Cognitive Screening Instruments: A Practical Approach

A. J. Larner (ed.)Cognitive Screening Instruments2013A Practical Approach10.1007/978-1-4471-2452-8_2© Springer-Verlag London 2013

2. The Mini-Mental State Examination (MMSE): An Update on Its Diagnostic Validity for Cognitive Disorders

Alex J. Mitchell¹, ²  

(1)

Department of Psycho-oncology, Leicestershire Partnership Trust, Leicester, UK

(2)

Department of Cancer Studies and Molecular Medicine, University of Leicester, Leicester, UK

Alex J. Mitchell

Email: ajm80@le.ac.uk

2.1 Introduction: History and Development

2.2 Structure and Reliability of the MMSE

2.3 Diagnostic Validity in Unselected Dementia

2.4 Diagnostic Validity in Early Dementia

2.5 Diagnostic Validity in Specific Dementias

2.6 Diagnostic Validity in MCI

2.7 Diagnostic Validity in Delirium

2.8 Conclusion: Implementation

References

Abstract

The Mini-Mental State Examination (MMSE) is the most commonly used brief cognitive tool in the assessment of a variety of cognitive disorders. The tool comprises a short battery of 20 individual tests covering 11 domains and totalling 30 points. Typical completion time is 8 min in cognitively unimpaired individuals rising to 15 min in those with dementia. Internal consistency appears to be moderate and test-retest reliability good. However, the main psychometric issue concerns the MMSE’s diagnostic validity against dementia, mild cognitive impairment, and delirium. This chapter updates previous meta-analytic summary analyses for the performance of the MMSE in specialist and nonspecialist settings. Summary sensitivity, specificity, positive, and negative predictive values are presented. Results suggest against dementia, mild cognitive impairment, and delirium it did not perform well as a confirmatory (case-finding) tool, but it did perform adequately in a rule-out (screening) capacity. In clinical practice, this means that a high score on the MMSE would lead to about a 10 % false negative rate, and further, a low (positive) score must be followed by more extensive neuropsychological or clinical evaluation. The MMSE is neither the most accurate nor more efficient tool with which to evaluate cognitive disorders, but it has provided a benchmark against which all newer tools can be measured.

Keywords

Mini-Mental State Examination (MMSE)DementiaMild cognitive impairmentDeliriumDiagnostic accuracyReliabilitySensitivityClinical utility

2.1 Introduction: History and Development

The Mini-Mental State Examination (MMSE) was published in 1975 by Folstein et al. [1] as a practical method of grading cognitive impairment and has become the most commonly used rapid cognitive screening instrument [2]. While it is true that the MMSE may never have been intended as a diagnostic tool, it has been very extensively investigated as a diagnostic test of dementia and related cognitive disorders. Many are attracted by the brevity of the instrument and the belief that it offers broad coverage of cognitive domains. Its ubiquitous use was no doubt helped by its royalty free distribution up to 2001 when copyright was acquired by Psychological Assessment Resources (http://www.minimental.com/). In clinical practice, the main applications of the MMSE are to help clinicians in the diagnosis of dementia and delirium [3]. It has been investigated in a case-finding role (i.e. confirmatory diagnosis) as well as in a screening role (largely rule-out application designed to minimize false negatives). Recent work has also investigated its performance in detecting mild cognitive impairment (MCI). A subsidiary aim, not discussed here, is grading severity of cognitive impairment in those with known disorders [4]. It is worth noting that while a typical application time of 10 min seems short to neuropsychologists, many working in primary care would consider this much too long [5, 6].

At least 100 validation studies exist, but most are underpowered and many lack an adequate criterion standard and, hence, can give a misleading impression of accuracy [7]. For example, Folstein, Folstein, and McHugh validated the MMSE in two samples of patients which included only 38 with dementia [1]. However, it can be argued that even with suboptimal accuracy, the large evidence base surrounding the MMSE is advantageous because scores on the MMSE are fairly well understood by health professionals. This is most applicable to normative data. Folstein et al. [1] tested a population sample in Baltimore and found 4.2 % of those aged 18–64 scored <24/30 compared to 20.8 % of those over 65. Crum et al. [8] tested an extensive group of 18,056 participants in US Epidemiologic Catchment Area (ECA) study and presented distributions by age and educational levels. Some groups have provided norms for each item on the MMSE by age group [9]. Yet there remains controversy about its clinical applications in case finding and screening, as well as the optimal cut-off threshold [10, 11]. A cut-off of <24/30 was recommended as significant by Folstein and colleagues in persons with at least 8 years of education [1]. But in reality, individuals with early dementia but with a background of extensive education are likely to experience a ceiling effect with the MMSE (see Sect. 2.4 on early dementia). Numerous other cut-offs have been calculated from receiver operating characteristic curve (ROC) analysis of specific populations together with adjustments for age and education [12, 13]. Here, I will review the accuracy of the MMSE when considering one of the common cognitive disorders in clinical practice: dementia, mild cognitive impairment, and delirium.

