I think MMSE can be used as a screening tool for disease patient, say neurological patients. You may be interested in reading the following:

Dick et al., JNNP 1984;47:496-499

Generally a MMSE's score below the cut-off of 24 is used as sign of possible dementia in healthy subjects, besides this tool was specifically constructed for neurological population. I would suggest to use also other neuropsychological tools, such as the Clock Drawing Test or the Coloured Progressive Matrices to have a more complete picture of the cognitive profile.

Thank you, Brandon. That was exactly my opinion by stating that loaded question. IMHO, the MMSE is a screening tool that is now becoming rapidly outdated. I continue to review submissions, however, that use the MMSE as a (or worse, THE) cognitive outcome measure. I typically respond to these with snarky comments, such as to the validity of a measure that yields significant differences if a person remembers what floor they are on (again) and that they rehearsed APPLE, PENNY, TABLE for two hours prior to their second session.

I also suggest the MoCA be investigated as an alternative screening measure and that researchers interested in cognitive outcomes should PLEASE speak with a neuropsychologist or cognitive scientist to obtain useful measures of the extremely broad term of cognitive function.

I agree with others who have pointed out the practice effects of the MMSE. Along with this come the floor and ceiling effects. Of course it also fails to provide assessment of many domains. I am currently using MoCA in a middle aged population of Mexican Americans . It works well even in comparatively young people . People might want to look at the NIH cognitive toolbox which is claimed to take only 30 minutes to administer and includes six domains. I think MMSE was really developed as a Quick screening tool for use in clinical situations and not for research purposes . Even so there are now better Screening tools .

First it strongly depends upon administering it. Without any experience the results are not trustworthy. It's good enough for detecting an ad hoc condition of the patient (so you need to repeat it after say three months. It's value is to distinguish between the several fields and to compare with the patient's history and information from a relative (or someone who knows the patient well).

Thanos Didaskalou

I can't help but agree with the comments about it's coarseness and relatively poor operating characteristics over repeated measurements. The tool began its life as a way for psychiatrists to standardize a mental status exam (McHugh & Folstein's original intention) but has developed many more applications over time. I can only add that it still remains the de facto standard for determining severity of dementia in almost all RCTs undertaken by pharmaceutical companys. On the plus side, pharmaceutical trials usually elect to not use it as a primary outcome measure, at least in recent years, although it usually is a secondary outcome measure. Preference for a primary outcome measure of cognitive change is generally given to the ADAS (Alzheimer's Disease Assessment Scale) in mild to moderate dementia and to the SIB (Severe Impairment Battery) for moderate to severe dementia. However, as attention shifts to early stages of illness (e.g. MCI or mild cognitive impairment), the value of the ADAS as a tool for tracking cognitive change has been called into question by myself and others. Other instruments are under development for milder forms of cognitive impairment, some with characteristics that make them useful both for classification and repeated measurement.

I agree with the others who have pointed out that the MMSE is useful as a rough screener (separating healthy from completely impaired individuals), but has a very poor track record for measuring change over time or measuring subtle cognitive differences. The test-retest reliability is terrible among healthy people. The scores are very heavily influenced by educational background, and are also unacceptably influenced by the administrator. (Some/most have their own quirky administration and scoring practices: Scoring "no if and or buts", scoring DLORW, giving 3-stage commands as pts perform 3-stage commands, having pts read "close your eyes" aloud, counting 100 as a correct response in serial sevens, substituting verbs and abstract nouns like "dignity" for the word recall, only administering backwards spelling, which is easier, not asking all orientation questions, giving credit for seasonal answers for season, etc.). My main criticism, however, is that the MMSE neglects processing speed and complex processing. Besides, the MoCA is free...

I agree with all of you that point out the low validity of MMSE as a cognitive outcome in older adults, mainly for the poor sensibility that has this tool for differentiate healthy older people from those at same age that could being on an early dementia stage but have a high education background and so for can easily get scores of >24.

So, if the MMSE is a screening test that just surround us to the need of a deeper/better evaluation, don't you think it its time to retire, becouse as you mention, there's a lot other test (builded on the basis of systematic investigation not just as clinical essay) that has better ROC curves on sensibility and specificity for differentiate dementia and normal elder, and furhtermore, to contrast MCI and early dementia onset (such as MOCA, MIS)??

