The relationship between the mini-mental state examination and cognitive functioning in normal elderly adults: a componential analysis

Age and Ageing, Sept, 1995 by Robert D. Hill, Lars Backman

Introduction

One of the most widely used screening instruments for the detection of cognitive impairment in older adults is the Mini-Mental State Examination (MMSE[1]). The MMSE is a 20-item instrument that is administered by a trained professional. It involves standardized questions that tap a range of cognitive abilities, including orientation for place and time, memory and attention, language skills, and visuospatial abilities. There are various scoring procedures for the MMSE but it is generally accepted that the instrument ranges from 0 to 30 points with a score below 24 indicating cognitive impairment [2]. Although it was originally designed to screen individuals with gross cognitive deficits, irrespective of disease, numerous references have appeared using the MMSE in a variety of contexts with both impaired and non-impaired older adult samples. For example, the MMSE has been used to measure progression of cognitive decline in cementing disorders [3-5], stage dementia severity [6], predict episodic memory performance in non-demented older adults [7, 8], assess the relationship of health and cognitive function in community dwelling elderly people [9], and predict the benefits of memory training in highly functioning older adults [10,11].

Several investigators have examined the degree to which specific items from the MMSE are differentially effective for predicting cognitive impairment in late life. These studies have attempted to reduce the number of questions and/or group items on the basis of empirically derived content domains. Braekus et al. [12] administered the MMSE to a heterogeneous sample of older adults in Norway. MMSE scores ranged from 11 to 29, with 43% of the sample scoring above 24. Their factor analysis yielded two factors; namely, `recent memory' and `praxis and orientation to place'. This factor structure, however, was qualified given that a number of items loaded on both factors making it difficult to distinguish different content domains (e.g. memory function vs. motor function).

In two reports examining a longitudinal sample of Alzheimer's disease (AD) patients with a mean MMSE score of 18.1 at entry [4, 5] two factor structures were identified; one related to a single point of measurement (static MMSE factor scores) and the other associated with change over time (dynamic MMSE factor scores). The dynamic factor scores were associated with increases in caregiver-reported memory and behaviour problems; however, the direction of these associations varied and was not easily interpretable. The authors concluded that although MMSE total score appears to be a global indicator of rate of cognitive decline in AD, considerable individual variation occurs for specific items within the MMSE in relation to disease severity and progression.

In normal elderly populations, no studies have examined the degree to which specific MMSE items are related to cognitive performance, although the MMSE total score has been used successfully to predict performance on a range of cognitive skills, including episodic memory and visuospatial performance. The fact that the MMSE is a potent predictor of cognitive performance in normal ageing in spite of its restricted range has been noted by several investigators [7, 8, 10, 11], although the nature of this range restriction has not been systematically explored.

The purpose of the current investigation is to examine the degree to which specific MMSE items differentially contribute to the prediction of performance on cognitive tasks in a population-based sample of very old adults in Sweden. It was hypothesized that, for normal older adults, the pattern of errors on the MMSE may be substantially different from that of subjects with cognitive impairment. For example, the probability of erring with regard to orientation for place and time, registering simple words, or identifying common physical objects may be relatively low in non-demented community-dwelling elderly adults; whereas variability may be greatest for tasks with higher cognitive demands (e.g., episodic memory, reproducing a geometric design). Thus, it is conceivable that the MMSE may be construed less as a measure of global cognitive function and more as an indicator of specific cognitive skill deficits of normal elderly adults.

Within the current study, three principal questions were addressed: (1) Are there differences in patterns of errors across specific MMSE items? (2) Do MMSE items have sufficient variability to cluster into empirically meaningful content domains when the range is restricted? and (3) Are these domains differentially related to performance on specific cognitive skills, namely, episodic memory and visuospatial performance.

Method

The sample was taken from all the inhabitants in the Kungsholmen parish of Stockholm, Sweden, aged 75 years and older who were included in a population survey on ageing and dementia. This longitudinal study was designed to investigate social and psychological issues associated with normal ageing and dementia. A more detailed description of the population and methods used has been reported elsewhere [13, 14]. Briefly, two phases of the study have been completed: In Phase I, 18 l 0 individuals from the whole study population were administered a questionnaire, which included the MMSE. In Phase II, all those with an MMSE score below 24 (n = 314) and a random sample of those with an MMSE score above 23 (n = 354) were assessed with a complete medical examination, social and family interviews, laboratory blood analyses, and a comprehensive cognitive test battery. The Swedish version of the MMSE was administered by trained nurses according to standardized procedures. If individual items were refused, or if they could not be completed because of physical disability, zero scores were given. A full description of the procedures and scoring rules are available in another report [15].