The role of culture and literacy in screening for major neurocognitive disorders for elderly patients on Reunion Island: A cross-sectional study on the concordance between the French Mini-Mental State Examination and an adapted version (MMS-Run)

Abstract

Background

On Reunion Island, literacy rates are estimated to be 62% for the elderly population. Education is a widely recognized protective factor for major neurocognitive disorders (MNCDs), including Alzheimer's disease, however during cognitive assessments by healthcare professionals, it can also act as a barrier.

Objective

This study aims to evaluate the concordance of the scores between the French version of the Mini-Mental State Examination (GRECO-MMSE) and an adapted version to the cultural context and literacy levels of the elderly Reunionese population (MMS-Run).

Methods

A cross-sectional, single-center study was conducted at the University Hospital of Reunion Island from April 14, 2023 to April 12, 2024. The GRECO-MMSE and MMS-RUN were performed one day apart to elderly participants who were hospitalized in the Department of Physical Medicine and Rehabilitation (PM&R) and the Department of Geriatrics. Analyses of concordance, correlation and distribution of scores as well as analyses by cognitive domains and by literacy levels (illiterate or literate) were carried out.

Results

Among the 61 participants, a discordance between the GRECO-MMSE and MSS-Run was observed (kappa=0.54), especially in the cognitive domains assessing orientation, attention and language (items that benefited most from adaptations). More than half of the participants were self-reported as illiterate (57%), and illiteracy was associated with increased discordance between the two tests.

Conclusions

The MMS-Run could be better suited for MNCD screening in the elderly Reunionese population. A validation study is necessary to determine appropriate threshold scores in the future.

Keywords

Alzheimer’s disease dementia geriatrics literacy mental status and dementia test neurocognitive disorders neuropsychological tests

Introduction

As populations around the world are aging, the prevalence of major neurocognitive disorders (MNCDs) is rising, and healthcare systems must strengthen their capacity to address this growing challenge.¹ Early screening for MNCDs enables timely support for the patient and their caregivers in establishing a safe and sustainable environment. This helps reduce complications related to loss of autonomy (falls, medication errors, malnutrition, infections) and, consequently, limits hospitalizations and institutionalization.^2,3 Finally, access to emerging treatments requires initiation at the early stage of the disease.⁴

MNCDs are mainly represented by Alzheimer's disease, vascular dementia and mixed dementia.⁵ They share common risk factors, such as age, history of head trauma, or genetic predisposition (specific to each etiology).^6–8 Conversely, level of education is a protective factor against their development. Individuals with a secondary level education are up to seven times less likely to develop an MNCD than those that are illiterate.^9,10 Moreover, studies have shown that this reduction in risk is not solely due to better performance on neuropsychological tests linked to reading, writing or arithmetic skills. Even on tests that do not require such skills, scores of individuals with the lowest education levels remain low.¹¹ In other words, education can contribute to the development of a cognitive reserve that delays the onset of disabilities.¹²

Neuropsychological tests cannot simply be translated from one culture to another without appropriate adaptation. Even among countries that share the same language, differences in test performance have been observed, partly due to variations in vocabulary use and cultural familiarity.¹³ Word frequency plays a critical role, as the most common words are more easily memorized, whereas rare words attract greater attention. Ultimately, words of medium frequency tend to be the most difficult to retain.¹⁴ Consequently, careful lexical selection is essential to ensure both comprehension and the validity of memory assessment. Knowledge, beliefs, and values also influence information processing, as the same piece of information activates different brain areas depending on an individual’s cultural background.¹⁵ Spatial and temporal orientation further depend on societal norms and systems, such as the use of landmarks instead of postal addresses or the reliance on different calendrical frameworks.¹⁶ Moreover, the reference framework used may affect the speed of information processing.¹⁷ Thus, if neuropsychological tests are adapted primarily to a majority cultural group, minority groups may be disadvantaged, with a risk of underestimating their health issues and limiting their access to care.¹⁸

Reunion Island, a French overseas department and region in the Indian Ocean, has a population derived from European colonization, African slavery and Asian migration.¹⁹ This diversity is reflected in the vernacular language, known as Reunionese Creole. The island's geographical isolation has also contributed to cultural differences with mainland France. Its education system, consistently developed only after 1946, limited access to generations over the age of 65 years due to insufficient school infrastructure.¹⁹

The effects of limited access to education on Reunion Island is evident to this day. In 2007, based on the Information and Daily Life survey, illiteracy rates in the region were three times higher among 18 to 65 year-olds (22.6%) than on mainland France (7.4%), and reached almost two-fifths of the elderly population (38.2%).²⁰ At the same time, this region is characterized by its young population, experiencing an increasing aging trend, with 25% of the population expected to be over 60 years old by 2050²¹ compared to 18.7% in 2021.²² This may suggest that there could be a rise in MNCD cases on the island over the next decade and screening should be focused on for early treatment.