MMSE is a useful tool for cognitive screening, and can differentiate healthy people from those with impairment. On the other hand, as a longitundinal measure of screening it is susceptible to learning effect, as occurs in almost of memory tests. Without discuss the importance of MOCA, my suggestion is that you use the second version of the MMSE (MMSE-II). This review sought to address the disabilities of the first version, creating two application forms: blue and red. Various changes were made, including the exclusion of problematic items, as to say the word DLROW. In addition, other changes were made and a expanded form of evaluation was included. The expanded version of MMSE-II consistis of tasks designed to screening the subcortical functioning. In Brazil, studies of adaptation were made by our research group, with permission of the publisher. It is worth do you confer.

The Mini-Mental State Examination is a very useful instrument, with international recognition. However, the transversal use in healthy adults, it is not very sensitive to find mild, initial deficits. In this case, the neuropsychological evaluation is indicated.

Yes, it is good enough as a global measure and for screening purpose. However, if you want to assess different cognitive domains such as memory, executive function, visuospatial function, et al, you will need neuropsychological tests. Also, remember MMSE has ceiling effect, the examinee only can get a full mark of 30 so it does not reflect the normal distribution of cognitive functioning in the population.

Sensitivity and specificity of the MMSE are poor. Even patients with beginning dementia can score 30 points. In the first year after publication other tests and from there on numerous tests demonstrated superiority. Unfortunately, these scientific findings are widely ignored. Moreover, an endless number of new tests is developed without demonstrating superiority over already existing tests.

There are many issues with using the MMSE as an outcome measure. First and foremost, if you are collecting multiple time points (e.g., pre, post, etc.), the MMSE has horrific practice effects. Scores will increase, particularly with relatively healthy older adults, simply by remembering the 30 items on the test. For example, it would not be uncommon to have an older adult in clinic tell you they went home and practiced WORLD backwards or serial 7s. They'll also practice all of the orientation questions with a caregiver/spouse on the way in. Finally, there are so many issues with with loading towards verbal strengths and not enough executive functioning assessment. I would stay away from it unless you are interested in a quick and dirty global functioning assessment. If you cannot construct a quick 30-40 minute neuropsych screening/assessment battery, I would recommend something a bit more comprehensive than the MMSE, such as the MOCA or Kokmen. Just my 2...

I agree with Ralph Ihl. The fact that the MMSE has a long tradition in screening leads to its inclusion in papers where clearly it does not provide added value.

Javid Khalique. Mental Health Services for Older People. Nottinghamshire Healthcare NHS Trust

The MMSE proved a very rough tool in pre - post measurement of any cognitive i benefits to Caribbean and Asian origin elders living with dementia who experienced culturally adapted cognitive stimulation therapy. Education background, language, numeracy skills and cultural understanding of the questions were points of consideration.

Any thoughts on this??

Hi, I agree with those not in favour of MMSE for all those reasons they have discussed! I would rather using MoCA,even though as a Clin Psych I am a bit hesitant to consider it a good outcome measure. However, MoCA has three versions that can be used alternatively to overcome the practice effect.

Hi All,

First let me admit that I agree completely with what others have said in terms of the limitations of the MMSE for outcome measure. Also, as Javid mentions, I have found its major limitation is education background in that it does not do a good job in indicating whether an illiterate individual is healthy vs pathological (plus it makes comparisons to literate individuals complicated).

Assumining the test is performed on literate adults (who have lived in the area for some time), the MMSE becomes a practical screening tool between pathological/healthy adults with emphasis on memory loss. Often times it is necessary to have a quick estimation of memory loss (potentially dementia???). For example, for my EEG studies, time is limited especially when working with older populations because the individuals may come in off of medications for some time and are willing to stay off of them for only so long. Also, I have been told by some that they need to eat or nap after a certain amount of time. It is true that a neuropsychological test battery can be performed on one day and testing on another, allowing the use of more thorough measures. On the other hand, I have also been told by participants that they would participate on the spot rather than making a second trip to the test site, as it may require organizing someone to drive them.

MMSE is a screening instrument, and if use with clock test is more complete for cognitive screening for literate clients.