2.2 Structure and Reliability of the MMSE

The MMSE has an internal structure of 20 individual tests covering 11 domains including orientation, registration, attention or calculation (serial sevens or spelling), recall, naming, repetition, comprehension (verbal and written), writing, and construction. Internal consistency appears to be moderate with Cronbach alpha scores reported between 0.6 to 0.9 [14, 15]. Test-retest reliability has been examined in several studies and in those where re-examination took place within 24-h reliability by Pearson correlation was usually above 0.85. Scoring emphasizes orientation (time – 5 points; place – 5 points) and attention/concentration/calculation (5 points) with lower emphasis on registration memory (3 points) and recall (3 points). Little weight is placed on naming (2 points), repetition (1 point), following a three-stage command (3 points), reading (1 point), writing (1 point), or copying intersecting pentagons (1 point). Factor-analytic and item-response studies suggest up to five factors [16, 17]. Using Rasch analysis, it is possible to grade the completion difficulty of each item on the MMSE. Relatively difficult items are the recall of three words, citing the correct date, coping the pentagon design, and spelling world backwards or completing serial sevens. Conversely, relatively simple items are naming the correct country, registering three words, following the command and naming an object (pencil).

A significant issue is that the individual questions are not particularly applied, which reduces acceptability of the test to those who suspect impairment. In other words, uptake of the test may be low in those with impairment. It is generally accepted that much of the content of the MMSE was derived from existing instruments [18]. All questions are designed to be asked in the order listed, with omissions scored as errors giving a maximum score of 30. However, there is some ambiguity in several items leading to the Standardized Mini-Mental State Examination from Molloy et al. [19] (see Sect. 3.2.1). The MMSE has helped in the development of newer potentially improved cognitive instruments discussed in other chapters (e.g. Chap. 4).

2.3 Diagnostic Validity in Unselected Dementia

This is probably the MMSE’s most common application and hence the most important question. Does the MMSE enable clinicians to accurately rule-in or rule-out dementia? Further, does this depend on prevalence of dementia, for example, when dementia is less common such as in primary care settings? O’Connor et al. [20] conducted one of the first adequately powered tests of the MMSE using a cut-off <24/30 in 2,302 primary care patients; 586 received a CAMDEX/CAMCOG interview as a gold standard (criterion reference). O’Connor et al. found that sensitivity of the MMSE was 86 % and specificity 92 % [20]. There have been at least eight other primary studies, and these documented somewhat differing results (see Table 2.1).

To clarify uncertainty, our group undertook a meta-analysis of MMSE dementia studies published prior to 2009 [21]. In the original paper, after excluding studies relying upon modified forms of the MMSE, as well as those focussing on specific sub-tests, there were 34 diagnostic validity studies against dementia (typically using DSM criteria, not robust post-mortem data). This is now updated to 45 studies (Table 2.1; [20, 22–64]) comprising 12 community studies, 7 primary care studies and 26 from specialist settings where the prevalence of dementia is relatively high. It is important to remember that the prevalence of a condition strongly influences test performance. High prevalence settings favor few false positives but at the expense of false negatives. Three studies were difficult to classify as they were conducted in the community but recruited from primary care lists. The most common reference standard in making a diagnosis of dementia was used in 20 studies. These results are now updated in Table 2.1 with the addition of eight new studies. A random effects meta-analysis model was used to calculate summary sensitivity, specificity, and PPV and NPV calculated using a prevalence of 25 %.

Table 2.1

Diagnostic validity of MMSE in diagnosing dementia

Updated from [21]

DSM-IV Diagnostic and Statistical Manual of Mental Disorders (DSM IV), CERAD Consortium to Establish a Registry for Alzheimer’s Disease, CDR Clinical Dementia Rating

Looking at specialist settings, meta-analysis showed that the MMSE’s sensitivity for diagnosing dementia was 76.9 % (95 % CI = 70.1–83.1 %) and its specificity was 89.9 % (95 % CI = 82.5–95.4 %). More meaningfully, that converts into a positive predictive value (PPV) of 89.3 % and a negative predictive value (NPV) of 74.8 % at a prevalence of 55 %, but 71.7 % (PPV) and 92.1 % (NPV) at 25 %. Thus, there would be a 25 % false negative rate for every 24 or above MMSE score in clinics where the prevalence of dementia was high, but only an 8 % error rate when prevalence was low. Clinical utility can be considered to be a function of both occurrence and discrimination of a test. MMSE may be suitable to be used as a first step screening tool in specialist