To date, the most commonly used screening test for MNCDs is the Mini-Mental State Examination (MMSE) which was developed in the United States, with a population considered to have a generally high level of education.²³ The MMSE mostly assesses cognitive functions such as spatiotemporal orientation, memory and language, which are the main functions impaired by Alzheimer's disease.²⁴ Because of its simplicity, speed of administration and validity, the MMSE has become one of the most globally used screening tests for MNCDs.²⁵

Interpretation thresholds for the MMSE vary according to age and level of education.²⁶ In France, the most widely used version of the MMSE is the consensus translation of the Groupe de Réflexion sur les Evaluations COgnitives (GRECO),²⁷ whose interpretation thresholds have been validated on the French metropolitan population and adapted to the education standards of the country.²⁸ However, due to the cultural differences as well as literacy and arithmetic difficulties of the Reunionese population, the MMSE may not be well adapted. Healthcare professionals (HCPs) who carry out cognitive assessments (including neurologists, geriatricians, physical medicine and rehabilitation (PM&R) physicians, general practitioners and neuropsychologists) have noted that lower test scores on Reunion Island, may not necessarily reflect actual cognitive impairment, and could affect the sensitivity and specificity of the GRECO version of the MMSE when carried out on Reunion Island natives.

Therefore, an adapted version to the socio-cultural and educational context of the island’s population (known as MMS-Run) was developed. Although it has been widely adopted in various hospital departments on Reunion Island, it has never undergone any formal validation process. Its routine use in real-world settings raises questions about its interpretation and diagnostic quality. The objective of this study was to assess the concordance between the GRECO-MMSE and MMS-Run for patients aged over 65 years who received a cognitive assessment during routine care on Reunion Island.

Methods

Study design and population

A cross-sectional, single-center study was conducted at the University Hospital of Reunion Island from April 14, 2023 to April 12, 2024. Participants were recruited using a convenience sampling approach, consecutively including all eligible patients seen during the investigator’s on-site days. The GRECO-MMSE and MMS-Run tests were administered to each participant by the same HCP practicing either in the Department of Geriatrics or in the PM&R Department of the hospital. GRECO-MMSE was always examined prior to the MMS-Run. The two tests were administered one day apart, in order to minimize the learning effect which occurs when the tests are taken too close together and to limit the risk of changes in cognitive status when the tests are administered too far apart. This choice is supported by Tombaugh’s review, which reported similar MMSE reliability coefficients across studies with test–retest intervals ranging from one day to several weeks.²⁹ In this study, no clinical decision was made on the basis of a lower score obtained on the MMS-Run because of a lack of scientific validation.

Inclusion/exclusion criteria

Participants were included if they were native, continuous residents, over 65 years old and because they had received a cognitive assessment as part of their usual healthcare during hospitalization. Exclusion criteria were the factors that could alter the stability of the cognitive performance of patients during the time between the two tests: i) hearing or visual problems, ii) Glasgow score<15, iii) history of psychiatric pathology with psychosis (altered perception of reality), and iv) active addiction to a psychoactive substance with the exception of tobacco. Participants deprived of their liberty by judicial or administrative decision or those subject to a legal protection measure were also excluded.

MMS-Run assessment

In 2000, a Reunionese version of the MMSE called MMS-Run was developed (Supplemental Material). This examination is a French-language adaptation that takes into account the specific cultural features of the island’s native population. The MMS-Run is based on the original MMSE, the GRECO version and the findings of the Hindi Mental State Examination, a version of the MMSE adapted for the illiterate population in India.¹⁶ Moreover, the MMS-Run differs from GRECO-MMSE by the replacement of tasks requiring reading and writing with an imitation test and an oral expression test, respectively. Mathematical calculations are simplified.

The cultural aspects of the test consist of a question regarding the current season on Reunion Island with a binary response format (dry or rainy season). Questions related to territorial entity/geography are replaced by asking for the name of the island (Reunion) and its neighboring island (Mauritius). The notion of floor is replaced by a local binary altimetric notion (‘highs and lows’). The three words to remember are ‘mango’, ‘key’, and ‘hat’. Additionally, the phrase to repeat is in Reunionese Creole. Scoring is identical, with a total of 30 possible points: 10 points for spatiotemporal orientation, 3 points for learning, 5 points for attention, 3 points for recall, 8 points for language and 1 point for constructional apraxia (Supplemental Material).

For both the MMSE-GRECO and the MMS-Run, if a participant was unable to answer, declined to answer, or answered an item incorrectly, the item was scored as 0. Total test scores were calculated as the sum of all items, and all participants were retained in the analysis, regardless of the number of items they responded to.

