%
Negative prediction = 80%
Reed et a17' 1. AD 21 70 20 NINCDS- Sensitivity = 57%
ADRDA
Black et a17*
Comparison 1 1. Cognitively intact 80 70+ Clinical assess- Sensitivity = 100%
2. Dementia(proba- 31 70+ ment plus Specificity = 66%
ble & definite) AGECAT#
Comparison 2 1. Cognitively intact 80 70+ Clinical assess- Sensitivity = 89%
2. Dementia(proba- 47 70+ ment Specificity = 70%
ble (definite &
mild)
Galasko et alZ1 1. Cognitively intact 74 70 29 NINCDS- Sensitivity = 68%
2. AD 74 71 20*** ADRDA Specificity = 100%
Positive prediction = 100
Negative prediction = 76%
Jorm et alZ9 1. Cognitivelyintact 45 80 DSM 111 Sensitivity = 76%
2. Dementia 24 80 Specificity = 73%
Murden et alZ2 1. Cognitively intact 148 60-99 Research crite- Sensitivity = 96%
2. Dementia 110 ria Specificity = 81%
Positive prediction = 79%
Negative prediction = 97%
Neurological and
Psychiatric Sub-
jects
DePaulo and 1. Cognitively intact 26 44 28 Clinical assess- Sensitivity = 50%
F01stein~~ 2. Cerebrallesions 42 56 23 ment Specificity = 100%
Positive prediction = 100%
Negative prediction = 55%
Dick et a147 1. Cognitively intact 93 49 Clinical assess- Sensitivity = 76%
2. Neurological & 50 54 ment Specificity = 96%
cognitive impair- Positive prediction = 90%
ment Negative prediction = 88%
Lautenschlaeger et 1. Non-organic 64-92 Clinical assess- Sensitivity = 76%
a175 2. Organic 64-92 ment Specificity = 64%
Schwamm et a176 1. CNS lesion (ex- 30 54 22 Clinical assess- Sensitivity = 52%
cluded dementia ment
& delirium)
Chandler and 1. Mixed psychiatric 102 28 DSM-III& Sensitivity = 33%
Gemdt77 without cognitive NINCDS- Specificity = 91%
impairment or ADRDA Positive prediction = 31 %
depression Negative prediction = 92%
2. Mixed organic 12 53 24
mental disorder
with cognitive
impairment
Faustman et a17' 1. Mixed psychiatric 76 29 Luria-Nebraska Sensitivity = 21 %
without cognitive Neuro- Specificity = 96%
impairment Psychological Positive prediction = 50%
2. Mixedpsychiatric 14 27 Battery Negative prediction = 87%
with cognitive
impairment
Note: See text for explanation of sensitivity, specificity, positive prediction, and negative prediction.
Note: DSM 111 (Diagnostic and Statistical Manual of the American Psychiatric Association-Third Edition); NINCDS-ADRDA (National Institute of
Neurological and Communicative Disorders and Stroke-Alzheimer's Disease and Related Disorders Association).
* Single age score represents mean age. Pair of age scores represents range of ages.
** Prediction ratios are unadjusted for estimated prevalence rates.
*** Estimated by authors.
# AGECAT is a computer program for analyzing scores on the Geriatric Mental Status (GMS) exam.
JAGS-SEPTEMBER 1992-VOL. 40, NO. 9 MINI-MENTAL STATE EXAM 927
and specificity data were cited or could be derived from ificity when psychiatric patients are included in the
information in the article. Only studies that used the comparison group. Davous et a165reported 100% spec-
criterion score of 23 or less (23/24) are included. ificity when the control group consisted of patients
Sensitivity: Dementia Subjects Anthony et a145 with neurological or "functional" disorders, but only
were the first to employ the 23/24 cut-off score to 82% when psychiatric patients were used. A similar
determine the sensitivity of the MMSE. The cut-off trend was reported by Folstein et all8 in a community
criterion was based on data originally reported by survey. Specificity was 62% for elderly subjects with-
Folstein et all that suggested that a high, if not perfect, out a clinical diagnosis and 46% for a mixed group of
level of sensitivity would occur if the cut-off criterion community-dwelling subjects diagnosed as having
was set at 23/24. Anthony et a145selected 99 of 101 some type of a DSM-I11 condition. As discussed later,
consecutive admissions to a general medical ward who the demographic characteristics of the sample also
were classified as either cognitively impaired or intact affect specificity levels.
based on the presence/absence of delirium or dementia
as assessed by a psychiatrist using DSM-II173criteria. Correlation with Other Tests*
The MMSE correctly identified 20 of the 23 impaired
patients (87% sensitivity). Table 2 shows that similar The degree to which the MMSE is correlated with
levels of sensitivity have been reported in approxi- other tests measuring cognitive functioning provides
mately 75% of comparisons using dementia patients. evidence of construct validity. These correlations are
A related and perhaps more critical clinical question is reviewed below.
how well does a positive test score predict the presence Cognitive Screening Tests Correlations ranging
of dementia. Inspection of the positive prediction data from -0.66 to -0.93 were obtained from the
reveals that in approximately 70% of the studies, a studies27,28,48,50.65 that compared the MMSE with
MMSE score of less than 23 was associated with the either the original 26-item Blessed Information-Mem-
diagnosis of dementia in at least 79% of the cases. ory-Concentration test (BIMC)" or the shortened 6-
The major variable that appears to differentiate be- item Blessed Orientation-Memory-Concentration test
tween high and low MMSE sensitivity is the level of (BOMC).83The negative sign occurs because the MMSE
cognitive impairment in the dementia groups. The adds the number of correct answers while the Blessed
probability of obtaining high levels of sensitivity in- sums the number of errors. Given the high degree of
creases as impairment increases. For example, all stud- overlap in items, the high correlation is not unexpected.
ies with a mean MMSE score of 15 or less for the Additional research has shown that correlations gen-
demented subjects report relatively high levels of sen- erally falling within the .70 to .90 range exist between
sitivity. The two studies with a mean MMSE score MMSE scores and those obtained from a representative
greater than 2067,70 for the impaired group reported sampie of other cognitive screening tests administered
low levels of sensitivity (44% and 68%). In addition, to a variety of different types of subjects.29,42-44,
higher sensitivity has been reported with hospital or Intelligence and Memory Tests Since the MMSE
clinic samples relative to community surveys. This oripally was designed to assess the construct of gen-
trend probably reflects the over-representation of mod- eral cognitive ability, Folstein et al' compared MMSE
erate and severe cognitive impairment compared to scores to those obtained on the Wechsler Adult Intel-
community samples. ligence Scale.94They found a correlation of .78 with
Sensitivity: Neurological and Psychiatric Subjects the Verbal Scale and .66 with the Performance Scale.
As shown in Table 2, the sensitivity of the MMSE for Comparable findings have been reported in several
general neurology and psychiatry patients usually is other studies using a variety of different types of
low, ranging from 21% to 76%. Two reasons are fre- subje~ts.25,47~60,78,95-97 The failure of some studies to
quently cited for this. Probably because of its bias obtain significant correlation^^^, 96 occurred where the
toward verbal items, the MMSE is relatively insensitive average MMSE score was 27 or above and probably
to damage in the right hemisphere, causing an increase reflects a range restriction due to the truncated distri-
in false negative^.^^,^^, 79,80 As well, the language items bution of scores. Moderate-to-high correlations were
are too simple to detect mild impairments.21,35, 47, 76,81 also obtained with the Wechsler Memory S~ale.9~. 97
In addition, according to Chandler and G e ~ m d the t~~ Neuropsychological Tests Modest-to-high corre-
heterogeneity of neurological patients makes it difficult lations between scores on the MMSE and those ob-
to identify cognitive impairment. It should be noted tained on various cognitive tests (eg, Trails B, WMS,
that a high degree of variability also exists in the digit span, story recall, word list recall) used in
positive predictive values, with scores ranging from neuropsychological assessments have been re-
31% to 100%. ported.37* 46, 78, 87, 95, 98, 99 These results are consistent

Specificity Most studies report moderate-to-high with a report by Morris et a15' that factor analysis of
levels of specificity, indicating the members of the several neuropsychological tests loaded on three dif-
control group are readily identified by a score greater ferent factors, and the MMSE loaded equally on aII
than 23. The negative predictive value of the MMSE is factors.
also relatively high due to the low number of false
negative cases. Table 2 shows that the composition of
the control group is an im ortant factor in determining
J2
specificity. Two studies", have shown lowered spec- * A summary table containing the correlations between MMSE scores and
those obtained on other tests is available on request.
