Score Disparity Between the MMSE and the SLUMS

Abstract

Background:

Multiple brief cognitive instruments are available to assess cognitive impairment in older adults. However, not all instruments demonstrate the same effectiveness when utilized with higher educated adults. This study evaluates the score disparity between the Mini-Mental State Examination (MMSE) and the St. Louis University Mental Status (SLUMS) Examination across the education spectrum. It was hypothesized that individuals with more years of formal education would produce higher scores on both the MMSE and SLUMS. Previous research was conducted to create a conversion scale used to compare and convert the MMSE scores to SLUMS scores. This research study provides additional data to add to the body of knowledge regarding a conversion scale for the MMSE and SLUMS.

Methods:

Seventy-five adults older than the age of 60 years were each administered the MMSE and SLUMS.

Results:

Contrary to our hypothesis, individuals with more years of formal education did not produce significantly greater scores on the MMSE or SLUMS. Likewise, education level analyzed as a continuous measure was not significantly correlated with the MMSE, r(75) = −0.191, or SLUMS, r(75) = 0.019. Interestingly, among participants with a high (but not low) education level, there was a marginal but significant difference in mean score between the MMSE (29.00 ± 1.47) and SLUMS (27.74 ± 3.08), t(64) = 3.70, P < .001.

Conclusion:

Other factors besides education may impact the performance of older adults on the MMSE and SLUMS, but it does appear that education level may moderate the score disparity between the 2 instruments. Additional studies are needed before using the MMSE to predict the score on the SLUMS and vice versa.

Introduction

Nearly 5.4 million Americans have been diagnosed with Alzheimer’s disease, and it is the sixth leading cause of death in the United States. ¹ The prevalence of dementia and Alzheimer’s disease will increase as the population of adults older than the age of 65 years continues to grow. ¹ The percentage of adults older than the age of 65 is estimated to reach 16% by the year 2020 in the United States. ² The cohort of adults born between the years 1946 and 1964, commonly known as Baby Boomers, will comprise the majority of this population. ² Unlike previous generations, Baby Boomers are expected to seek diagnosis and treatment of dementia more frequently. ² Baby Boomers are comfortable with utilizing technology to obtain medical information and are more apt to report mental issues to healthcare providers. ² In addition, advances in research have brought greater awareness to the general public about dementia and Alzheimer’s disease. ² All of these factors are expected to raise the anxiety level of Baby Boomers about developing dementia, thus preempting them to contact their healthcare providers when cognitive issues arise. ² Although treatment options have been devised to manage symptoms and slow the progression of the disease, there is no cure for dementia and Alzheimer’s disease. ¹ Furthermore, a diagnosis of dementia or Alzheimer’s can lead to psychological, physical, and financial strain on both the individual and his or her family. ¹

To better meet the needs of this growing population, it will be necessary for healthcare professionals to have access to fast and reliable cognitive screening assessments. ³ Brief cognitive screening assessments are designed to provide a quick evaluation of an individual’s cognitive status to determine if further clinical testing is necessary to diagnose memory impairment. ⁴ The assessments can provide physicians, psychologists, nurses, social workers, and other healthcare professionals with early warning signs of cognitive impairment, such as mild neurocognitive disorder (mNCD). The mNCD is a stage of cognitive decline experienced before a clinical diagnosis of dementia or Alzheimer’s disease occurs. ^5,6

One of the earliest and most widely used brief cognitive assessment tools is the Mini-Mental State Examination (MMSE). ⁷ Developed in 1975, the MMSE is used to assess cognitive impairment in the older adult population. ⁸ Given that older patients with symptoms of delirium and dementia often do not have long attention spans, Folstein et al ⁸ designed the MMSE to take a minimal amount of time to administer. The MMSE has also been used to assess other populations, including individuals with Parkinson disease, depression, and multiple sclerosis, as well as stroke survivors. ⁸