The MMSE is surely NOT a valid test of cognitive function. In addition to what has been said already in that vein, it. Is a fact that at least 18 of the 30 points on the test are devoted ONLY to RECALL. Recall is but one form of retrieval from memory. Whether it is recalling the day if the week or the name of an object, it is recall. Cognitive abilities include far more than recall and far more than recall of what is demanded by this screening device. Calling the MMSE a valid test of cognitive function is, itself, a cause for concern about the cognitive functioning of the speaker.

Just to add to Brandon Ally's comment: Yes, we should indeed all stay away from the MMSE; it's embarrassing to stick to it, both for researchers and certainly for clinicians.

I'm not sure it should be used as a screening tool at all. To build on Brandon's answer, I've tested patients with amnesia who managed to get an adequate score. I'm thinking of one in particular who aced serial 7's (many people will get at least one wrong), happened to remember one of the words, and was mostly oriented (partially because we were testing him in a hospital he'd been going to since well before the onset of his amnesia). Plus, you lose the "surprise" factor in the recall portion--examinees know they will be tested and will come out with the words before you even ask. In my view, if someone gets a low score on the MMSE, there's probably a problem, but if they get a high score, you don't know anything at all. I tend to use Cognistat, but there are a number of good choices out there.

In my experience as both a physician and a Master's Level therapist the MMSE is a horrible test of cognitive function, from my experience. I tend to agree with Daniel low scores indicate the need for further investigation, however high scores tell you absolutely nothing. I too have had better luck with Cognistat.

i used MMSE to compare 3 groups (AD, old and young adults) and i found some young subject have 27/30 on MMSE (unexpected result !!!). If you focus only on this result, we can say that maybe they can develop AD!!! but the explanation is :young people are very bad in mental calculation (because of calculator...)that's why they have -3 in the subtraction exercise

==> for my opinion MMSE is not up to date, and we need to use specific tests and to spend more time , than to use just one and is not reliable...

Hi, I agree with the critical statements before.

According to my experience of neuropsychological diagnostics of dementia in an ambulant setting about 13,7 per cent out of 132 persons achieving 27 until 30 points in the MMSE showed substantial clinical signs of a dementia in a detailed neuropsychological examination. In this sample there is also a significant correlation between MMSE and depression.

It depends also on the sample how many persons with cognitive decline are detected. I found that a clinical setting like a hospital means having a threshold for participation behaviour by psychological factors. In an ambulant setting you get more persons with mild cognitive impairments or a beginning dementia.

Using MMSE as single measure for healthy older adults or persons with beginning deficits (mild cognitive impairment) means that the whole study is not only a waste of time and energy but also contestable. Especially various treatment effects of a beginning dementia cannot be revealed by using MMSE as there are different premorbid starting points of the cognitive abilities which are not seen in the MMSE.

It´s better to use the competence of a neuropsychologist and work out a special set of psychological tests in order to answer your specific scientific question.

If you want to test for a cognitive outcome, then you should use tests and tasks for the outcome you are working on specifically. I mean: if you are working on verbal memory, then find a test for verbal memory (Rivermead Behavioral Memory Test); if you are working on executive functioning, then look for a test of executive functioning (Behavioral Assessment of Dysexecutive Syndrome), and so on. As a clinical neuropsychologist and researcher, I can't see how the MMSE score can be regarded as a cognitive outcome. However, it is unfairly needed for research purposes very often.

As an aside, it doesn't matter if there are practice effects. What is important is to know whether your patient's test-retest scores are significantly higher than what would be expected based on practice effects. And there are some ways to calculate that. See Crawford and Garthwaite (2007) and Kevin Duff (2012).

Best

Any opinions on the BCAT (Mansbach)? It's another one that is over in 12 minutes and can be administered by a robot and thus is not going into clinical depth, but I don't hear as much about it, good or bad, as we're hearing about the MMSE.

Back to the original question about outcome measures: It seems that any brief instrument will be subject to practice effects, and despite Mr. Oltra-Cucuarella's point about eliminating practice effects statistically, in clinical practice I think people are just taking a guess at whether improvement is due to intervention or practice effects (and have a bias toward giving credit to interventions).