928 TOMBAUGH AND MCINTYRE jAGS-SEPTEMBER 1992-VOL. 40, NO,9

Activities of Daily Living (ADL) Measures The functioning that occurs with AD. Several studies have
relationship between cognitive deficits and ability to reported that scales evaluating functional competence,
function independently is critical in dementia. Impaired such as ADL measures, assess a wider range of func-
occupational or social function has been shown to tions than the MMSE and are more appropriate for
covary with severity levels, and the degree of impair- longitudinal studies involving severely demented pa-
ment may serve as a criterion for the diagnosis of tients.19,80,100,121,122
dementia (eg, DSM-111). Thus, MMSE scores should
correlate with measures of functional capacity, such as Other Evidence of Construct Validity
those obtained with the Blessed Dementia Rating Scale
(BDRS),82that assess an individual’s everyday activi- MMSE scores also correlate with other measures that
ties, habits, and personality. Correlations between reflect severity of AD, thus providing additional evi-
MMSE scores and those obtained from ADL scales dence of construct validity. These include urinary
generally range from .40 to .75, indicating that lower inc~ntinence,’~~, 123 -ma~-tality,~~~,lZ4* 125 changing
MMSE scores are related to decreased indepen- health status,’05abnormal behavioral change,1o4*118, 126, 127
dence.19,49.50, 100-107 The ECA Piedmont Health hearing impairment,56* I2O length of time in hospi-

Survey107provides the most extensive ADL data. tal,128-130and extrapyramidal signs.131However, some
MMSE scores from 1,637 community-dwellingindivid- studies have shown that neuropsychologicalmeasures,
uals were correlated (.48) with instrumental activities such as drawing a 3-dimensional cube, were related to
(eg, cooking, caring for finances) but not with physical behaviors such as wandering and urinary incontinence,
activities (eg, dressing, eating). The higher correlation which were not correlated with MMSE scores.98’132
obtained with instrumental, rather than h sical, ADLs
has been reported in several st~dies.6~. p,3x08p Thus, Influence of Demographic and Social Variables
scores on the MMSE are sensitive to a decline in more Educational Level MMSE scores repeatedly have
cognitively demanding functional behavior that is in- been shown to be related to educational attainment.
dependent of physical health and mobility. This association between years of education and MMSE
Finally, since the BDRS originally was validated performance has been reported in studies using hos-
against postmortem neuropathological changes,82, pital patients,47a mixed sample of dementia and cog-
the high degree of relationship between the MMSE nitively intact subjects,133individuals randomly sam-
and BDRS provides indirect evidence that MMSE pled from the community,2.6.16,17,30,31.36,40,107,128,134
scores are correlated to histopathological findings. This and sub’ects screened to exclude delirium and demen-
speculation is supported by the -.70 correlation that tia.17*”, ‘08, 135 The importance of education was re-
existed between MMSE scores and plaque counts.”’ vealed by regression analyses as well, showing that
education accounted for more variance than other de-
mographic variables including gender, race, and social
Longitudinal Studies class.’, 133, 134 In addition, education levels affected
Longitudinal studies with Alzheimer’s patients pro- the distribution of errors across individual items and
vide additional evidence of construct validity. Since categories of items.22. 31.36.45,133,134,136 some studies,
AD is a progressive disease where cognitive func- however, have failed to fiid such a relation.”, 60*90, 137
243

tions decline over time, MMSE scores should decline This may be due, at least in part, to sampling biases
with serial testing. Longitudinal studies using test- caused by an overrepresentation of individuals with
retest intervals ranging from 1 month to 3 years show either high education or severe cognitive impairment,
that MMSE scores for dementia patients, the major- the latter causing a restricted range of very low MMSE
ity classified as AD, significantly declined over scores.
time.50,52,56,58.71,80,92,112-120 Although the rate of de- A central issue emerging from these results is
cline varied between and within studies, it generally whether the effects attributed to education represent a
fell between 2 and 5 points per year. Moreover, the measurement error or a risk factor. The prevalent view
rate of decline in 1 year was not correlated with that is that education introduces a psychometric bias lead-
occurring during the following year.” ing to a misclassification of individuals from different
A substantial degree of variability occurred in those educational backgrounds, and this bias should be cor-
studies employing Alzheimer’s patients. While some of rected by employing norms stratified for education.
the test-retest variability is related to the uneven pro- This position assumes that education reflects a stable
gression of AD and the heterogeneity of subjects pro- characteristic that is not associated with any type of
duced by differences in the duration of illness, age of underlying pathology and, as such, does not constitute
onset, and subclass of AD, the content and psycho- a risk factor. To a large degree, this view is based on
metric properties of the MMSE are contributing factors evidence showing that low educational levels increase
as well. The verbal items (eg, recall of three words, 7s/ the likelihood of misclassifying normal subjects as cog-
WORLD) that make the MMSE sensitive to the pro- nitively impaired (ie, false positives). This is particularly
found decline in memory that occurs in moderately evident when subjects have fewer than 9 years of
demented patientsso.112,118,121 lose their discriminabil- education.22* 45 Higher education levels also may pro-
ity as the severity of the illness increases. Thus, as the duce classification errors. Fillenbaum et al.lo7specu-
lower limits of the scale are approached, the MMSE lated that higher education levels may mask mild im-
becomes less sensitive to the progressive decline of pairment, and O’Connor et aP2 found that all dementia
JAGS-SEPTEMBER 1992-VOL.40, NO. 9 MINI-MENTAL STATE EXAM 929
patients with an MMSE score of 24 or greater (ie, false been reported in community surveys employing ran-
negative) had relatively high levels of education. Thus, dom sampling procedures,zP18,313 34,36,40,107,134 studies
evidence from cross-sectional studies shows that num- in which subjects did not suffer from dementia, delir-
ber of years of education affects both sensitivity and ium, or depre~sion,’~,45, 51, 137 and studies that tested
“
3

specificity. hospital or clinic patients.I7*35, Most of this age-

Results from longitudinal studies provide further related change begins about age 55 or 60 and then
evidence that while education is associated with low dramatically accelerates over the age of 75 or 80.16,
MMSE scores, it is not associated with either the diag- These age effects persist when subjects are stratified
nosis of AD or the rate of cognitive decline that occurs by education level, 6, 31, demonstrating that the age
as the severity of AD increases. Three studies”, 13‘, 13’ effect is not simply due to cohort differences in edu-
have reported that although educational levels and cational attainment.