Data collection

Baseline characteristics including socio-demographic data such as age, sex and hospital department visit were collected. The reasons for testing were grouped into the following categories: i) screening for MNCDs; ii) follow-up for an MNCD, iii) cognitive assessment following cerebrovascular disease (although testing for other brain disorders was considered, no participant required such assessment). Included participants were classified as literate or illiterate according to a self-reported assessment concerning their reading, writing or arithmetic skills. The Sociocultural level (SCL) of participants was determined according to a self-administered questionnaire, following the criteria defined for the GRECO-MMSE in which: i) SCL1 was incomplete primary education, ii) SCL2 was incomplete secondary education (up to Year 9), iii) SCL3 was schooling completed between Year 10 and Year 13 (without obtaining a high school/A-level diploma), and iv) SCL4 was High school/A-level diploma or higher.

Study outcomes

The main outcome of this study was the concordance between the total scores obtained on the GRECO-MMSE and the MMS-Run as part of a cognitive assessment for patients during routine care. We aimed to describe the distribution of the scores obtained (total score and according to the cognitive domains) for the two tests, evaluate the concordance of the scores between both tests at the level of each cognitive domain and examine the concordance according to the level of literacy of participants.

Sample size calculation

As the primary outcome was the linearly weighted kappa coefficient (Cicchetti–Allison weighting), no established method was found for determining the minimum required sample size. Consequently, we performed a series of simulations varying both kappa values (0.00 to 1.00) and sample sizes (10 to 100 participants). The results showed that, beyond 50 participants, the 95% confidence interval for kappa stabilized, with fluctuations of less than ±0.03. Therefore, a sample size of 50 participants was deemed sufficient to obtain a stable and reliable estimate of the coefficient.

Statistical analysis

Concordance between the GRECO-MMSE and MMS-Run total scores was assessed using a linearly weighted kappa coefficient, Bland-Altman plot, and a Spearman correlation study. The test results were considered discordant for any linearly weighted kappa coefficient strictly below 0.80. Although lower thresholds have been reported in the literature (e.g., 0.61–0.80, considered substantial agreement), it is now advised not to use them, since they allow high rates of discrepant interpretations between the two tests for the same sample.³⁰ Moreover, by selecting a high threshold, the evaluation of the MMS-Run adaptations becomes more stringent, making it easier to conclude that there is no meaningful effect. The Bland-Altman plot was used to visualize the presence of a systematic difference. This difference was expected to be zero if the scores matched. The level of correlation according to Spearman's coefficient was described as follows: negligible: 0.00 to 0.20, weak: 0.21 to 0.40, moderate: 0.41 to 0.60, strong: 0.61 to 0.80 and very strong: 0.81 to 1.00.³¹

The distributions of overall scores and the scores according to the cognitive domains were described in terms of mean, standard deviation, confidence interval of the mean, median, range and interquartile range. The distributions were compared using the Wilcoxon rank test. Given that the median MMSE-GRECO score in the French mainland population is 28/30 across all ages and education levels, floor and ceiling effects were assessed as the proportion of participants scoring in the lowest (≤2/30) and highest (≥28/30) 10% of the scale, respectively.²⁸ Concordance between the scores of the two tests for each cognitive domain and within the literate and illiterate groups was assessed similarly as the primary outcome. Confidence intervals were calculated at 95% (95%CI). Statistical significance was set at p < 0.05. Analyses were performed using R V4.3.2 software.

Ethical considerations

The study has received ethical approval by the Medical Ethics Committee (Comité de Protection des Personnes Ouest III) under Registration No. 23.00625.000169

Results

A total of 61 participants were included in the study, with most visiting the Geriatrics Department of the hospital (n = 45). The mean age of included participants was 81 years old and the sex ratio was 1.1 (Table 1). Most participants were self-declared as illiterate (57.4%). Of the illiterate, 32 were classified as SCL1 (91.4%), 2 as SCL2 (5.7%), and 1 as SCL3 (2.9%). The motive for taking the test was mostly for MNCD screening (63.9%) and most participants visiting the PM&R Department had a previous history of stroke (75.0%) (Table 1).

Table 1.

Baseline characteristics

Characteristics	Hospital department for the visit		Total, (n = 61)
Characteristics	PM&R, (n = 16)	Geriatrics, (n = 45)
Age (years), mean [min - max]	71 [65–80]	85 [65–97]	81 [ 65 - 97 ]
Sex, n, (%)
Male	14 (87.5)	18 (40.0)	32 ( 52.5)
Female	2 (12.5)	27 (60.0)	29 ( 47.5)
Motive for the test, n (%)
MNCD screening	2 (12.5)	37 (82.2)	39 ( 63.9)
MNCD monitoring	2 (12.5)	7 (16.6)	9 ( 14.8)
Stroke, n (%)	12 (75.0)	1 (2.2)	13 ( 21.3)
SCL, n (%)
1	8 (50.0)	33 (73.3)	41 ( 67.2)
2	1 (6.3)	8 (17.8)	9 ( 14.8)
3	6 (37.5)	3 (6.7)	9 ( 14.8)
4	1 (6.3)	1 (2.2)	2 ( 3.3)
Self-declared literacy level, n (%)
Literate	9 (56.2)	17 (37.8)	26 ( 42.6)
Illiterate	7 (43.8)	28 (62.2)	35 ( 57.4)

PM&R: physical medicine and rehabilitation; MNCD: major neurocognitive disorder; SCL: sociocultural level.