Researchers have acknowledged that the MMSE is an effective screening tool for dementia, but limitations exist when it comes to assessing individuals with higher levels of education who may have mNCD. ^9,10 Individuals with higher levels of education can continue to score within the normal cognitive range on the MMSE despite the looming diagnosis of dementia. ⁹ While it is atypical, individuals who have been diagnosed with dementia have scored a 30/30 on the MMSE. ¹¹ One explanation for this is the cognitive reserve hypothesis. According to the hypothesis, higher levels of cognitive reserve may serve as a preventative cushion for the onset of cognitive impairment, especially during periods of mNCD. ^{12 -14} Cognitive reserve can accumulate across the life span as a result of a variety of factors with educational attainment being one of them. ¹⁵ A study conducted by Roe et al ¹⁶ found that each additional year of formal education that was obtained by the participants decreased their chances of receiving a clinical diagnosis of dementia or Alzheimer’s disease. The study also indicated that individuals with higher levels of education and a neuropathological diagnosis of Alzheimer’s disease were more likely to score higher on the MMSE 1 year prior to death. ¹⁶

With the educational bias in mind, O’Bryant et al ¹⁷ conducted a study to establish norm scores for the MMSE for highly educated individuals with a minimum of 16 years of education. Participants varied in gender, ethnicity, and illness such as dementia and mNCD. ¹⁷ They concluded that a score of anything below 27 points for individuals with at least a college degree was an indicator of an increased risk for being diagnosed with dementia or Alzheimer’s disease. ¹⁷ Although norm scores for the MMSE have been identified for educational levels, the scoring norms are not identified on the copyrighted version of the MMSE. A limitation to the educational norms for the MMSE is that not all clinicians are aware of, or use, the norms that have been established. ⁹ Despite the documented limitations, the MMSE is commonly used worldwide to assess cognitive impairment. ⁷ Even now that the MMSE is copyrighted, the familiarity and ease of administration makes it a popular data collection instrument for research studies and for the use of healthcare professionals. ¹⁷

An individual with mNCD may be independent with skills such as driving, cooking, managing finances, and personal care, despite some noticeable cognitive impairment. ⁶ Individuals with mNCD, who experience memory problems, are at a higher risk for developing dementia and Alzheimer’s disease. ¹⁸ For example, individuals with mNCD are 12% to 15% likely to progress to a diagnosis of Alzheimer’s disease versus the 1% to 2% of individuals without mNCD. ¹⁸ Additionally, progression rates differ among individuals living in the community and in clinical settings. ¹⁸ Evidence suggests that individuals with mNCD in clinical settings progress to dementia at a 10% to 15% annual rate, and community-dwelling individuals progress at a lower annual rate of 3% to 4%. ¹⁸ Despite the numerous cognitive screening assessments available to quickly screen patients for cognitive impairment, not all assessments produce accurate results for patients who are in the very early stages of the disease process. ³ With this limitation in mind, new cognitive assessments have been created and have proven to be more accurate for detecting mNCD, or predementia symptoms. ^3,19

The St. Louis University Mental Status (SLUMS) Examination was designed to compensate for several of the limitations found with the MMSE: mainly, the ability to detect mNCD in individuals with higher levels of education. ¹⁹ The original research study introducing and validating the SLUMS was a comparison study with the MMSE and included a sample of veterans older than the age of 60 years who were mostly male. ¹⁹ The SLUMS was later studied with a more diverse sample of nonveteran older adults to further test the validity. ²⁰