MMSE scores were correlated, number of years of The finding that older adults tend to have lower
education was not related to the diagnosis of dementia. MMSE scores than younger adults suggests that fixed,
This is consistent with other reports showin that age-independent cut-off scores (eg, 23/24) may under-
education is not associated with dementia.’39, Ad- estimate cognitive impairment with younger adults or
ditionally, several longitudinal studies have reported over-estimate it with older adults. Evidence showing
that education was not associated with the rate of that dementia frequently is overestimated in older nor-
11’ Similar effects have been reported from mals supports this notion.18,45, However, since
studies using measures other than the MMSE to assess age generally is viewed as a risk factor for de-
rate of cognitive change in AD patient^.^^*-"^ While mentia,134,146.147 all age effects cannot be merely dis-
these results do not conclusively demonstrate that missed as representing only psychometric bias. The
education solely reflects a psychometric bias, they matter is further complicated by the finding from lon-
do suggest that substantial potential for misclassifica- gitudinal studies that rate of cognitive deterioration in
tion exists, prompting several authors to recom- Alzheimer’s patients appears to be independent of
mend that education levels, particularly 8 vears or age.’”
less, be considered when- interpriting ‘ MMSE Conflicting evidence exists concerning the degree to
scores.31,36, 40, 45, 51, 79, 107, 133, 144, 145 which age interacts with education. While some
Although little doubt exists that educational levels studiesl6.31report that age-related impairment in scores
lead to potential misclassification, several authors cau- is greater for subjects with lower educational levels,
tion that education may not exclusively represent a othersI7 conclude that declines for age and education
detection bias but also may reflect etiologic factors are independent. In addition, several studies have re-
critical in a process eventually resulting in de- ported that age and education do not uniformly affect
mentia.17,134,146-148 Lower education levels could, for all iterns.’, 16, 17, 24, 31. 36, 45, 134 These and other results
example, contribute to the incidence of dementia be- suggest that the association between total MMSE per-
cause education is associated with various biological formance, age, and education is complex, but never-
risk factors typically associated with multi-infarct de- theless one that should be considered in interpreting
mentia, such as hypertension, obesity, and serum cho- MMSE performance.
lesterol. Moreover, the failure of education to be etio-
logically related to one type of cognitive impairment,
Gender In a summary report containing data from
say Alzheimer’s disease, does not necessarily mean approximately 20,000 individuals participating in the
that education will be unrelated to other types of ECA study, George et algOconcluded that no meaning-
cognitive impairments, such as vascular dementia. ful gender differences existed in the prevalence of
Thus, modification of the cut-off score to compensate cognitive impairment. Other studies indicate that
for educational biases in one sample does not justify even when differences were statistically significant,
using the same criterion for another sample, unless they did not account for a substantial portion of
justified by empirical results for that sample. variance and enerally were not substantively im-
portant.16-18,24,$5, 51, 107, 118, 150
BerkmanIg6suggests that adjusting for education level
at the time of initial screening decreases, if not elimi- Race/Ethnicity and Social Class The most widely
nates, the possibility of exploring the hypothesis that cited study on the effects of race/ethnicity on MMSE
education constitutes a risk factor for dementia. Fur- performance analyzed interview responses from over
thermore, it reduces the possibility of investigating the 3,000 English and Hispanic residents of Los A n g e l e ~ . ~ ~
relationship between education and cognitive impair- Both English and Spanish versions of the MMSE were
ment caused by factors other than dementing illness. used. Analysis of individual items revealed that His-
The importance of Berkman’s warning is underscored panics performed significantly lower on many items.
by preliminary results from the East Boston study A more recent report of the results from all ECA
suggesting that rate of AD may be higher among centers“ indicates that race/ethnicity exerts a signifi-
individuals with few years of schooling.’47 Thus, it cant effect on the distribution of MMSE scores. The
appears likely that education represents both a psycho- finding that racial effects tend to be maintained within
metric bias and a risk factor, differing only in degree different educational levels suggests that education
(for further discussion see Refs. 30, 40, 144, 146-148). cannot explain all race/ethnicity effects. However,
Age Numerous studies have shown that MMSE Murden et a1” reported contradictory findings that
scores decrease as age increases. This relationship has MMSE performance was not affected by race. Effects
930 TOMBAUGH AND MCINTYRE IAGS-SEPTEMBER 1992-VOL.40, NO. 9

of social class and socioeconomic status on MMSE contained in each factor varied among the studies.
133
scores also have been ob~erved.’~,
34, Nevertheless, the results from the studies are important
since they show that the set of cognitive domains
ANALYSIS OF INDIVIDUAL ITEMS measured by the MMSE is certainly less than the seven
Various analyses, including frequency distributions categories into which the questions usually are
of the errors, part-total correlations, item analyses, and grouped.
stepwise regression analyses, have been employed to Finally, studies have determined if serial 7s and
investigate the relationship between individual items reverse spelling of WORLD represent equivalent tasks.
and total MMSE score. Analyses of individual items The overwhelming weight of the data shows they are
from studies containing normal subjects (ie, large com- not comparable. Spelling WORLD backward consist-
munity surveys or ones in which participants were ently produces higher scores than does counting back-
screened to eliminate obvious cases of dementia and ward by sevens.16,19,21,33,36,37,45,137,155-158 Moreover
delirium) revealed that most errors generally occurred Holzer et all6 reported a correlation between serial 7s
in only four of the seven cognitive domains: recall of and WORLD of only .37, with WORLD having a higher
three words, serial 7s/WORLD, pentagon, and orien- correlation with the total score than serial 7s (.47 vs
tation to time.2, 16, 21,24,31.34,40.137 Although the relative .39). Finally, studies on the effects of age, education,
contribution of each domain varied from study to and gender on performance on these two items have
been inconsistent.22.32,36, 45, 155. 156
study, recall of three words usually produced the great-
est number of errors. Errors rarely occurred for orien-
tation to place, registration, and individual language SUGGESTED MODIFICATIONS TO THE MMSE
questions. Several attempts have been made to improve the
Analyses of individual items for Alzheimer’s sensitivity and specificity of the MMSE. Some studies
patients again indicated that the same four cognitive have explored altering the cut-off score, but without
domains consistently produced the greatest number of much success. In general, changing the cut-off points
errors.2,19, 21, 27, 28, 135, The relative difficulty of these alters both the sensitivity and specificity of the test,
items vaned across studies, with the exception that the increasing one while decreasing the other.21,32, 34, 45, 72, 77
greatest percentage of errors generally occurred for A second approach has been to compensate in
recall of three words. The failure to observe a consistent some manner for various ages and/or educational
rank order of difficulty for the other three domains is levels.2,22, 45, 51, 133, 145, ls9 One strategy is to assume
243

attributable, at least in part, to differences in dementia that the MMSE is valid only if the person has 9 or
severity that existed between studies. Although there more years of schooling.22,lo5*133, 145 Another tack is to
are reports that orientation to place is highly predictive generate normative data, stratified for age and/or ed-
of total MMSE scores,21this finding is not consistent u c a t i ~ n . ~Although
~, using different cut-off scores for
and may depend on whether subjects were tested in a various ages and educational levels has its merits, it
familiar (eg, home) or unfamiliar (eg, hospital) environ- also suffers the same problem noted above-attempts
ment.37,56,57 to increase sensitivity usually decrease specificity.