Concordance between the GRECO-MMSE and MMS-Run total scores

The linearly weighted kappa coefficient was 0.54 (95%CI [0.44 to 0.63]), indicating significant discordance between the GRECO-MMSE and MMS-Run total scores (Table 2). We observed a significant systematic difference of -3.11 points in the total score of the GRECO-MMSE compared with that of the MMS-Run (Figure 1). The correlation coefficient of 0.89 indicated a very strong correlation between the total scores of the two tests.

Figure 1.

Concordance analysis of the total mean scores between the GRECO-Mini-Mental State Examination (MMSE) and an adapted version for Reunion Island (MMS-Run) by Bland-Altman plot (n = 61).

Table 2.

Concordance analysis for total score and cognitive domain-specific scores of the GRECO-MMSE and MMS-Run using linearly weighted Kappa coefficients, systematic differences and Spearman correlation coefficients (n = 61, Reunion Island).

	Concordance		Systematic Difference		Correlation
Cognitive domain	κ	[95%CI]	d	[95%CI]	ρ	[95%CI]
Orientation	0.49	[0.38 to 0.61]	−1.36	[−1.83 to −0.89]	0.81	[0.71 to 0.88]
Learning	0.58	[0.36 to 0.79]	−0.02	[−0.18 to 0.15]	0.61	[0.42 to 0.75]
Attention	0.46	[0.32 to 0.60]	−1.07	[−1.45 to −0.68]	0.78	[0.65 to 0.86]
Recall	0.36	[0.20 to 0.53]	0.02	[−0.28 to 0.31]	0.48	[0.27 to 0.66]
Language	0.41	[0.27 to 0.56]	−0.74	[−1.03 to −0.44]	0.67	[0.50 to 0.79]
Constructional apraxia	0.65	[0.45 to 0.86]	0.05	[−0.05 to 0.15]	0.66	[0.48 to 0.78]
Total	0.54	[0.44 to 0.63]	−3.11	[−3.93 to −2.30]	0.89	[0.82 to 0.93]

κ: linearly weighted kappa coefficient; 95%CI: 95% confidence interval; d: systematic difference; ρ: Spearman correlation coefficient.

Distribution of the total scores

The median total scores were 15 and 19 out of 30 for the GRECO-MMSE and MMS-Run, respectively. The largest divergence occurred at the third quartile, with scores of 21 and 26 out of 30 (Supplemental Figure 1). This difference was particularly notable near the established diagnostic thresholds for the GRECO-MMSE (from 22 to 26, depending on participants’ SCL). Moreover, we observed a narrower dispersion of scores with the GRECO-MMSE test (interquartile range=10) than with the MMS-Run (interquartile range=13). Overall, the difference in the distribution of total scores was statistically significant (p = 0.012).

No significant floor effect was observed, as only one participant scored below 2 points on the GRECO-MMSE. A greater ceiling effect was observed for the MMS-Run (11.5% of participants scoring above 28) than for the GRECO-MMSE (3.3%). However, only one participant achieved the maximum score on the GRECO-MMSE, and none on the MMS-Run.

Concordance between GRECO-MMSE and MMS-Run cognitive domains

The assessment of concordance between each cognitive domain is presented in Table 2. All cognitive domains except constructional apraxia showed significant discordance. We also found significant systematic differences for orientation, attention and language (Figure 2). Correlations between the two tests were strong for attention and orientation, moderate for learning, language, and constructional apraxia, and weaker for recall.

Figure 2.

Concordance analysis of cognitive domains (orientation, attention, and language) between the GRECO-Mini-Mental State Examination (MMSE) and an adapted version for Reunion Island (MMS-Run) by Bland-Altman plot (n = 61).

Distribution of scores according to the cognitive domains

Comparing cognitive domain scores between the GRECO-MMSE and MMS-Run (Supplemental Table 1), distribution differences were observed for orientation, attention, and language, with lower mean scores on the GRECO-MMSE. In addition, for orientation, half of the participants scored at least five points on the GRECO-MMSE, compared with three quarters on the MMS-Run. For attention, half of the participants failed all subtraction calculations on the GRECO-MMSE, while half achieved at least two correct responses on the MMS-Run. Regarding language, all participants succeeded at least two MMS-Run tasks, which was not systematically the case with the GRECO-MMSE.

Analyses of literacy

Concordance analysis revealed significant discordance between GRECO-MMSE and MMS-Run total scores in both literate and illiterate participants (Table 3). It was more pronounced among the illiterate group.

Table 3.

Concordance analysis for total score and cognitive domain-specific scores of the GRECO-MMSE and MMS-Run based on the literacy level of elderly participants on Reunion Island (literate: n = 26, illiterate: n = 35).