Similar to the MMSE, the SLUMS was created to assess individuals’ orientation, memory, and attention span, but it uses different tasks. ¹⁹ In contrast to the MMSE, the SLUMS was designed to be more sensitive to detecting deficits in executive functioning, an area that is commonly affected in the early stages of dementia and mNCD. ¹⁹ Executive functioning involves a myriad of cognitive processes that includes regulation of emotions, processing, planning, and organizing information to enable the performance of everyday tasks. ²¹ The SLUMS uses a clock drawing and animal recall task to assess executive functioning, both of which have been incorporated to help identify mNCD. ¹⁹ To assist the clinician in scoring the SLUMS, a table is located at the bottom of the assessment instrument. The table provides cutoff scores for educational attainment (high school and less than high school education) as well as cognitive functioning (normal, mNCD, and dementia). ¹⁹ Since the SLUMS has better sensitivity for detecting mNCD, the within-patient scores differ when paired with the MMSE. ^3,19 The difference in mean scores have been compared in other studies. ^3,20 The other samples included nonveteran community-dwelling older adults and individuals residing in long-term care facilities. ^3,20 Although the samples differ, the results are similar in that scores on the MMSE tend to be higher when compared with the scores on the SLUMS. ^3,20

Buckingham et al ³ conducted a study to create a conversion scale that could be used to compare and convert the scores of the MMSE and the SLUMS. The purpose of the conversion scale is to take old MMSE scores and convert them to SLUMS scores without having to assess the patient again. ³ Unlike the MMSE, the SLUMS has a scoring scale that provides cutoff scores for mNCD and dementia based on educational attainment. ¹⁹ By converting old MMSE scores to SLUMS scores, health-care professionals will be able to differentiate between normal cognitive functioning, mNCD, and dementia. ³ Furthermore, health-care professionals may find the SLUMS a more attractive assessment tool for 2 reasons. The SLUMS has higher validity for detecting mNCD, and it is free to use. ¹⁹ The MMSE is copyrighted and costs over a dollar per assessment.

To date, the Buckingham et al’s ³ study has been the only one conducted to determine a predictive change score between the MMSE and the SLUMS for the purposes of creating a conversion scale. It was determined that on average participants scored 4.56 points higher on the MMSE, compared with the SLUMS. ³ It was acknowledged that the score difference may be influenced by cognitive reserve, the ability of the brain to continue working efficiently despite the presence of brain damage. ³ Individuals with higher levels of cognitive reserve may be able to employ mental strategies that enable them to score well on cognitive assessments that are less sensitive to mNCD, such as the MMSE. ³ Cognitive reserve can be obtained throughout the life span from experiences such as formal education, exercise, stimulating occupations, participation in leisure activities, and social networks. Buckingham et al ³ suggested that individuals with higher levels of education might have a larger mean difference score between the 2 assessments. Importantly, they examined the average MMSE and SLUMS scores as a function of their living environment (independent living, assisted living, skilled nursing, and other) and not by education. ³

The present study has 2 aims. The first is to validate the MMSE–SLUMS conversion factor proposed by Buckingham et al ³ in a more heterogeneous population including community-dwelling older adults. Second, the effect of education level on the score disparity between the MMSE and SLUMS will be evaluated. In the present study, it was hypothesized that individuals with higher levels of education would score higher on both the MMSE and the SLUMS when compared to individuals with lower levels of education. For the purpose of this study, formal education was measured as a means to investigate the participants’ cognitive reserve.

Methods

Participants

This study was approved by the Longwood University and Nova Southeastern institutional review boards. The participants were recruited from 4 primary research sites in a small rural town in central Virginia, including a public liberal arts university, a recreation center, a continuing care retirement community, and the local area agency on aging. Convenience and snowball sampling methods were used to recruit participants aged 60 years and older without severe visual or hearing impairments or a known diagnosis of dementia or Alzheimer’s disease. No participants reported a known diagnosis of dementia or Alzheimer’s disease. Examples of severe visual or hearing impairments included the inability to read large font, hear questions, write, or verbalize answers.