We are aware of only two studies that directly com- A third avenue is using multivariate proce-
pared the performance of normals and Alzheimer’s dures to differentially weight existing MMSE
patients on each item. Brayne and C a l l ~ w a yreported
’~~ items.2,16.21,31,36,38,137,144,160 For example, Cullum et

that “normal”subjects scored significantly higher than al,137using a stepwise regression analysis with highly
dementia patients on all MMSE items except “no ifs, educated normal subjects aged 50 to 80, reported that
and, or buts” and naming two objects (watch, pencil). recall of three words and orientation to time correlated
Galasko et a1,” employing a paired-comparison pro- .87 with total score. Galasko et al,” employing a logis-
cedure for different age and education levels, compared tic-regression model with AD patients, found that the
scores from healthy controls with those obtained by sum of the scores for recall of three words and orien-
Alzheimer’s patients. As expected, the differences be- tation to place resulted in sensitivity and specificity
tween the “normal” and inoderate-to-severe groups levels that were similar to those produced by total
increased on the same items that had previously best MMSE scores. Magaziner et a131 generated a series of
discriminated between mild AD and normals. In addi- prediction equations for different age groups and
tion, a significant difference occurred for the language educational levels. However, results from several
items. Thus, it appears that although the language studies make the generality of these equations ques-
items are useful in distinguishing between normals and ti~nable.’~l-’~~
Alzheimer’s patients, they are much less discriminating A fourth alternative is to modify the content of the
than are the other four categories and are most sensitive MMSE.~, 21,22, 36,45,81, 144 0
ne way to modify the MMSE
to individual differences among patients with moder- would be to exclude questions particularly sensitive to
ate-to-severe AD. This finding is consistent with pre- age, education, and culture or to add questions less
vious findings showing that items are differentially sensitive to these demographic variables. Another ap-
sensitive to disease severity.”, ‘19, 152-154 proach would involve eliminating items with little di-

‘
Three studies were undertaken to determine if the
seven rationally derived cognitive cate ories could
be validated through factor analysis.27, l9 Although
each study yielded a two factor solution, the items
agnostic utility and/or adding items known to be sen-
sitive to cognitive impairment. The need for more
adequate language items, for example, has been men-
tioned by several 47 In some instances, sup-
1AGS-SEPTEMBER 1992-VOL. 40, NO.9 MINI-MENTAL STATE EXAM 931

plementary items have been employed. For example, specific types of cognitive abilities, and activities of
Galasko et alZ1found that adding a word fluency task daily living. Other evidence shows the MMSE to be
decreased error rates and increased the sensitivity of correlated with AD pathology. Longitudinal research
the MMSE from 79.2% to 87.5% for mild AD. Mayeux with dementia patients illustrates its ability to serially
et al" created a modified version of the MMSE document cognitive change.
(mMMS) by adding digit span, recall of four US presi- However, the MMSE is not without problems. Per-
dents, confrontational naming, sentence for repetition, haps the most frequently cited shortcoming relates to
and copy of two additional designs. The mMMS cor- its lack of sensitivity to mild cognitive impairment and
related .89 with the MMSE, has a test-retest reliability its failure to adequately discriminate patients with mild
of .95,'64 and has been employed in several dementia AD from normal patients. Attempts to improve the
studies.81,103,131,165 sensitivity, including altering the cut-off scores, differ-
The most extensive revision of the MMSE was un- entially weighting existing items, modifying the con-
dertaken by Teng et al.'35 The Modified Mini-Mental tent of the items, or using supplementary items, have
State Examination (3MS) maintains the MMSE's basic met with mixed success. The MMSE also has received
format while extensively modifying its content. It con- its share of criticism because of its insensitivity to
tains four new test items, an expanded range of scores progressive changes occurring with severe Alzheimer's
(0-100 rather than 0-30), and modified scoring pro- disease. Moreover, inconsistencies in the way it is
cedures allowing assignment of partial credit on some administered, scored, and interpreted make cross-study
items. Subsequent research5' demonstrated that the comparison difficult. The content of the MMSE is
3MS possesses higher reliability and validity than the highly verbal, lacking sufficient items to adequately
MMSE. Additional information on the psychometric measure visuospatial and/or constructional praxis.
properties of the 3MS should be forthcoming since it Hence, its utility in detecting impairment caused by
currently is being used to assess cognitive impairment focal lesions, particularly those residing in the right
in the Canadian Study on Health and Aging,'66 an hemisphere, is uncertain. Items designed to measure
epidemiological survey sponsored by the Health and language functions also tend to be overly simplistic
Welfare Canada. and tend to be insensitive to mild linguistic deficits,
A fifth alternative has been to include the MMSE as hence increasing the number of false negative errors.
part of a battery of tests. For example, Pfeffer et Some of the shortcomings may have occurred be-
developed the Mental Function Index (MFI) by using a cause the MMSE, in its current form, is expected to
discriminant function analysis to obtain weighted provide too many different types of screening func-
scores for the MMSE, the Symbol Digit Modalities tions. It may be unrealistic to expect a single version of
Tests,I6*and the Raven Subtest B.'69 The combined the test to meet all of these demands. Different types
sum of the weighted scores yielded 93% sensitivity for of screening applications (eg, AD) may require different
demented patients and 80% specificity for normals. versions of the MMSE. Thus, it also may be more
Similar values were obtained in a follow-up efficient to employ one set of cut-off scores for a
However, substantially lower levels of sensitivity have hospital-based geriatric clinic and a different set for
been reported.26T 70 In addition, Mowry and Burvi1lZ6 epidemiological surveys. Since altering cut-off scores
noted that high refusal rates and administrative time typically increases either the sensitivity or specificity
also contraindicate the MFI's use as a screening device. while the other decreases, the selection of the specific
criterion score may depend on a cost-benefit analysis
SUMMARY AND RECOMMENDATIONS to assess the relative importance of increased numbers
The MMSE was developed as a screening test to of false positives or false negatives.
quantitatively assess the severity of cognitive impair- In view of the above, we feel the following recom-
ments and to document cognitive changes that occur mendations are warranted when the MMSE is em-
over time. The research reviewed over the past 26 ployed in a clinical setting.
years indicates, to a large degree, that the MMSE has (1)The MMSE should be used as a screening device
been able to fulfill these goals. Examination of its for cognitive impairment or a diagnostic adjunct in
psychometric properties shows moderate-to-high lev- which a low score indicates the need for further eval-
els of reliability, with test-retest reliability higher than uation. It should not serve as the sole criterion for
measures of internal consistency. Items measuring re- diagnosing dementia or to differentiate between var-
call of three words, copy pentagon, 7s/WORLD, and ious forms of dementia. However, MMSE scores may
orientation to time appear to be the most sensitive to be used to classify the severity of cognitive impairment
both normal aging and dementing illnesses. The or to document serial change in dementia patients.
MMSE, like many measures of cognitive ability, is (2) The following three cut-off levels should be em-
affected by demographic factors. Of these, age and ployed to classify the severity of cognitive impairment:
education exert the greatest effect. Criterion validity no cognitive impairment = 24-30; mild cognitive im-
measures show high levels of sensitivity for moderate- pairment = 18-23; severe cognitive impairment = 0-
to-severe levels of dementia. Construct validation stud- 17.
ies demonstrate that MMSE scores correlate highly (3) The MMSE should not be used cIinicalIy unIess
with those obtained from other types of cognitive the person has at least a grade eight education and is
screening tests as well as from psychological and neu- fluent in English. While this recommendation does not
ropsychological tests measuring intelligence, memory, discount the possibility that future research may show
932 TOMBAUGH AND MCINTYRE JAGS-SEPTEMBER 1992-VOL. 40, NO.9

that number of years of education constitutes a risk 15. McKhann G, Drachman D, Folstein M, et al. Clinical diagnosis of Alz-
factor for dementia, it does acknowledge the weight of heimer’s disease: Report of the NINCDS-ADRDA Work Group under
the auspices of Department of Health and Human Services Task Force
evidence showing that low educational levels substan- on Alzheimer’s disease. Neurology 1984;34:939-944.
tially increase the likelihood of misclassifying normal 16. Holzer CE 111, Tischler GL, Leaf PJ et al. An epidemiologicassessment of
subjects as cognitively impaired. cognitive impairment in a community population. In: Greenley JR, ed.
Research in Community Mental Health, Vol 4. London England: JAI
(4)Serial 7s and WORLD should not be considered Press, 1984, pp 3-32.
equivalent items. Both items should be administered 17. OConnor DW, Pollitt PA, Treasure FP et al. The influence of education,
social class and sex on Mini-Mental State scores. Psychol Med
and the higher of the two should be used. In scoring 1989;19:771-776.
serial 7s, each number must be independently com- 18. Folstein MF, Anthony JC, Parhad J, et al. The meaning of cognitive
pared to the prior number to insure that a single impairment in the elderly. J Am Geriatr SOC1985;33:228-235.
19. Ashford JW, Kolm P, Colliver JA et al. Alzheimer patient evaluation and
mistake is not unduly penalized. WORLD should be the Mini-Mental State: Item characteristic curve analysis. J Gerontol
spelled forward (and corrected) prior to spelling it 1989;44:P139-146.