	Literate		Illiterate
Cognitive domain	κ	95% Confidence Interval	κ	95% Confidence Interval
Orientation	0.44	[0.25 to 0.62]	0.52	[0.38 to 0.66]
Learning	0.57	[0.21 to 0.93]	0.58	[0.32 to 0.84]
Attention	0.52	[0.30 to 0.75]	0.31	[0.15 to 0.47]
Recall	0.56	[0.37 to 0.76]	0.20	[−0.02 to 0.42]
Language	0.57	[0.40 to 0.75]	0.29	[0.11 to 0.47]
Constructional apraxia	0.70	[0.42 to 0.97]	0.46	[0.06 to 0.86]
Total	0.63	[0.51 to 0.74]	0.43	[0.31 to 0.55]
	d		d
Total	−2.54	[−3.74 to −1.34]	−3.54	[−4.65 to −2.43]
	ρ		ρ
Total	0.88	[0.81 to 0.93]	0.86	[0.78 to 0.92]

κ: linearly weighted kappa coefficient; d: systematic difference; ρ: Spearman correlation coefficient.

Analysis by cognitive domain allowed us to identify which literacy levels amplified differences between the two tests. Greater discordance among illiterate participants was observed for attention, recall, and language. Similar levels of discordance were found across literacy groups for orientation, while no significant discordance was observed for learning and constructional apraxia, suggesting that these three domains may not be influenced by literacy level.

Compared with the MMS-Run total score, the GRECO-MMSE total score showed a more important systematic difference among illiterate participants than among literate ones (Figure 3). The correlation between total scores remained strong regardless of literacy level.

Figure 3.

Concordance analysis of the total mean scores between the GRECO-Mini-Mental State Examination (MMSE) and an adapted version for Reunion Island (MMS-Run) by Bland-Altman plot, based on literacy levels (illiterate: n = 35, literate: n = 26).

Discussion

Main findings

Consistent with clinical observations at the University Hospital of Reunion Island, this study revealed a discordance between the GRECO-MMSE and the MMS-Run. This was primarily due to a systematic difference between them, with lower scores observed for the GRECO-MMSE. Such findings indicate that the two tests may lead to different conclusions for the management of MNCDs. Nevertheless, the strong correlation suggest that both tests assess the same cognitive domains even though the GRECO-MMSE is more stringent. Orientation, attention and language were identified as the cognitive domains contributing to the observed discrepancies. These included items most adapted to the context of Reunionese culture and literacy level. Learning and recall showed significant discordance according to kappa values, but their distributions were similar and the systematic differences were not significant. Constructional apraxia was the only domain to show no discordance (the items were identical in both tests).

This study also provided an initial estimate of the illiteracy rate among people over 65 years on Reunion Island. In the absence of published data for this age group, the observed rate (57%) far exceeded our initial projection of 38.2%. Most of the illiterate individuals were classified as SCL1 (91%). Therefore, this population differs from the sample used for the standardization and calibration of the MMSE by the GRECO, where only 14% were classified as SCL1, as reported in Kalafat et al. (2003).²⁸

The impact of literacy level on our findings was emphasized by the analysis of illiterate and literate participants. Although discordances were present in both groups, they were more pronounced among illiterate participants, particularly in attention, recall, and language, which require skills in calculation, lexical memory, writing, and reading. These domains are known to be influenced by literacy level.¹¹ In contrast, orientation, learning, and constructional apraxia did not appear to be affected by the level of literacy, consistent with previous findings by Reis et al. (2003).³³

The influence of Reunionese culture on test performance can also be highlighted. Indeed, literacy level alone does not account for the observed discordance between the two tests, as it is also present among literate participants who received schooling comparable to that in mainland France. Orientation—a domain not influenced by literacy level—shows a significant discordance across the entire sample. Notably, it includes the greatest number of adaptations to Reunionese culture, particularly regarding spatial references. The systematic difference observed in this domain accounts for a substantial proportion of the overall systematic difference in total scores (−1.36 out of −3.11). Interestingly, orientation—together with the three-word learning task—is also among the domains most frequently adapted, particularly in various Asian versions of the MMSE.³⁴

By selecting a hospital-based population, in which cognitive impairments are more prevalent, we likely limited the ceiling effect commonly observed with the GRECO-MMSE in the general population. This approach is recommended, as ceiling effects tend to reduce the internal consistency of testing.³² Here, only 3.3% of participants scored above 28 out of 30 with the GRECO-MMSE and 11.5% with the MMS-Run, compared with 50–75% in the GRECO-MMSE standardization study conducted in the general population.²⁸ This limited ceiling effect enhanced the quality of our analyses, particularly those concerning concordance and correlation.