Instruments

The MMSE and SLUMS were employed to assess cognitive function. Copyrighted versions of the MMSE were purchased from Psychological Assessment Resourced (2001). There was no charge to use the SLUMS, and permission was granted from St. Louis University. Both the assessments use a scoring scale totaling 30 points. The MMSE contains 11 questions and is divided into 2 sections and takes approximately 5 to 7 minutes to administer. ⁸ The first section requires the participant to respond verbally to questions focusing on memory, attention, and orientation. ⁸ The second section assesses the patient’s ability to follow verbal and written commands and involves writing a sentence and drawing a polygon figure for a total of 9 points. ⁸ A score between 24 and 30 points is considered normal cognitive functioning for an individual with at least 8 years of formal education; a score of 19 to 23 is borderline cognitive impairment, and anything below a score of 19 is an indicator of cognitive impairment. ²² The validity of the MMSE as an assessment of cognitive functioning was verified in the original study in 1975. ⁸ When using the cutoff score of 24 for normal cognitive functioning, the MMSE has a sensitivity of 87% and specificity of 82%. ⁶ In the original study, the MMSE was shown to be reliable when administered 24 hours later and 28 days following the initial assessment. ⁸ When the same examiner administered the MMSE 24 hours later, there was a significant correlation between the 2 scores, r(22) = 0.887, P < .0001. ⁸ Additionally, a sample of patients with a combination of dementia and depression was given the MMSE 28 days following the initial assessment; the second set of scores showed very little difference from the first set of scores as represented by r(23) = 0.988, P < .0001. ⁸ In conclusion, the MMSE is a widely used tool for research related to cognitive functioning, and there are nearly 70 studies that have been conducted to validate it. ²³ However, many of the samples in the studies lack power and therefore may give a misrepresentation of the level of accuracy. ²³

The SLUMS is also comprised of 11 questions and takes on average 7 minutes to administer. ¹⁹ The assessment includes questions related to orientation, memory, number calculations, immediate and long-term recall, figure recognition, and executive functioning. ¹⁸ Cutoff scores are based on the educational level of the participant. For individuals with less than a high school education, the cutoff score for normal cognitive functioning is 25 to 30, the nMCD cutoff score is 20 to 24, and the dementia cutoff score is 1 to 19. ¹⁹ For individuals with a high school education or higher, the cutoff score for normal functioning is 27 to 30, the nMCD cutoff score is 21 to 26, and the dementia cutoff score is 1 to 20. ¹⁹ The SLUMS was validated in the original study by Tariq et al ¹⁹ by comparing the SLUMS and the MMSE in a sample of older adult veterans. The results of the study indicated that both the MMSE and SLUMS are similar in regard to detecting dementia; however, the SLUMS is more effective than the MMSE in detecting mNCD from normal cognitive functioning. ¹⁹ The accuracy of the SLUMS was determined by examining the area underneath the curve of the receiver operator curves when examining 2 groups: participants with less than a high school education with nMCD and dementia and participants with a high school education with nMCD and participants with dementia. ¹⁹ In the original study, the SLUMS was able to outperform the MMSE in terms of identifying participants with dementia and mNCD. ¹⁹ The SLUMS was able to identify 98.3% to 99.8% of participants with dementia and 92.7% to 94.1% of participants with mNCD. ²⁴ The SLUMS was also validated using a sample of nonveteran community-dwelling older adults. ²⁰ Finally, a limitation of the SLUMs is that there is little psychometric data related to reliability. ²⁴ There is no data to support test–retest reliability or interrater–interrater reliability. ²⁴

A supplementary demographic questionnaire was also utilized to obtain age, gender, ethnicity, uncorrected sensory impairments such as visual and hearing, and diagnosis of dementia or Alzheimer’s disease. Dichotomous markers, or yes and no response options, were utilized for the questions pertaining to uncorrected sensory impairments and the diagnosis of dementia or Alzheimer’s disease.