20. Beatty W, Goodkin DE. Screening for cognitive impairment in multiple
backward. The scoring procedures employed by either sclerosis: An examination of the Mini-Mental State Examination. Arch
Morris et all3 or Teng et alZ3are recommended. Neurol 1990;47297-301.
(5) The words apple, penny, and table should be 21. Galasko D, Klauber MR, Hofstetter CR et al. The Mini Mental State
Examination in the early diagnosis of Alzheimer‘s dementia. Arch Neurol
used for registration and recall. If necessary, the words 1990;4749-52.
may be administered up to three times in order to 22. Murden RA, McRae TD, Kaner S et al. Mini-Mental State Exam scores
obtain perfect registration, but the score is based on vary with education in blacks and whites. J Am Geriatr SOC1991;39:149-
155.
the first trial. 23. Teng EL, Chiu HC. The Modified Mini-Mental State (3MS) examination.
( 6 ) The “county”and “where are you” orientation to J Clin Psychiatry 1987;48:314-318.
place questions should be modified. The name of the 24. Bleecker ML, Bolla-Wilson K, Kawas C et al. Age-specific norms for the
Mini-Mental State Exam. Neurology 1988;33:1565-1568.
county where a person lives should be asked rather 25. Farber JF, Schmitt FA, Logue PE. Predicting intellectual level from the
than the name of the county where the testing site Mini-Mental State Examination. J Am Geriatr Soc 1988;36:509-510.
26. Mowry BJ, BurvilL BW. A study of mild dementia in the community
resides, and the name of the street where the individual using a wide range of diagnostic criteria. Br J Psychiatry 1988;153:328-
lives should be asked rather than the name of the floor 334.
where the testing is taking place. 27. Fillenbaum GG, Heyman A, Wilkinson WE et al. Comparison of two
screening tests in Alzheimer’s disease. Arch Neurol 1987;44:924-927.
28. Zillmer EA, Fowler PC, Gutnick HN et al. Comparison of two cognitive
ACKNOWLEDGMENTS bedside screening instruments in nursing home residents: A factor ana-
lytic study. J Gerontol 1990;45:69-74.
We wish to thank Drs. Bryan A. Bernard, William G. 29. JormA, Scott R, Cullen JS et al. Performance of the Informant Question-
Snow, and Robert S. Wilson for the helpful suggestions naire on Cognitive Decline in the EIdrly (IQCODE) as a screening test
made on an earlier version of the article. for dementia. Psychol Med 1991;21:785-790.
30. Jorm AF, Scott R, Henderson AS et al. Educational level differences on
the Mini-Mental State: The role of test bias. Psychol Med 1988;18:727-
REFERENCES 731.
31. Magaziner J, Bassett SS, Hebel JR. Predicting performance on the mini-
1. Folstein MF, Folstein SE, McHugh PR. ‘Mini-mental State’. A practical mental state examination: Use of age- and educational-specificequations.
method for grading the cognitive state of patients for the clinician. J J Am Geriatr SOC1987;35:996-1000.
Psychiatr Res 1975;12:189-198. 32. O’Connor DW, Pollitt PA, Hyde JB et al. The reliability and validity of
2. Bird HR, Canino G, Stipec MR et al. Use of the Mini-Mental State the Mini-Mental State in a British community survey. J Psychiatr Res
Examination in a probability sample of Hispanic population. J Nerv Ment 1989;23:87-96.
Dis 1987;175:731-737. 33. Olin IT, Zelinskin M. The 12-month reliabiltiy of the Mini-Mental State
3. Engedal K, Haugen P, Gilje K, Laake P. Efficacy of short mental tests in Examination. Psychol Assessment 1991;3:427-432.
the detection of mental impairment in old age. Compr Gerontol 34. Kay DWK, Henderson AS, Scott R et al. Dementia and depression among
1988;2:87-93. the elderly living in the Hobart community: The effect of the diagnostic
4. Heeren TJ, Lagaay AM, Beek WAC et al. Reference values for the Mini- criteria on the prevalence rates. Psychol Med 1985;15:771-788.
Mental State Examination in octo- and nonagenarians. J Am Geriatr SOC 35. Tsai L, Tsuang MT. The Mini-Mental State Test and computerized
1990;38:1093-1096. tomography. Am J Psychiatry 1979;136436-439.
5. Kamo M, Audrey MA, Escobar JI, et al. Development of the Spanish- 36. Escobar J, Bumam A, Kamo M et al. Use of the Mini-Mental Status
language version of the National Institute of Mental Health Diagnostic Examination (MMSE) in a community population of mixed ethnicity:
Interview Schedule. Arch Gen Psychiatry 1983;40:1183-1188. Cultural and linguistic artifacts. J Nerv Ment Dis 1986;174:607-614.
6. Katzman R, Zhang M, Ya-Qu 0 et al. A Chinese version of the Mini- 37. Livingston G,Sax K, Willison J et al. The Gospel Oak study stage I 1 the
Mental State Examination: Impact of illiteracy in a Shanghai dementia diagnosis of dementia in the community. Psychol Med 1990;20881-891.
survey. J Clin Epidemiol 1988;41:971-978. 38. Klein LE, Roca RP, McAuthur J et al. Diagnosing dementia: Univariate
7. Park JH, Ha JC. Cognivitive impairment among the elderly in a Korean and multivariateanalyses of the Mental Status Examination.J Am Geriatr
rural community. Acta Psychiatr Scand 1988;7752-57. SOC1985;33:483-488.
8. Rocca WA, Bonaiuto S, Lippi A, et al. Prevalence of clinically diagnosed 39. Fillenbaum GG, George LK, Blazer DG. Scoring nonresponse on the
Alzheimer’s disease and other dementing disorders: A door-to-door Mini-Mental State Examination. Psychol Med 1988;18:1021-1025.
survey in Appignano, Macerata Province, Italy. Neurology 1990;40626- 40. George LK, Landerman R, Blazer -bGet al. Cognitive impairment. In:
631. Robbins LN, Regier DA, eds. Psychiatric Disorders in America. New
9. Salmon D, Riekkinen PJ, Katzman R, et al. Cross-cultural studies of York Free Press, 1991, pp 291-327.
dementia. Arch Neurol1989;46:769-772. 41. Franssen EH, Reisberg B, Kluger A et al. Cognition-independent neuro-
10. National Institute of Mental Health. The Diagnostic Interview Schedule. logic symptoms in normal aging and probable Alzheimer’s disease. Arch
Washington, DC: NlMH Center for Epidemiological studies, 1979. Neurol 1991;48:148-154.
11. Robins LN, Regier DA, eds. Psychiatric Disorders in America: The Epi- 42. Reisberg B, Ferris SH, de Leon MJ et al. Age-associated cognitive decline
demiologic Catchment Area Study. New York The Free Press, 1991. and Alzheimer’s disease: Implications for assessment and treatment. In:
12. Roth M, Tyme E, Mountjoy Q et al. CAMDEX A standardized instrument Bergener M, Ennini M, Stahelin HB, eds. Thresholds in Aging. New
for the diagnosis of mental disorder in the elderly with special reference York Academic Press, 1985.
to the early detection of dementia. Br J Psychiatry 1986;149:698-709. 43. Reisberg B, Ferris SH, Anand R et al. Functional staging of dementia of
13. Moms JC, Mohs RC, Rogers H et al. Consortium to establish a registry the Alzheimer type. Ann NY Acad Sci 1984;35:481-483.
for Alzheimer’s disease (CERAD) clinical and neurological assessment of 44. Foreman MD. Reliability and validity of mental status questionnaires in
Alzheimer‘s disease. Psychopharmacol Bull 1988;24:641-652. elderly hospitalized patients. Nurs Res 1987;36:216-220.