With the GRECO-MMSE, over three quarters of participants scored below 21, the pathological threshold for the lowest SCL, according to Kalafat et al. (2003).²⁸ While cognitive impairment in the sample cannot be excluded, the improved performance observed on the adapted version cannot be attributed merely to task facilitation. For example, in the attention domain, which involves performing five successive subtractions, half of the participants scored zero on the GRECO-MMSE, whereas they managed at least two correct answers with the MMS-Run. However, the purpose of this task is to assess short-term information processing rather than arithmetic skills. Thus, the MMS-Run preserves the underlying construct to better assess the desired criteria within the Reunionese context. Other culturally and literacy-adapted MMSE versions have been shown to enhance specificity, thereby reducing the likelihood of false positives in these populations.¹¹

Limitations

Although our sample size exceeds the minimum estimated, it remains relatively small, which limits the generalizability of our findings. This is particularly true for subgroup analyses by literacy level, where the groups consisted of 26 literate and 35 illiterate participants. The study is also monocentric, and because it relies on consecutive sampling, the sample lacks randomization.

This study had a selection bias since participants came from a hospitalized population, and illiteracy itself constitutes a risk factor for hospitalization.^37,38 Therefore, the observed illiteracy rate may be overestimated compared to the general population. A measurement bias was also present in the classification of literacy levels since this classification was based on subjective self-reporting of difficulties in writing, reading, or calculation, which could lead to errors in participant categorization. Moreover, by assigning a score of zero in cases of missing or refused responses, we may have introduced a bias that could alter the evaluation of the cognitive function assessed by the item.

We did not consider the use of the Reunionese Creole language as a study criterion because its impact has been shown to be negligible on French-language neuropsychological tests in previous literature.^35,36

Furthermore, since the MMS-Run was systematically administered after the MMSE, the question of the learning effect arises, particularly given the overall improvement observed in total scores. However, this improvement was not consistent across all cognitive domains. For instance, performance on the constructional apraxia task—the only one identical in both tests—did not show systematic improvement. Additionally, efforts were made not to correct participants’ errors during the tests, thereby minimizing the learning effect and only encouraging spontaneous correct responses. Since the GRECO-MMSE was the only validated test, we chose to conduct it first to preserve the integrity of its score. No randomization was performed since this would have altered standardized care practices.

Lastly, within the MMS-Run, the reading and writing tasks, which assess written comprehension and expression, were replaced by tasks involving non-verbal comprehension and oral expression. Therefore, the domains explored were not strictly identical, limiting the comparability of results.

Recommendations

Our results call into question the use of the GRECO-MMSE for the Reunionese population over 65 years. By improving the relevance of the MMSE for Reunionese individuals, the MMS-Run may reduce the number of false positives in MNCD screening, thereby increasing the specificity and predictive value of the test.

One item required modification in both tests. The phrase to repeat used in the language evaluation is a French or Reunionese Creole translation of the original phrase, ‘No ifs, ands, or buts’. This idiomatic expression, common in English, lacks meaning in the French or Creole version. Werner et al (1999) conducted a study in Hebrew which showed that using a phrase deeply rooted in the local culture improves diagnostic performance and reduces the influence of participants’ educational level.³⁹

As the MMS-Run appears better suited for the Reunionese population, further studies are required. Construct validity could be improved by comparing it again with the GRECO-MMSE, while randomizing the order of test administration to eliminate potential learning effects. Any future comparative study should also account for multiple comparisons across cognitive domains. It would be important to assess inter-rater reliability as well as test-retest reliability. Finally, establishing validated normative values is essential for accurate interpretation of the results. This requires a study on a larger sample, including participants with known cognitive impairments as well as cognitively intact individuals.

In the meantime, the use of the MMS-Run, or the GRECO-MMSE on Reunion Island, should be approached with caution, and scores should not be interpreted based on a single administration. Nonetheless, a decline in scores on the same test over repeated assessments conducted several months apart should raise concern and warrant further investigation.

More generally, this study reinforces the importance of adapting cognitive tests to the population being assessed, going beyond simple language translation. As illustrated by the case of Reunion Island, such adaptations may be necessary for specific territories. The same consideration should also apply to populations with immigrant backgrounds, for whom tests reflecting their original cultural context should be preferred whenever available.

Conclusion

The psychometric properties of the MMS-Run appear to better meet the needs of the Reunionese population, including individuals with a high SCL, due to the various cultural adaptations it has undergone. A validation study of the MMS-Run should be conducted within the Reunionese population in the future.

Supplemental Material

sj-docx-1-alz-10.1177_13872877261418292 - Supplemental material for The role of culture and literacy in screening for major neurocognitive disorders for elderly patients on Reunion Island: A cross-sectional study on the concordance between the French Mini-Mental State Examination and an adapted version (MMS-Run)

Supplemental material, sj-docx-1-alz-10.1177_13872877261418292 for The role of culture and literacy in screening for major neurocognitive disorders for elderly patients on Reunion Island: A cross-sectional study on the concordance between the French Mini-Mental State Examination and an adapted version (MMS-Run) by Alexis Chevy, Antoine Tisseaux, Cyril Ferdynus, Mirella Moutounaïck, Jean-Marc Eychène, Elena Morello, Rosanne Ufkes, Stéphane Renaud, Jean-Marc David, Véronique Martins, Léa Bruneau and Bénédicte Le Quang in Journal of Alzheimer's Disease

Footnotes

Acknowledgements

The authors would like to thank all those who participated in this study from the Department of Physical Medicine and Rehabilitation and the Department of Geriatrics at the Centre Hospitalier Universitaire de La Réunion. They would also like to thank the INSERM CIC 1410 for their support in the methodology, Ms Emilie Techer for her administrative assistance as well as Marty Brucato and Sarina Yaghobian (AcaciaTools) for their reviewing and proofreading services.