Study Design

Following informed consent and completion of the demographics questionnaire, participants completed the MMSE and SLUMS. When administering the MMSE and SLUMS, the order was counterbalanced with half of the participants taking the MMSE before the SLUMS. All tests were administered in a private room to decrease test anxiety and distractions. Additionally, the investigator followed several general rules. First, each participant was asked if all hearing and visual devices or aids were in place prior to the start of testing. Next the investigator recited, “I am going to ask you several questions and give you some problems to solve. Please answer them to the best of your ability.” Questions were not repeated more than 3 times, and questions were read exactly as stated on the paper version of the MMSE and SLUMS. The participants were not provided with any hints to encourage correct answers nor was any feedback given about performance between the assessments. The entire process of collection of consent forms, demographic information, and cognitive assessments took 20 to 30 minutes per participant.

Data Analysis

Statistical analyses were conducted using SPSS version 22.0. The data collected from the demographic questionnaire were summarized using descriptive statistics. Paired samples t-tests were utilized to test for an overall difference in mean scores between the MMSE and SLUMS. Additionally, a Pearson product-moment correlation test was used to investigate the correlation between MMSE scores and SLUMS scores. Inspection of the plotted relationship between MMSE and SLUMS prompted a segmented Pearson product-moment correlation analysis between MMSE scores and SLUMS scores of 27 or less, and between MMSE scores and SLUMS scores of 28 or more.

Years of formal education was dichotomized to represent low (less than 12 years of formal education) and high (12 or more years of formal education). An independent samples t test was used to compare the mean difference in MMSE and SLUMS scores between participants with low and high education levels, and paired samples t-tests were used to compare the mean difference in MMSE and SLUMS scores within each education level. A Pearson product-moment correlation test was used to assess the association between years of formal education and cognitive assessment scores.

Results

Of the 75 participants, the mean age was 74.80 years of age (standard deviation [SD] = 7.61). Of all, 15 participants were African American and 60 were Caucasian. Additionally, the average years of formal education attained by the sample was 15.53 years (SD = 3.91). Of the 75 participants, 65 were in the high education category. Table 1 shows the demographic information organized by educational level.

Table 1.

Demographics and Cognitive Assessment Scores of a Function of Years of Formal Education.^a

	Low (n = 10)		High (n = 65)
	Mean	SD	Mean	SD
Demographics
Age	74.80	7.60	71.90	9.00
Gender	10 Females	0 Male	54 Females	11 Males
Race	4 African American	6 Caucasian	11 African American	54 Caucasian
Cognitive assessment scores
MMSE	29.60	0.52	29.00	1.47
SLUMS	28.50	1.18	27.74	3.08

Abbreviations: MMSE, Mini-Mental State Examination; SD, standard deviation; SLUMS, St. Louis University Mental Status.

^aLow indicates <12 years of formal education and high indicates ≥12 years of formal education.

The overall mean score among the entire sample on the MMSE (29.08 ± 1.39) was significantly greater than the mean score on the SLUMS (27.84 ± 2.90), t(74) = 3.89, P < .001. Although statistically significant, the 1.24 mean difference in the present study was rather marginal. Collectively, the MMSE scores were not significantly correlated with the SLUMS scores (Table 2). Interestingly, visual inspection of the correlation plot revealed a noticeable change in the relationship between MMSE and SLUMS at a SLUMS score of 27, which happens to be the cutoff score for normal functioning individuals with at least a high school education. Although there was no significant correlation between MMSE and SLUMS for the subset of participants scoring at least 28 on the SLUMS, a positive correlation was found for participants with SLUMS scores of 27 and below, r(20) = 0.639, P < .01.

Table 2.

Association Between Cognitive Assessment Scores.^a

	MMSE
	r	P Value
SLUMS	0.343	.161
SLUMS (score ≤27)	0.639	.002
SLUMS (score >27)	0.262	.054

Abbreviations: MMSE, Mini-Mental State Examination; SLUMS, St. Louis University Mental Status.

^a r is Pearson product-moment correlation.

