1. Mini-Mental State Examination - 2nd Edition

2. Overview
Original MMSE is one of the most widely used brief screening instruments for cognitive impairment
Has been used in a variety of settings, including screening individual patients, tracking progress over time, screening for large populations, and clinical trials

3. Goals for the Revision
Revise some of the original items to better standardize its administrations, particularly for translations
Provide an even briefer version that could be used for rapid assessment
Provide a slightly longer version that would be more sensitive to subcortical dementia and that would not have a ceiling effect
Develop equivalent alternate forms to decrease practice effects in serial administration
2. Very useful for large clinics and large clinical/drug trials

4. Goal #1: Revise Original Items
The MMSE-2 Standard Version (MMSE-2:SV) maintains the same structure and scoring as the original MMSE
Changes were made to improve problematic items and to better standardize administration in other languages (e.g., penny, no if, ands, or buts)
Wanted to make translation to other languages easy.

5. MMSE-2: Standard Version
Task
Description
Identical task on Original MMSE
Revision on
MMSE-2:SV
Registration & Recall
Ability to repeat and retain three unrelated words, and then recall after a short intervention task
Words have been made slightly more difficult and easier to translate
Orientation to Time
Identify current year, season, month, day of the week, and date X
Orientation to Place
Identify state, county, city/town, building, and floor currently in
Attention & Calculation (Serial 7s)
Count backwards by 7s
No longer can use WORLD spelled backwards as alternate task
Naming
Ask to identify body parts when pointed to by examiner
Change from “watch” and “pencil” to body parts allows for translation and no use of external materials
Repetition
Required to repeat a sentence that contains words not often said together
Revised to include a sentence that is easier to translate and difficulty slightly decreased
Comprehension
Understand and carry out a three-stage verbal command
Removed the reliance on motor responses.
Reading
Read and follow instructions
Writing
Asked to write a sentence
Drawing
Asked to copy intersecting pentagons

6. Equivalency Between the MMSE and MMSE-2:SV
Like the MMSE, the MMSE-2:SV has a raw score range of 0-30
The generalizability coefficient (n = 411) between the MMSE and the MMSE-2:SV total raw score was .97
Therefore it is possible to switch from MMSE to the MMSE-2:SV without compromising longitudinal data and without any change in the normal range of scores

7. Goal #2: Develop Briefer Version
Can be used for quick cognitive screener, specifically when an individual has not been referred for specific cognitive impairment
Composed of Registration/Recall, Orientation to Time, and Orientation to Place
Raw score ranges from 0-16 points
Tasks were selected based on literature review, use in the MMSE, and their sensitivity and specificity to detect dementia

8. MMSE-2: Brief Version AD = Alzheimer’s (Dementia)
Subcortical dementia = milder form of dementia. Less impaired than AD populations.
Of all the options for a Brief Version, Option C, which is the first three subtests differentiates both severe and mild populations from healthy controls more effectively.

9. Goal #3: Develop Expanded Version
Consists of all of the items on the MMSE-2:SV plus two new tasks:
Story Memory: An immediate recall of a brief story
Processing Speed: A symbol-digit coding task
Theory behind the decision:
Wanted an additional memory task due their importance in this area.
Added subtests that statistically differentiate mild dementia from normals
Wanted to make it more difficult to fake

10. MMSE-2: Expanded Version
Improves the clinical utility of the MMSE by:
Extending the test’s ceiling
Increasing the range of raw scores (0-90)
Increasing the sensitivity for individuals with less severe cognitive impairment (subcortical dementia, MCI)

11. The higher the score, the closer to normal.
The Expanded Version demonstrated sensitivity was improved over the standard version for mild dementia problems (>22)

12. Expanded was also better than standard for mild subcortical (>22)

13. Goal #4: Develop Equivalent Alternate Forms
Two forms (Blue and Red) were developed for each of the 3 versions of the MMSE-2
Based on the results of the equating study, the accuracy of the equating process was confirmed

14. Equating: MMSE & MMSE-2:SV

15. Administration Issues
18 years and older
Relatively easy to administer, typically one training session is sufficient
Test Materials:
User’s Manual
Pocket Norms Guide
Scoring Templates for Processing Speed
Administration Forms :
MMSE-2:BV Blue and Red Form
MMSE-2:SV Blue and Red Form
MMSE-2:EV Blue and Red Form
All instructions are on the form. There is not a form with just EV so can’t add it on to the SV. Must order the EV forms which include it all.