14. Zaudig M, Mittelhammer J, Hiller W et al. SIDAM-A structure interview 45. Anthony JC, LaResche L, Niaz U et al. Limits of the ’Mini-Mental State’
for the diagnosis of dementia of the Alzheimer type, multiinfarct demen- as a screening test for dementia and delirium among hospital patients.
tia and dementias of other aetiology according to ICD-10 and DSM-III- Psychol Med 1982;12:397-408.
R. Psychol Med 1991;21:225-236. 46 Pfeffer RI, Kurosaki TT, Chance JM et al. Use of the Mental Function
IAGS-SEPTEMBER 1992-VOL. 40, NO. 9 MINI-MENTAL STATE EXAM 933
Index in older adults. Am J Epidemiol 1984;120:922-935. in a psychiatric sample. Acta Psychiatr Scand 1990;81:126-131.
47. Dick JPR, Guiloff RJ, Stewart A et al. Mini-mental state examination in 79. Nelson A, Fogel BS. Faust D. Bedside cognitive screening instruments: A
neurological patients. J Neurol Neurosurg Psychiatry 1984;47496-499. critical assessment. J Nerv Ment Dis 1986;174:73-82.
48. Thal LJ, Grundman M, Golden R. Alzheimer’s disease: A correlational 80. Salmon DP, Thal LJ, Butters N et al. Longitudinalevaluation of dementia
analysis of the Blessed Information-Memory-Concentration Test and the of the Alzheimer’s type: A comparison of 3 standardized mental status
Mini-Mental State Exam. Neurology 1986;36:262-264. examinations. Neurology 1990;401225-1230.
49. Kafonek S, Ettinger WH, Roca R et al. Instruments for screening for 81. Mayeux R, Stem Y, Rosen J et al. Depression, intellectual impairment
depression and dementia in a long-term care facility. J Am Geriatr SOC and Parkinson‘s disease. Neurology 1981;31:645-650.
1989;3729-34. 82. Blessed G, Tomlinson BE, Roth M. The association between quantiative
50. Moms JC, Heyman A, Mohs RC, et al. The Consortium to Establish a measures of dementia and of senile change in the cerebral grey matter
Regishy for Alzheimer’s Disase (CERAD). Part I. Clinical and neuropsy- of elderly subjects. Br J Psychiatry 1968;114:797-811.
chological assessment of Alzheimer’s disease. Neurology 1989;39:1159- 83. Katzman R, Brown T, Fuld P et al. Validation of a short orientation-
1165. memory-concentration test of cognitive impairment. Am J Psychiatry
51. Teng EL, Chui HC, Gong A. Comparisons between the Mini-Mental 1983;140:734-739.
State Exam (MMSE) and its modified version-the 3MS test. Psychoger- 84. Blessed G, Black SE, Butler T et al. The diagnosis of dementia in the
iatrics: Biomedical and Social Advances. Tokyo: Excerpta Medica, 1990, elderly. Br J Psychiatry 1991;159:193-198.
pp 189-192. 85. Block Rl, Devoe M, Russell M et al. Clinical ratings: Relationship to
52. van Belle G, Uhlmann RF, Hughes JP et al. Reliabdtiy of estimates of objective psychometric assessment in individuals with dementia. Psychol
changes in mental status test performance in senile dementia of the Rep 1985557 183-189.
Alzheimer type. J Clin Epidemiol 1990;43:589-595. 86. Burch EA, Andrews SR. Cognitive dysfunction in psychiatricconsultation
53. Lesher EL, Whelihan WM. Reliability of mental status intruments admin- subgroups: Use of two screening tests. South Med J 1987;80:1079-1082.
istered to nursing home residents. J Consult Clin Psychol 1986;54:726- 87. Horn L, Cohen Cl, Teresi J. The EASI: A self-administered screening test
727. for cognitive impairment in the elderly. J Am Geriatr SOC1989;37848-
54. Marascuilo LA, Serlin RC. Statistical Methods for the Social Sciences. 855.
New York WH Freeman and Company, 1988, pp 87-91. 88. JormAF, Scott R, JacombPA. Assessment of cognitive decline in dementia
55. Keating JJ Ill. “Studying”for the Mini-Mental Status Exam. J Am Geriah by informant questionnaire. Int J Geriatr Psychiatry 1989;4:35-39.
SOC1987;35:594-597. 89. Kokmen E, Smith GE, Peterson RC et al. The Short Test of Mental Status.
56. Peters CA, Potter JF, Scholer SG. Hearing impairment as a predictor of Arch Neurol 1991;48:725-728.
cognitive decline in dementia. J Am Geriatr SOC1988;36:981-986. 90. Pearson JL, Chemer M, Tem L. The Mini-Mental State Exam and the
57. Ward HW, Ramsdell JW, Jackson JE, et al. Cognitive function testing in Mental Status Questionnaire: Depression in Alzheimer’s patients. Clin
comprehensive geriatric assessment. J Am Geriatr SOC1990;38:1088- Gerontol 1989;8:31-37.
1092. 91. Strain JJ, Fulop D, Lebovits A et al. Screening devices for diminished
58. Becker JT, Huff J, Nebes RD et al. Neuropsychological function in cognitive capacity. Gen Hosp Psychiatry 1988;10:16-23.
Alzheimer’s Disease: Pattern of impairment and rates of progression. 92. Teitelbaum L, Ginsburg L, Hopkins RW. Cognitive and behavioral im-
Arch Neurol 1988;45:263-268. pairment among elderly people in institutions providing different levels
59. Lopez OL, Swihart AA, Becker JT et al. Reliabiltiy of NINCDS-ADRDA of care. Can Med Assoc J 1991;144:169-173.
clinical criteria, for the diagnosis of Alzheimer‘s disease. Neurology 93. Yesavage ]A, Poulsen SL, Sheikh J et al. Rates of change of common
1990;40:1517-1522. measures of impairment in senile dementia of the Alzheimer’s type.
60. Mitrushina M, Satz P. Reliability and validity of the Mini-Mental State Psychopharmacol Bull 1988;24:531-534.
Exam in neurologically intact elderly. J Clin Psychol 1991;47:537-543. 94. Wechsler D. Manual for the Wechsler Adult Intelligence Scale. New York
61. OConnor DW, Pollitt PA, Hyde JB et al. A follow-up study of dementia Psychological Corporation, 1955.
diagnosed in the community using the Cambridge Mental Disorders of 95. Giordani B, Boivin MJ, Hal AL et al. The utility and generality of Mini-
the Elderly Examination. Acta Psychiatr Scand 1990;81:78-82. Mental State Examination scores in Alzheimer’s disease. Neurology
62. Sackatt DL, Haynes RB, Tugwell P. Clinical Epidemiology: A Basic 1990;40:1894-1896.
Science for Clinical Medicine. Toronto: Little, Brown and Company, 96. Myers BA. The Mini-Mental State in those with developmental disabili-
1985. ties. J Nerv Ment Dis 1987;175:85-89.
63. Siu, AL. Screening for dementia and investigating its causes. Ann Intern 97. DePaulo JR, Folstein MF, Gordon B. Psychiatricscreening on a neurolog-
Med 1991;115:122-132. ical ward. Psychol Med 1980;10:125-132.
64. Goldschmidt TJ, Malli R, Still CN. Recognition of cognitive impairment 98. Henderson VW, MacK W, Williams BW. Spatial disorganization in Alz-
in urimarv
a , care outuatients. Southern Med 1 1983:761264-1270.
I .
heimer’s disease. Arch Neurol 1989;46:391-394.
65. Davous P, Lamour Y, Debrand E et al. A comparative evaluation of the 99. Selnes QA, Carson K, Rovne B et al. Language dysfunction in early- and
short orientation memory concentration test of cognitive impairment. J late-onset possible Alzheimer’s disease. Neurology 1988;38:1053-1056.