ORCID iDs

Alexis Chevy

Antoine Tisseaux

Cyril Ferdynus

Mirella Moutounaïck

Jean-Marc Eychène

Elena Morello

Rosanne Ufkes

Stéphane Renaud

Jean-Marc David

Véronique Martins

Léa Bruneau

Bénédicte Le Quang

Ethical considerations

The study has received ethical approval by the Medical Ethics Committee (Comité de Protection des Personnes Ouest III) under Registration No. 23.00625.000169.

Consent to participate

Verbal non-opposition to the use of the data was obtained from all patients after they were provided with an information sheet about the study.

Consent for publication

Not applicable

Author contribution(s)

Alexis Chevy: Conceptualization; Investigation; Methodology; Supervision; Writing – original draft.

Antoine Tisseaux: Conceptualization; Formal analysis; Methodology; Writing – review & editing.

Cyril Ferdynus: Conceptualization; Formal analysis; Methodology; Writing – review & editing.

Mirella Moutounaïck: Investigation; Writing – review & editing.

Jean-Marc Eychène: Investigation; Writing – review & editing.

Elena Morello: Investigation; Writing – review & editing.

Rosanne Ufkes: Investigation; Writing – review & editing.

Stéphane Renaud: Investigation; Writing – review & editing.

Jean-Marc David: Investigation; Writing – review & editing.

Véronique Martins: Investigation; Writing – review & editing.

Léa Bruneau: Conceptualization; Methodology; Project administration; Writing – original draft.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supplemental material

Supplemental material for this article is available online.

References

Emmady PD, Schoo C and Tadi P. Major neurocognitive disorder (dementia). Treasure Island (FL): StatPearls Publishing. https://www.ncbi.nlm.nih.gov/books/NBK557444/ (2022, accessed 9 December 2025)

N° 2454 - Rapport sur la maladie d’Alzheimer et les maladies apparentées au nom de l’Office parlementaire d’évaluation des politiques de santé (Mme Cécile Gallez). https://www.assemblee nationale.fr/12/rap-off/i2454.asp (2005, accessed 9 December 2025)

Bennett

Thomas

AJ.

Depression and dementia: cause, consequence or coincidence?

Maturitas 2014; 79: 184–190.

van Dyck

Swanson

Aisen

, et al. Lecanemab in early Alzheimer's disease. N Engl J Med 2023; 388: 9–21.

Petretto

Carrogu

Gaviano

, et al. Dementia and major neurocognitive disorders: some lessons learned one century after the first Alois Alzheimer’s clinical notes. Geriatrics (Basel) 2021; 6: 5.

Alzheimer’s Association. 2013 Alzheimer’s disease facts and figures. Alzheimers Dement 2013; 9: 208–245.

Brett

Gardner

Godbout

, et al. Traumatic brain injury and risk of neurodegenerative disorder. Biol Psychiatry 2022; 91: 498–507.

Campbell

Unverzagt

LaMantia

, et al. Risk factors for the progression of mild cognitive impairment to dementia. Clin Geriatr Med 2013; 29: 873–893.

Dartigues

Gagnon

Michel

, et al. [The Paquid research program on the epidemiology of dementia. Methods and initial results]. Rev Neurol (Paris) 1991; 147: 225–230.

10.

Zhang

Katzman

Salmon

, et al. The prevalence of dementia and Alzheimer’s disease in Shanghai, China: Impact of age, gender, and education. Ann Neurol 1990; 27: 428–437.

11.

Julayanont

Ruthirago

The illiterate brain and the neuropsychological assessment: From the past knowledge to the future new instruments. Appl Neuropsychol Adult 2018; 25: 174–187.

12.

Pettigrew

Soldan

Defining cognitive reserve and implications for cognitive aging. Curr Neurol Neurosci Rep 2019; 19: 1.

13.

Buré-Reyes

Hidalgo-Ruzzante

Vilar-López

, et al.

Neuropsychological test performance of Spanish speakers: is performance different across different Spanish-speaking subgroups?

J Clin Exp Neuropsychol 2013; 35: 404–412.

14.

Lohnas

Kahana

MJ.

Parametric effects of word frequency in memory for mixed frequency lists. J Exp Psychol Learn Mem Cogn 2013; 39: 1943–1946.

15.

Park

Huang

CM.