16. Overview of Administration Forms
Task
MMSE
MMSE-2:BV
MMSE-2:SV
MMSE-2:EV
Registration & Recall X
Orientation to Time
Orientation to Place
Attention & Calculation (Serial 7s)
Naming
Repetition
Comprehension
Reading
Writing
Drawing
Story Memory
Processing Speed

17. Determining Which Version is Appropriate
MMSE-2:BV
Adequate for screening large populations; screening individuals in practice who have not been referred because of cognitive complaints
MMSE-2:SV
Used first if referred because of complaint of cognitive decline or if patient indicates memory is not as good as it use to be; depending on results may want to supplement with MMSE-2:EV
MMSE-2:EV
Same as above + well educated (ceiling effect); suspected subcortical dementia

18. Scoring Mean raw total scores are presented by age and education level
T scores are also presented by age and education level
Pocket Guide
Reliable Change Scores
Although T scores are provided, it is the raw scores that are usually interpreted

19. Reliable Change Scores
Reliable change refers to the extent to which the change in test performance shown by an individual falls beyond the range that can be attributed to practice effects or to measurement variability that is inherent to the instrument itself
The approach used here is a method developed by Iverson (2001)

20. Interpretation
A cut score of 22/23 is typically used with the original MMSE
Because the MMSE-2:SV is equivalent to the MMSE, the same cut score is suggested
The authors have not provided specific recommendations for the new forms, however ranges of raw score cut scores are provided for the dementia, AD, and subcortical samples by form
22/23 is cutoff for Standard Version. If below this raw score, likely cognitive impairment.

21. Example of Cutoff Table
A good cutoff score will make the least sacrifices on both sensitivity and specificity. The 23/24 and 22/23 scores have almost identical SENS and SPEC scores and add to the validity of using this range as a cutoff score for dementia.

22. Development Task Development – 5 additional tasks were tested
Bias Panel – assessed potential bias and offensiveness to protected groups
Expert Review:
2 neuropsychologist, 1 geriatric psychologist, 1 geriatric psychiatrist
Aided in selection of tasks, provided feedback on content, and assisted with refining items and instructions for the pilot and standardization versions

23. Pilot Testing

24. Standardization Sample
n = 1,531 healthy controls
Hard to get low education samples since so many get post HS training
Geographic region is not as important in neuro, especially for adults

25. Cognitively Impaired Samples

26. Effects of Age and Education
In reality, EV is really a better predictor for both age and education. High age spread shows it but education does not.

27. Development of the Norms
Because of the importance of age and education on MMSE-2 scores norms were developed for several different age and education ranges
Two resources for age and education adjustments are provided:
Means and standard deviations of total raw scores by age and education groups
Age- and education-adjusted T scores (continuous norming method)

28. Reliability: Internal Consistency
Why are these only moderate?
The normals all have high scores so there is little variability
The test does not measure one trait per se. It is a variety of tasks with memory having the highest correlation with dementia.

29. Reliability: Test Retest
Higher scores from BV to SV to EV because there are progressively more items.

30. Reliability: Interrater

31. Validity
Content Validity – similar items on other tests (e.g., Serial 7s similar to attention and concentration task on WMS-III)
Intercorrelations among task and total scores (presented for both the normative and clinical samples)
Diagnostic Validity: Prior work on the MMSE using 22/23 or 23/24
There is an appendix in the manual with all the article references to validate a 22/23 and 23/34 cutoff score.
Also an appendix that documents studies looking at effect size by disorder.
Table: ADL is a 7 question daily living skills questionnaire. Low scores are more impaired.
All versions significantly differentiated severe (0-4) from mild (5-7). SV differentiated best.

32. Validity Convergent Validity: WMS-III subtests Category Naming Test
COWA
BNT
TMT
WAIS-R subtests
JOLO
HVLT-R
Stroop Color and Word Test