There is a practical issue here: The best screening instrument may not be the best outcome measure, but we may have to err on the side of using good outcome measures right from the start, so we can establish a baseline. This means doing more elaborate screening than is often performed, more than clinicians may feel they have time for. I wonder how the use varies in different settings; in general psychiatric practice and gerontology practice I've seen a lot of reliance on those overlapping pentagons.

Thanks for your words, Mr. James Foley. I really appreciate them.

I didn't mean "eliminating", but instead "taking into account" practice effects. If you know that a population retest score is 3 points higher than test scores, then you have to take that into account when you analyze your patients test-retest scores. And, more important, is the question about "statistically" significant change. A change may not be statistically significant but clinically significant. If a discrepancy between test-retest scores is higher than 94% of a population (healthy controls, peopel with brain injury without intervention,...), then probably it won't be statistically different, but obviously is clinically significant as it is very uncommon. Again, it is worth taking a look at Crawford and Garthwaite papers about reliable change index.

To add to the discussion there are plenty of limitations and problems with the MMSE, but one big issue is that although it can arguably distinguish dementia quite well it is very poor at detecting early, more subtle changes associated with the stage of mild cognitive impairment, which can precede a diagnosis of dementia. I have administered the MMSE to patients with clear cognitive deficits in memory, executive function etc, compared to age-matched norms but their MMSE might be 28/30 - so not a helpful marker at all. It is also very biased towards those with a higher education level.

You make a fair point Sarah Montgomery, yet the MMSE is widely used clinically to identify and screen for cognitive impairment in older adults pre-diagnosis (not those already at an advanced stage of Alzheimer's or Dementia). In these cases, as Elizabeth Kehoe has mentioned, the MMSE often fails to identify subtle cognitive impairments, which are precursors to numerous age-related disorders including dementia.

Furthermore, from my own experience working with community-dwelling older adults, some appear to be aware of the exact types of questions (and answers) on the MMSE before I administer it, through discussion with peers who have completed the task after visiting their GP with concerns over memory loss etc. In these instances, the MMSE is even less likely to efficiently assess cognitive decline (and rather just recall of known correct answers).

Some tests have an A form and a B form in order to avoid remembering. So a B form of the MMSE should be developed. On the other Hand, the shortcomings of the present MMSE as outlined, can render false results but opnly in one direction, namely that the Patient may appear better than he actually is. But this is sometimes even helpful in practise when it comes to gettimng the allowance from the health insurance for a special medicament. Best wishes Yours Lüder Deecke, Vienna

Should the MMSE (Mini Mental State Exam) be limited to use as a screening tool for only healthy populations? Well, if there are concerns regarding the sensitivity and specificity of the MMSE, it seems that one should be cautious using it as a sole screening instrument that could be derived from other means. If we are screening a primarily healthy population of adults, with no sign of cognitive impairment, It might be best to use background information such as health/medical history, age and family history (including family members with dementia, and other issues), and a brief interview, a brief narrative recall (two trials of repeating the same narrative with 10 minutes between the first and second retell to determine any rapid deterioration in recall of important details/plot, and a vision/hearing screening

Not only are sensitivity and specificity important, but also positive and negative predictive values. What is the probability that one of your clients/patients actually has dementia if he scores positive on the test? This is the important question, not only if people with dementia scores positive in the test. however, I think that clinical information along with a detailed neuropsychological assessment is the best way, even if scores are in normal ranges. There are so many factors influencing performance that I don't think that a brief test can identify cognitive impairment with few questions

MMSE may be used as a part of screening cognitive function but may not be a good tool for pre and post test as it has practice effects. The scores too are influenced by age as well as educational level.

Please have a look on the scores based on age and education by Crum et al. 1993.

Crum, R. M., Anthony, J. C., Bassett, S. S., Folstein, M. F. (1993). Population-based norms for the mini-mental state examination by age and educational level. JAMA, 18, 2386-2391.

I don't think that practice effects are an important issue if there are data that can be used to study test-retest. I mean: if you have summary data from the group with your patient's age and educational level you can build a regression equation and test whether the difference between test and retest is abnormally infrequent in the sample. Obviously, if you have test-retest data only for the whole group, then you cannot interpret test-retest. Let's take an example. Imagine that your patient scores 26 on the test and 28 three months later. You can interpret improvement as practice effects. OK. But, what if a difference of +2 or less in three months test-retest is observed only in 3% of people same age and educational level? 97% of a sample of peers would score 3 or more three months later. Then you can interpret as something is going wrong, even if practice effects exist.

Best,

Javi

MMSE could be used together with other screenings (e.g. Clock drawing). Actually, in practice I recommend to the clinicians to use three screening instruments in order to diagnose the cognitive function in older adults. MMSE should never be used as the sole instrument for a psycho-diagnosis on cognitive function.

See some papers and books too: http://www.cfmal.com/Guardianship%20related%20references-1-11-10/Tombaugh%201992%20The%20MMSE%20-%20A%20comprehensive%20review.pdf

and: xa.yimg.com/kq/groups/27461618/536209845/name/geraitry_psyquiatry-2002.pdf#page=150

As a practicing psychogeriatrician I would like to see the MMSE abandoned. There are far better tests out there such as the MOCA or the ACE-III. For use in primary care there is the GPCOG. The MMSE is poor in terms of frontal lobe function. I think that the Folstein's contributed significantly to cognitive screening but sadly they will be remembered for using the MMSE for financial gain (for those who don't know, the MMSE was free for many years but has now been licenced and a fee is payable for its use)!

I don't think that the MMSE would pick up anything in healthy ageing! Even MCI is not that easily picked up by the MMSE. The MoCA (Montreal Cognitive Assessment) is better in distinguishing between MCI and the early stages of dementia, but maybe even this test may not pick up healthy ageing deficits, or only in a limited fashion.

MMSE doesnt assess the executive functions, its better to use MoCA or ACE.

In the Gerontopsychiatry in Germany we usually use CERAD+ (The Consortium to Establish a Registry for Alzheimer's Disease) to get a personal profile of cognitive impairment in Alzheimer's Disease. However, you can use this test also to detect above-average cognitive performance. Z-Scores above 1.3 are considered above-average in relation to the norm sample. You can check the following paper: Morris JC et al, Neurology. 1989.

Dear Ahmed, how long does it take to examine a patient with CERAD+? The MMSE takes 5min. As I suggested earlier, the MMSE should come in 2 Versions which are validated: Form A and Form B. Many psychological tests do it that way to vercome the Problem of remembering the test when you do it a second time, e.g. to quantify effects of treatment. Best! Yours Lüder Deecke Vienna

MMSE VS. MoCA

The MMSE has never been designed for research settings and shows very poor test-retest performance so should indeed never be considered for follow-up studies. The MMSE also relies too much on memory and not enough on other cognitive functions that are known to be altered with age. On the other hand, the MoCA has much better psychometric properties. It is free of use and downloadable from internet (http://www.mocatest.org). I still do not understand why clinicians are still using the MMSE. Especially that Psychological Assessment Resources is claiming copyrights over the MMSE and pretends only its payable version is to be used.

Given repeated measures of cognitive functions are indeed biased by long term memory. The idea of developing two version would be good and should be considered for the MoCA.

References:

Nasreddine ZS, Phillips NA, Bedirian V, Charbonneau S, Whitehead V, Collin I, Cummings JL, Chertkow H: The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc 2005, 53(4):695-699

Bernstein IH, Lacritz L, Barlow CE, Weiner MF, DeFina LF: Psychometric evaluation of the Montreal Cognitive Assessment (MoCA) in three diverse samples. Clin Neuropsychol 2011, 25(1):119-126.

Hoops S, Nazem S, Siderowf AD, Duda JE, Xie SX, Stern MB, Weintraub D: Validity of the MoCA and MMSE in the detection of MCI and dementia in Parkinson disease. Neurology 2009, 73(21):1738-1745

Luis CA, Keegan AP, Mullan M: Cross validation of the Montreal Cognitive Assessment in community dwelling older adults residing in the Southeastern US. Int J Geriatr Psychiatry 2009, 24(2):197-201

McLennan SN, Mathias JL, Brennan LC, Stewart S: Validity of the montreal cognitive assessment (MoCA) as a screening test for mild cognitive impairment (MCI) in a cardiovascular population. J Geriatr Psychiatry Neurol 2011, 24(1):33-38