Culture wires the brain: a cognitive neuroscience perspective. Perspect Psychol Sci 2010; 5: 391–400.

16.

Ganguli

Ratcliff

Chandra

, et al. A Hindi version of the MMSE: The development of a cognitive screening instrument for a largely illiterate rural elderly population in India. Int J Geriat Psychiatry 1995; 10: 367–377.

17.

Bassetti

Clarke

Trenkic

The linguistic transparency of first language calendar terms affects calendar calculations in a second language. Acta Psychol (Amst) 2018; 186: 81–89.

18.

Manly

Echemendia

RJ.

Race-specific norms: using the model of hypertension to understand issues of race, culture, and education in neuropsychology. Arch Clin Neuropsychol 2007; 22: 319–325.

19.

Lucas

La Mise En Place de l’école Dans Une Société Esclavagiste Puis Coloniale: La Réunion 1815-1946. Université de Paris VIII; 1994. https://theses.fr/1994PA080923

20.

Michailesco F and Le Grand H. 116 000 personnes en situation d’illettrisme en 2011 à La Réunion. INSEE. https://www.insee.fr/fr/statistiques/1291760 (2013, accessed 9 December 2025)

21.

Baktavatsalou R, Chaussy C and Sui-Seng S. Trois fois plus de seniors en perte d’autonomie en 2050: Projections de personnes âgées en perte d’autonomie à La Réunion. INSEE. August 27, 2020. https://www.insee.fr/fr/statistiques/4643075 (2020, accessed 9 December 2025)

22.

INSEE. Évolution et structure de la population en 2021: Département de La Réunion (974). INSEE; 2024. https://www.insee.fr/fr/statistiques/8200783?geo=DEP-974 (2024, accessed 9 December 2025)

23.

Folstein

McHugh

PR.

“Mini-mental state.” J Psychiatr Res 1975; 12: 189–198.

24.

Bondi

Edmonds

Salmon

DP.

Alzheimer’s disease: past, present, and future. J Int Neuropsychol Soc 2017; 23: 818–831.

25.

Tsoi

KKF

Chan

JYC

Hirai

, et al. Cognitive tests to detect dementia: a systematic review and meta-analysis. JAMA Intern Med 2015; 175: 1450.

26.

Tangalos

Smith

Ivnik

, et al. The Mini-Mental State Examination in general medical practice: clinical utility and acceptance. Mayo Clin Proc 1996; 71: 829–837.

27.

Derouesne

Poitreneau

Hugonot

, et al. [Mini-Mental State Examination: a useful method for the evaluation of the cognitive status of patients by the clinician. Consensual French version]. Presse Med 1999; 28: 1141–1148.

28.

Kalafat

Hugonot-Diener

Poitrenaud

Standardisation et étalonnage français du “Mini Mental State” (MMS) version GRÉCO [French standardization and range for the GRECO version of the “Mini Mental State”

(MMS)]. Rev Neuropsychol 2003; 13: 209–236.

29.

Tombaugh

McIntyre

NJ.

The mini-mental state examination: a comprehensive review. J Am Geriatr Soc 1992; 40: 922–935.

30.

McHugh

ML.

Interrater reliability: the kappa statistic. Biochem Med (Zagreb) 2012; 22: 276–282.

31.

Prion

Haerling

KA.

Making sense of methods and measurement: Spearman-Rho ranked-order correlation coefficient. Clin Simul Nurs 2014; 10: 535–536.

32.

Kukull

Larson

Teri

, et al. The mini-mental state examination score and the clinical diagnosis of dementia. J Clin Epidemiol 1994; 47: 1061–1067.

33.

Reis

Guerreiro

Petersson

KM.

A sociodemographic and neuropsychological characterization of an illiterate population. Appl Neuropsychol 2003; 10: 191–204.

34.

Shim

Yang

Kim

, et al. Characteristic differences in the mini-mental state examination used in Asian countries. BMC Neurol 2017; 17: 141.

35.

Soulaimana

Bruneau

Tisseaux

, et al. Psychometric properties of a Creole version of Medical Outcome Study – Short Form 36 among type II diabetes patients on Reunion Island. Front Public Health 2023; 11: 1186153.

36.

Safieddine

Bruneau

Soulaimana

, et al. Quality of life assessment in diabetic patients: validity of the creole version of the EQ-5D-5L in Reunion Island. Front Psychol 2023; 14: 1185316.

37.

Berkman

Sheridan

Donahue

, et al. Low health literacy and health outcomes: an updated systematic review. Ann Intern Med 2011; 155: 97.

38.

Easton

Entwistle

Williams

Health in the “hidden population” of people with low literacy. A systematic review of the literature. BMC Public Health 2010; 10: 459.

39.

Werner

Heinik

Lin

, et al. ‘Yes’ ifs, ands or buts: examining performance and correlates of the repetition task in the Mini-Mental State Examination. Int J Geriatr Psychiatry 1999; 14: 719–725.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB