Article Text

Download PDFPDF

The validity of using the mini mental state examination in NICE dementia guidelines
  1. R J Davey,
  2. S Jamieson
  1. Department of Neurology, Leeds General Infirmary, Great George Street, Leeds LS1 3EX, UK
  1. Correspondence to:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The mini mental state examination (MMSE) is widely used as a rapid means of quantifying cognitive function.1 The National Institute for Clinical Excellence (NICE) guidelines concerning the use of cholinesterase inhibitors (CI) in Alzheimer’s disease recommend using the MMSE as quantifiable measure to inform decisions regarding initiation and continuation of drug treatment.2 Our study questions whether poor interrater reliability of the MMSE makes it an inappropriate tool for monitoring drug response.

A postal survey evaluating the MMSE section termed “attention and calculation” was conducted among all consultant neurologists in the UK. The original instructions regarding this section involve asking the patient to count backwards in sevens from 100 for five subtractions, or, if the patient “cannot or will not perform this task” to spell “WORLD” backwards, scoring the number of letters in the correct order.1

Of the 407 questionnaires sent, there were 234 (58%) responses.

The MMSE was used in clinical practice by 91% of respondents, with 51% of respondents describing their use as “frequent”. Test choice and method of scoring this section of the MMSE are shown in fig 1.

Only 10% of respondents were aware of schemes describing standardised scoring of mistakes when spelling WORLD backwards. Raters were asked to score a sample incorrect response “DRLOW”. We did not allocate a “correct” score for this example since we believe the original guidance on how to score errors1 is imprecise. Out of a maximum of five points, 51% assigned a score of three and 25% a score of one. Other scores included two respondents assigning a score of five. When scoring “93–85–78–71–64”, 59% assigned a score of four points and 23% a score of one.

This survey of consultant neurologists confirmed substantial variability in the use and scoring of the serial sevens/WORLD backwards section of the MMSE. This interrater error leads to a potential score difference of up to four points for this section alone. We focused on this section of the MMSE alone to ensure brevity of the survey, as we perceived that the scoring method assigned to these questions is particularly dependant on the rater’s interpretation. Rogers et al found the mean improvement in MMSE score after 12 weeks of treatment with CI to be 1.3 points.3 NICE guidelines for CI prescription in Alzheimer’s disease give specific recommendations to stop these drugs if a patient’s score deteriorates when the MMSE is repeated two to four months after commencement, or if the score falls below 12 points. Our results show that typical patterns of scoring the MMSE are too inconsistent to detect such small improvements in cognition and so may lead to inappropriate cessation of CI treatment.

The findings are in keeping with the large interrater variability previously demonstrated among a small group of psychiatrists,4 suggesting that the results likely to be generalisable to most doctors. Our study’s incomplete response rate and finding that only half the respondents use the MMSE frequently introduces sources of selection bias, however, we believe that numbers were large enough to derive meaningful conclusions.

Should NICE use the MMSE as their recommended cognitive rating scale in Alzheimer’s disease? Few of the studies showing benefit from CI use in Alzheimer’s disease use the MMSE as a primary outcome measure. However, cognitive tests used in CI trials such as Alzheimer’s disease assessment scale (ADAS-cog) are time consuming. Using a modified version of the MMSE, a standardised scheme to score errors made, or an alternative test, introduces new problems.

Although the MMSE has a number of limitations, awareness of these highlights some simple ways of improving its reliability in practice. Physicians should rigorously standardise their own testing methods. Local MMSE scoring guidelines could be instituted. It is helpful to retain completed score sheets in patient notes and to document in clinic letters whether serial sevens and/or WORLD were tested.

Because NICE estimate there to be 400 000 sufferers of Alzheimer’s disease currently in England and Wales, the cost implications to the NHS regarding the appropriate use of CI’s is vast. We call for inclusion of more detailed guidance on MMSE use in NICE recommendations and for further studies of new dementia treatments to use the MMSE as an outcome measure.

Figure 1

Use of counting backwards from 100 in sevens and spelling WORLD backwards in clinical practice.


The Association of British Neurologists kindly provided contact details for UK neurologists.



  • Funding: Eisai-Pfizer supported the postal costs for this survey.