Smith T, Gildeh N, Holmes C: The Montreal Cognitive Assessment: validity and utility in a memory clinic setting. Can J Psychiatry 2007, 52(5):329-332

Videnovic A, Bernard B, Fan W, Jaglin J, Leurgans S, Shannon KM: The Montreal Cognitive Assessment as a screening tool for cognitive dysfunction in Huntington's disease. Mov Disord 2010, 25(3):401-404

The MMSE is a measure of mental status and has ceiling effects for high performing or well educated persons. Anyone investing in a longitudinal study should invest time in finding cognitive tests suitable for adults who perform well. The MMSE is useful as adjunct measure but should never be used alone. Either do the longitudinal study with a reasonable battery of tests, or not at all. Look at the longitudinal studies in the literature. Try for example the batteries used in the Maine Syracuse Longitudinal Study or the Framiingham stueies.

I do not think that Folstein tried to replace the full psychiatric mental state, he named it a mini mental state examination and aimed at a duration of 5 minutes. In clinical praxis it is quite helpful, and the health insurance institutions in Austria postulate that it is compulsary for the prescription of certain medicaments for dementia. For me it is a very valuable rough orientation. The ceiling effect on intellectual people, Merryl mentioned, is correct. Einstein would have made 30 points and not more! But this is not of particular relevance in the settings the MMSE is used, namely to diagnose mild cognitive impairment and dementia. The only thing that has to be improved, I think I suggested this before, is the repeated use of the MMSE, and the problem that patients remember items from the previous test. Therefore, there are many psychological tests that come in a Form A and a Form B. So my suggestion is to develop and standardize a Form B for the mini mental state examination. Best wishes! Yours Lüder Deecke, Vienna

The authors of MoCA say the following regarding it's use in measuring cognitive impairment:

MoCA is better for mild stages

MMSE is better for more advanced stages

(advanced = AD pts with more functional impairment)

Nasreddine, Z. S., Phillips, N. a., Bédirian, V., Charbonneau, S., Whitehead, V., Collin, I., … Chertkow, H. (2005). The Montreal Cognitive Assessment, MoCA: A brief screening tool for mild cognitive impairment. Journal of the American Geriatrics Society, 53(4), 695–699.  For the above comments, see bottom of pg 697, top of pg 698.

Still I think it is a valuable rough orientation, and with 28 to 30 Points one can say there is no dementia. With 20 to 28 Points there is a light dementia, from 19 on below there is a moderate dementia and from 10 Point down there is e severe dementia. This orientation is necessary for legal judgements of the mental abilities. best wishes Yours Lüder Deecke Vienna

I´m sorry to read that persons with a normal MMSE-Score are still considered as non demented.

In a clinical polulation in an ambulant setting I made a different experience: In comparison to a detailed neuropsychological testing (including  an anamnesis by the caregivers) and a subsequent rating with the Global Deterioration Scale (GDS, Reisberg, 1991) it appeared that 13,7 per cent of persons achieving  a MMSE-Core of 27 - 30 showed already a dementia (N=132; data collection 2008-2012). In aiming diagnostics at an early stage the best fitting choice of the screening method is crucial.

In a study with a healthy population a neuropsychological assessment is valuable.

Best wishes to all from Steyr, Austria.

MMSE is an old test to evaluate cognitive decline in older adults. Other tools like ACE-3 or MOCA offer more balanced evaluation of cognitive decline, as some geriatricians explained to me.

I think that there are better options. The MMSE has a number of weaknesses….half the test items are look at recall, it largely ignores executive function….The impact of education level is important to consider but age grouped norms are rarely considered… In clinical practice I would think we should be assessing a range of cognitive domains so consider the MoCA or RUDAS better options.

Yes, but this scale has various cut points stated in litrature for describing the normal or mild congnitive imparment (MCI) state, for example some studies reported a score of 24 points or less for describing the MCI and other stated a cut point of 26

The only reason why people still use the MMSE nowadays is that many colleagues have used it in the past. As some of the commentators have noted, it is not to be used for dementia screening (for various reasons; e.g. a severe fronto-temporal dementia, behavioral variant, will not be reflected in MMSE scores), inter-rater reliability is low, cut-offs poorly defined, executive functions not explicitly assessed (but will influence scores anyway) - I recommend the MoCA instead.