Individuals with more years of formal education did not produce significantly greater scores on the MMSE or SLUMS (Table 1). Likewise, education level analyzed as a continuous measure was not significantly correlated with the MMSE, r(75) = −0.191, SLUMS, r(75) = 0.019, or the difference in score between the MMSE and SLUMS, (r(75) = −0.031; Table 3). Among participants with a high education level, there was a marginal but significant difference in mean score between the MMSE (29.00 ± 1.47) and SLUMS (27.74 ± 3.08), t(64) = 3.70, P < .001. There was not a significant difference between MMSE and SLUMS scores for the low educational-level group.

Table 3.

Association Between Years of Formal Education and Cognitive Assessment Scores.^a

	Years of Formal Education
	r	P Value
MMSE	−0.191	.100
SLUMS	0.019	.869
MMSE–SLUMS	−0.031	.794

Abbreviations: MMSE, Mini-Mental State Examination; SLUMS, St. Louis University Mental Status.

^a r is Pearson product-moment correlation.

Discussion

Score Disparity Between MMSE and SLUMS

The SLUMS was created to help offset the educational bias of the MMSE by using cutoff scores for varying educational levels and enhanced questions and tasks related to detecting impairment of executive functioning. ¹⁹ Because of this difference, researchers have suggested that the SLUMS is a superior tool when used to detect individuals with mNCD and possible cases of early dementia. ^3,19,20

Consistent with Buckingham et al ³ and Tariq et al, ¹⁹ the entire sample of participants in this study scored significantly lower on the SLUMS than on the MMSE. However, the score difference in this study of 1.24 is negligible and inconsistent when compared with the 4.56 difference previously reported by Buckingham et al. ³ In addition, the Tariq et al’s ¹⁹ study found an approximate 2.34 difference in scores between the MMSE and SLUMS for the 440 participants with normal cognitive functioning. The lack of a consistent score difference may argue against using a standardized conversion factor. There are 2 possible explanations for the variation in score disparities. First, the sample from the present study differed in gender and age ranges when compared with other studies. The present study produced a rather homogenous sample of older mostly women, which contrasts the primarily male sample in Tariq et al ¹⁹ and the broad age spectrum (41-96 years) in Buckingham et al. ³ Next, the majority of the participants in the present study were community-dwelling individuals, and there were no participants with a self-reported diagnosis of dementia or Alzheimer’s disease. Although the Tariq et al’s ¹⁹ study did include a portion of participants with diagnosed mNCD and dementia, the researchers excluded this data from the comparison and used only the data that was reported for the participants with normal cognitive functioning. The Buckingham et al’s ³ study did not indicate if any participants had diagnosed cognitive impairment; however, they did report that some participants were recruited from assisted living and skilled nursing facilities. It is widely known that cognitive impairment, including dementia, can lead to functional disability and thus increase the likelihood for being placed in a long-term care facility. ¹⁸ This factor alone could have resulted in the greater difference in scores that was observed between the MMSE and SLUMS in the Buckingham et al’s ³ study.

Interestingly, a greater difference between scores on the MMSE and SLUMS was detected when participants scored a 27 or lower on the SLUMS. A score of 27 on the SLUMS is the cutoff for normal cognitive functioning for individuals with at least a high school education. Although it is beyond our scope of practice to diagnose mNCD, since the SLUMS is more sensitive to detecting impairments in executive functioning, it is within reason to suggest participants with cognitive impairment may have scored lower on the SLUMS and contributed to the greater observed variation. ^19,21 This greater variation in scores between the MMSE and SLUMS is more in line with the findings from Tariq et al ¹⁹ and Feliciano et al ²⁰ when assessing individuals with mNCD. In further support, a recent study by Stewart et al ²⁴ concluded that several of the participants’ cognitive impairment was not detected on the MMSE; however, they did score within a range indicating mNCD or dementia on the SLUMS and the other brief cognitive assessment tool. It is important to note that the sample in the Stewart et al’s ²⁴ study differed from the present study in that all of the participants resided in a long-term care facility and had cognitive impairments.