Neurol Neurosurg Psychiatry 1987;50:312-317. 100. Ashford J, Hsu LN, Becker M et al. Mini-mental status and activities of
66. Fisk AA, Pannill FC 111. Assessment and care of the community-dwelling daily living: Cross validation by scalogram and item analysis techniques.
Alzheimer’s disease patient. J Am Geriatr SOC1987;35:307-311. Gerontologist 1986;26:143.
67. Huff FJ, Becker JT, Belle SH et al. Cognitive deficits and clinical diagnosis 101. Bondareff W, Raval J, Woo B et al. Magnetic resonance imaging and the
of Alzheimer’s disease. Neurology 1987;371119-1124. severity of dementia in older adults. Arch Gen Psychiatry 1990;4747-
68. Pfeffer R, Afifi AA, Chance JM. Prevalence of Alzheimer‘s disease in a 51.
retirement community. Am J Epidemiol 1987;125:420-436. 102. Breen AR, Larson EB, Reifler BV et al. Cognitive performance and
69. lAckson MD, Ramsdell JW. Use of the Mini-Mental State Examination functional competence in coexisting dementia and depression. J Am
(MMSE) to screen for dementia in elderly outpatients. J Am Geriatr SOC Geriatr SOC1984;32:132-137.
1988;36:662. 103. Stem Y, Hesdorffer D, Sano M et al. Meausrement and prediction of
70. Knopman DS, Ryberg S. A verbal memory test with high predictive functional capacity in Alzheimer‘s disease. Neurology 1990;408-14.
accuracy for dementia of the Alzheimer‘s type. Arch Neurol1989;46:141- 104. Ten L, Larson EB, Reifler BV. Behavioral disturbance in dementia of the
145. Alzheimer’s type. J Am Geriatr SOC1988;361-6.
71. Reed BR, Jagust, WJ, Seab JP et al. Memory and regional cerebral blood 105. Uhlmann RF, Larson EB, Buchner DM. Correlations of Mini-Mental State
flow in mildly symptomatic Alzheimer‘s disease. Neurology and Modified Dementia Rating Scale to measures of transitional health
1989;39:1537-1539. status in dementia. J Gerontol 1987;42:33-36.
72. Black SE, Blessed G, Edwardson JA et al. Prevalence rates of dementia 106. Warren El, Grek A, Conn D et al. A correlation between cognitive
in an ageing population: Are low rates due to the use of insensitive performance and daily functioning in elderly people. J Geriatr Psychiatr
instruments? Age Ageing 1990;19:84-90. Neurol 1989;2:96-100.
73. American Psychiatric Association Committee on Nomenclature and Sta- 107. Fillenbaum GG, Hughes DC, Heyman A et al. Relationshipof health and
tistics. Diagnostic and Statistical Manual of Mental Disorders. 3rd Ed. demographic characteristics to Mini-Mental State Examination score
Washington DC: APA, 1980. among community residents. Psychol Med 1988;18:719-726.
74. DePaulo JR, Folstein F. Psychiatric disturbances in neurological patients: 108. Bassett SS, Folstein MF. Cognitive impairment and functional disability
Detection, recognition, and hospital course. Ann Neurol1978;4:225-228. in the absence of psychiatric diagnosis. Psychol Med 1991;21:77-84.
75. Lautenschlaeger E, Meier HMR, Donnelly M. Folstein vs. Goldfarb mental 109. Mahurin RK, DeBettignies BH, Piozzolo FJ. Structured Assessment of
status exams. Clin Gerontol 1986;4:41-42. Independent Living Skills: P r e l i a r y report of a performance measure
76. Schwamm LH, Van Dyke C, Kieman RJ et al. The Neurobehavioral of functional abilities in dementia. J Gerontol 1991;46:58-66.
Cognitive Status Examination: Comparison with the Cognitive Capacity 110. Tomlinson BE. Morphological changes and dementia in old age. In: Smith,
Screening Examination and the Mini-Mental State Examination in a WL, Kinsboume, M, eds. Aging and Dementia. Holliswood Ny: Spec-
neurosurgical population. Ann Intern Med 1987;107486-491. trum, 1977, pp 25-56.
77. Chandler JD, Gemdt J. Cognitive screening tests for organic mental 111. Martin EM, Wilson RS, Penn RD et al. Cortical biopsy results in Alz-
disorders in psychiatric inpatients. J Nerv Ment Dis 1988;176675-681. heimer’s disease: Correlation with cognitive deficits. Neurology
78. Faustmen WO, Moses JA Jr, Csemansky JG. Limitation of the Mini- 1987;371201-1204.
Mental State Examination in predicting neuropsychological functioning 112. Bums A, Jacoby R, Levy R. Progression of cognitive impairment in
934 TOMBAUGH AND MCINTYRE JAGS-SEPTEMBER 1992-VOL.40, NO. 9
Alzheimer’s disease. J Am Geriatr Soc 1991;39:39-45. 146. Berkman LF. The association between educational attainment and mental
113. Bums A, Lewis G,Jacoby R, et al. Factors affecting survival in Alzheimer’s status examinations: Of etiologic significance for senile dementias or not?
disease. Psychol Med 1991;21:363-370. J Chron Dis 1986;39:171-174.
114. Forette FF, Henry JF, Orgogozo JM et al. Reliability of clinical criteria for 147. Evans DA, Scherr PA, Cook NR et al. Estimated prevalence of Alz-
the diagnosis of dementia. Arch Neurol 1989;46:646-648. heimer’s disease in the United St6:es. Milbank Q 1990;68:267-289.
115. Lopez OL, Boller F, Becker JT et al. Alzheimer’s disease and depression: 148. Gurland BJ. The borderline of dementia: The influence of sociocultural
Neurological impairment and progression of the illness. Am J Psychiatry characteristics on rates of dementia occurring in the senium. In: Miller
1990;147855-860. NE, Cohen GD, eds. Clinical Aspects of Alzheimer’s Disease and Senile
116. OConnor DW, Pollitt PA, Hyde JB et al. The progression of mild Dementia, Aging, Vol15. New York Raven Press, 1981, pp 61-84.
idiopathic dementia in a community population. J Am Geriatr SOC 149. Folstein MF, McHugh PR. Dementia syndrome of depression. In: Katz-
1991;39:246-251. man R, Terry RD, Bick KL, eds. Alzheimer’s Disease: Senile Dementia
117. Small GW, Kuhl DE, Riege WH et al. Cerebral glucose metabolic patterns and Related Disorders, Aging, No. 7. New York Raven Press, 1978.
in Alzheimer’s disease. Arch Gen Psychiatry 1989;46:527-532. 150. Cavanaugh S, Wettstein RM. The relationship between severity of depres-
118. Ten L, Hughes JP, Larson EB. Cognitive deterioration in Alzheimer’s sion, cognitive dysfunction, and age in medical inpatients. Am J Psychia-
disease: Behavioral and health factors. J Gerontol 1990;45:P58-63. try 1983;140:495-496.
119. Tinklenberg 1, Brooks JO, Tanke ED et al. Factor analysis and preliminary 151. Brandt J, Folstein SF, Folstein MF. Differential cognitive impairment in
validation of the Mini-Mental State Examination from a longitudinal Alzheimer’s disease and Huntington’s disease. Ann Neurol 1988;23:555-
perspective. Int Psychogeriatr 1990;2:123-134. 561.
120. Uhlmann RF, Larson EB, Koepsell TD. Hearing impairment and cognitive 152. Folstein MF, Whitehouse PJ. Cognitive impairment of Alzheimer’s dis-
decline in senile dementia of the Alzheimer’s type. J Am Geriatr Soc ease. Neurobehav Toxic01 Teratol 1983;5:631-634.