Education as a Moderator of MMSE and SLUMS Scores

It was suggested that MMSE and SLUMS would be directly related to the level of formal education in that the participants in the high education group would produce better scores on both the MMSE and SLUMS when compared to the low education group. ³ The findings from the present research are not consistent with Buckingham et al ³ and Tariq et al, ¹⁹ where participants with more years of formal education scored higher on both the MMSE and SLUMS. Additionally, educational level was not significantly correlated with the MMSE, SLUMS, or the difference found between 2 sets of scores. Finally, there was not a significant difference between the scores on the MMSE and SLUMS for the low education group. However, our sample of participants with less than 12 years of formal education was small (n = 10), compared with the 216 participants studied by Tariq et al, ¹⁹ a potentially more representative sample.

One explanation for the lack of difference of scores among the 2 educational groups is that there were no reports of dementia or Alzheimer’s disease among the participants. It is not impossible for an individual with less than 12 years of formal education to score very well or even 30/30 on both the MMSE and SLUMS. The research suggests that formal education serves as a protective factor for cognition and therefore can vary from individual to individual. ^{12 -15} Also, although formal education has been documented as a means to accumulate cognitive reserve, it is not the only method of building one’s cognitive reserve. ^{12 -15} Cognitive reserve can accumulate across the life span as a result of engaging in aerobic exercise, leisure activities, socializing, obtaining formal education, and working in stimulating environments. ^{12 -15} Therefore, it is possible that the low education group in this study built up their cognitive reserve utilizing other means over the life span by remaining physically and socially active, and engaging in cognitively stimulating jobs or activities. The accumulation of cognitive reserve over one’s lifetime may be a contributing factor why there was little variability between the mean scores for the 2 groups of participants.

Limitations

There are a few limitations worthy of mention. First, despite attempts to recruit a diverse sample of individuals, our study population was a rather homogenous sample of mostly well-educated Caucasian women, so the results may not be generalizable. Second, the low education group was relatively small, so caution should be used when making assertions about education level as a possible moderator in the MMSE to SLUMS scores disparity. Third, participants were administered both the MMSE and SLUMS, one right after the other, on the same day. Additionally, no distractor tasks were implemented between assessments, and therefore a practice effect may have been introduced.

Finally, and potentially most significant, convenience sampling may have introduced bias into the sample. Participants with self-reported memory complaints may have been nervous about taking a cognitive assessment and therefore may have not volunteered for the research study. ²⁵ During the informational sessions in the recruitment phase, many individuals expressed their reluctance to take the cognitive assessments by verbally voicing a fear of not doing well on the assessments. As a result, this may have introduced threats to external validity, with the sample not providing an accurate representation of the population. ²⁶

Conclusion

In support of our hypothesis, MMSE scores were significantly greater than SLUMS scores across all education levels, though the difference was marginal compared with previous studies. Importantly, a greater score disparity was observed between the MMSE and SLUMS for participants who scored 27 or less on the SLUMS, which suggests the SLUMS may be more sensitive to identifying cognitive impairment. Still, the lack of consistency in the MMSE and SLUMS scores disparity between studies indicates it is premature to use the score on one to predict the score on the other.

It is the authors’ recommendation that several other factors besides educational level be investigated before validating a conversion scale for the MMSE and SLUMS. Cognitive reserve may have played a role in the lack of score disparity observed between the high and low educational groups in this study. In future studies, it is recommended to obtain data on other possible sources of cognitive reserve (activity engagement, occupational attainment, etc) that participants have accumulated to further investigate the effect of cognitive reserve on cognitive performance. Furthermore, individuals with mental health disorders such as depression, anxiety, and substance abuse disorders may present with cognitive impairment. ² Therefore, investigating the correlation between mental health disorders and cognition may also shed more light on how these variables may impact MMSE–SLUMS discrepancy. Due to the numerous factors that can impact cognition in older age and given the accessibility and ease of use of both instruments, we recommend administering both until further research supports a valid conversion scale for the MMSE and SLUMS.