1986;34:207-210. 153. Heyman A. Aprasia/apraxia and familial aggression in Alzheimer’s dis-
121. Galasko D, Corey-Bloom J, Thal LJ. Monitoring progression in Alz- ease. Ann Neurol 1984;15:615.
heimer’s disesase. J Am Geriatr SOC1991;39:932-942. 154. Kaszniak AW, Fox I, Gandell DL et al. Predictors of mortality in presenile
122. Berg L, Miller JP, Storandt M et al. Mild senile dementia of the Alzheimer and senile dementia. Ann Neurol1978;3:246-252.
type: 2. Longitudinal assessment. Ann Neurol 1988;23:477-484. 155. Bernard BA, Wilson RS, Gilley DW et al. Backward spelling as an
123. Ouslander JG, Zarit SH, Orr NK et al. Incontinence among elderly alternative to serial subtraction on the Mini-Mental State. Paper presented
community-dwelling dementia patients: Characteristics, management at the 20th Annual meeting of the International Neuropsychological
and impact on caregiver. J Am Geriatr SOC1990;38440-445. Society, San Diego, CA, 1992.
124. Berg S, Jeppson L. Cognitive functioning and survival in psychogeriahic 156. Ganguli M, Ratcliff G,Huff FJ et al. Serial sevens versus word backwards:
patients. Acta Psychiatr Scand 1991;84160-162. A comparison of the two measures of attention from the MMSE. J Geriatr
125. Heyman A, Wilkinson WE, Hurwitz BJ et al. Early-onset Alzheimer’s Psychiatry Neurol 1990;3:203-207.
disease: Clinical predictors of institutionalition and death. Neurology 157. Tombaugh TN. The Mini-Mental State (MMSE): Normative data over a
1987;37980-984. 60-year-age range. Unpublished research, Carleton University, 199 1.
126. Cooper JK, Mungas D, Weiler PG. Relation of cognitive status and 158. Watkins P, Gouvier WM, Callon E et al. Equivalence of items on the
abnormal behaviors in Alzheimer’s disease. J Am Geriatr Soci mini-mental state. Arch Clin Neuropsychol 1989;4:381-384.
1990;38:867-870. 159. Yesavage ]A. Degree of dementia and improvement with memory train-
127. Petry S, Cummings JL, Hill MA et al. Personality alterations in dementia ing. Clin Gerontol 1982;1:77-80.
of the Alzheimer type. Arch Neurol1988;45:1187-1190. 160. Garcia CA, Tweedy JR, Blass JP. Underdiagnosis of cognitive impairment
128. Binder EF, Robbins LN. Cognitive impairment and length of hospital stay in a rehabilitation setting. J Am Geriatr Soc 1984;32:339-342.
in older persons. J Am Geriatr Soc 1990;38:759-766. 161. Bucknam M, Kanar SM. Predicting performance on the MMSE. J Am
129. Fields SD, MacKenzie DR, Charlson ME et al. Cognitiveimpairment. Can Geriatr SOC1988:361072.
it predict the course of hospitalized patients? J Am Geriatr SOC 162. Kessler J, Beil C. Predicting performance on the MMSE? J Am Geriatr SOC
1986;34:579-585. 1988;36:1072-1073.
130. Johnston M, Wakeling A, Graham N et al. Cognitive impairment, emo- 163. Solomon DH. Anti-sense of prediction equations. J Am Geriatr SOC
tional disorder and length of stay of elderly patients in a district general 1988;36:1072.
hospital. Br J Med Psychol 1987;60:133. 164. Stem Y, Sano M, Paulson J et al. Modified Mini-Mental State Examina-
131. Mayeux R, Stem Y, Spanton S. Heterogeneity in dementia of the Alz- tion. Neurology 1987;37(Suppl 1):179 (Abstract).
heimer type: Evidence of subgroups. Neurology 1985;35:453-461. 165. Stem Y, Mayeux R, Sano M et al. Predictors of disease course in patients
132. Davidson HA, Bonie MJ, Crilly RG. Copy task performance and urinary with probable Alzheimer’s disease. Neurology 1987;37:1649.
incontinence in Alzheimer’s disease. J Am Geriatr Soc 1991;39:467-471. 166. Gauthier S, McDowell I, Hill G. Canadian study of health and aging
133. Uhlmann RF, Larson EB. Effect of education on the Mini-Mental State (CaSHA). Psychiatr J Univ Ott 1990;15:227-229.
Examination as a screening test for dementia. J Am Geriatr SOC 167. Pfeffer RI, Kuroski TT, Harrah CH et al. A survey diagnostic tool for
1991;39:876-880. senile dementia. Am J Epidemiol 1981;114:515-527.
134. Brayne C, Calloway P. The association of education and socioeconomic 168. Smith A. Symbol Digit Modalities Test. Los Angeles: Western Psycholog-
status with the Mini Mental State Examination and clinical diagnosis of ical Services, 1973.
dementia in elderly people. Age Ageing 1990;19:91-96. 169. Raven JC. Coloured Progressive Matrices. London: HK Lewis and Co,
135. Teng EL, Chiu HC, Schneider LS et al. Alzheimer’s dementia: Perform- 1969.
ance on the Mini-Mental State Examination. J Consult Clin Psychol
1987;55:96-100.
136. OConnor DW, Pollitt PA, Hyde JB et al. The prevalenec of dementia as APPENDIX
measured by the Cambridge Mental Disorders of the Elderly Examination.
Acta Psychiatr Scand 1989;79:190-198. Mini-Mental State Examination (MMSE)
137. Cullum CM, SMemoff EN, Lord SE. Utility and psychometric properties
of the Mini Mental-State Examination in healthy older adult. Paper Questions Points
presented at the 19th Annual meeting of the International Neuropsycho-
logical Society, San Antonio, TX 1991. 1. What is the: Year? Season? Date? Day? 5
138. OConnor DW, Pollitt PA, Treasure FP. The influence of education and Month?
social class on the diagnosis of dementia in a community population.
Psychol Med 1991;21:210-224. 2. Where are we: State? County? Town 5
139. Kay DWK, Beamish P, Roth M. Old age mental disorder in Newcastle or City? Hospital? Floor?
Upon Tyne. 11. A study of possible social and medical causes. Br J
Psychiatry 1964;110:668-682. 3. Name three objects (Apple, Penny, 3
140. Parsons PL. Mental health of Swansea‘s old folk. Br J Prev Soc Med Table), taking one second to say each.
1965;19:43-47.
141. Berg L, Danziger WL, Storandt M et al. Predictive features in mild senile Then ask the patient to tell you the
dementia of the Alzheimer types. Neurology 1984;34:563-569. three. Repeat the answers until the
142. Filley M, Brownell HH, Albert AL. Education provides no protection patient learns all three.
against Alzheimer‘s disease. Neurology 1985;35:1781-1784.
143. Katzman R, Brown T, Thal LJ et al. Comparison of rate of annual change 4. Serial 7s. Subtract 7 from 100. Then 5
of mental status score in four independnet studies of patients with subtract 7 from that number, etc. Stop
Alzheimer‘s disease. Ann Neurol 1988;24:384-389.
144. Kittner SJ,White LR, Farmer ME. Methodological issues in screening for after five answers.
dementia: The problem of education adjustment. J Chron Dis Alternative: Spell WORLD backwards.
1986;39:163-170.
145. Naugle RI, Kawczak K. Limitationsof the Mini-Mental State Examination. 5. Ask for the names of the three objects 3
Clevel Clin J Med 1989;56277-281. learned in # 3.
JAGS-SEPTEMBER 1992-VOL. 40, NO. 9 MINI-MENTAL STATE EXAM 935

6 . Point to a pencil and watch. Have the 1 10. Have the patient write a sentence of 1
patient name them as you point. his or her own choice.
7. Have the patient repeat "No ifs, and, 3 11. Have the patient copy the following 1
or buts". design (overlapping pentagons).
8. Have the patient follow a three-stage 3
command: "Take the paper in your
right hand. Fold the paper in half. Put
the paper on the floor".
9. Have the patient read and obey the 1
following: "CLOSE YOUR EYES".
(Write it in large letters). Total points = 30