G Model

AGG-2719; No. of Pages 6

Archives of Gerontology and Geriatrics xxx (2012) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Archives of Gerontology and Geriatrics

journal homepage: www.elsevier.com/locate/archger

Cross-cultural difference and validation of the Chinese version of Montreal

Cognitive Assessment in older adults residing in Eastern China: Preliminary
findings
Jian-bo Hu, Wei-hua Zhou, Shao-hua Hu, Man-li Huang, Ning Wei, Hong-li Qi, Jin-wen Huang, Yi Xu *
Department of Mental Health, The First Affiliated Hospital, College of Medicine, Zhejiang University, No. 79 Qingchun Road, Hangzhou, China

A R T I C L E I N F O A B S T R A C T

Article history: To evaluate the psychometric properties of the Chinese Montreal Cognitive Assessment (MoCA-C) and
Received 25 March 2012 assess cross-cultural differences in a community-based cohort residing in the Eastern China. The study
Received in revised form 14 May 2012 included 72 patients with Alzheimer’s disease (AD), 84 patients with mild cognitive impairment (MCI)
Accepted 17 May 2012
and 146 cognitively normal controls. Sensitivities and specificities were calculated using the
Available online xxx
recommended cut-off scores. Receiver operator characteristic (ROC) curve analyses were performed
to determine optimal sensitivity and specificity. Criterion validity, inter-rater, test–retest reliability and
Keywords:
internal consistencies of the MoCA-C were examined, and clinical observations made. The influence of
Mild cognitive impairment
age, education level and gender on MoCA score was examined. Using the recommended cut-off score of
Montreal Cognitive Assessment
Validity 26, the area under the ROC (AUC) for predicting MCI groups using the MoCA-C was 0.930 (95%CI: 0.894;
MMSE 0.965). The MoCA-C demonstrated 92% sensitivity and 85% specificity in screening for MCI. Cultural
Screening differences from the original MoCA affected the test response rate. The MoCA-C appears to have utility as
Cultural differences a cognitive screen for early detection of AD and for MCI and warrants further investigation regarding its
applicability in primary care settings in elderly Chinese people. It will be necessary to revise the contents
of the questionnaire to account for by local characteristics.
! 2012 Elsevier Ireland Ltd. All rights reserved.

1. Introduction The Montreal Cognitive Assessment (MoCA), developed and

validated by Nasreddine et al., 2005, is a brief and potentially
MCI is defined as cognitive decline greater than expected for an useful screening tool with high sensitivity and specificity for
individual’s age and education level, that does not interfere notably detecting MCI in persons performing in the normal range on the
with activities of daily life. The prevalence of MCI in population- Mini-Mental State Examination (MMSE). The validity of the MoCA
based epidemiological studies ranges from 3% to 19% in adults has been studied in various clinical settings (Cumming, Bernhardt,
older than 65 years (Gauthier et al., 2006). MCI is recognized as a & Linden, 2011; Gaviria, Pliskin, & Kney, 2011; Gill, Freshman,
risk factor for AD (Levey, Lah, Goldstein, Steenland, & Bliwise, Blender, & Ravina, 2008; Nazem et al., 2009; Olson, Chhanabhai, &
2006), prompting many researchers to screen for MCI in order to McKenzie, 2008; Popovic, Seric, & Demarin, 2007; Videnovic et al.,
provide early treatment and reduce the risk of progression to 2010). In Western countries the MOCA has been shown to have
dementia (Ritchie, Artero, & Touchon, 2001). However, owing to good sensitivity and specificity in detecting MCI and is widely used
the subtle decline in cognitive function during the initial stages of in various fields. However, because of cultural background and
MCI, improvements in recognition and diagnosis are still needed if lifestyle differences in Eastern and Western countries, it is
the economic and psychosocial burdens associated with AD are to necessary to assess the scale in patients from different cultures.
be reduced. This will only be possible if treatment is initiated prior Studies undertaken in Asian countries (including Japan, Korea,
to the onset of full-blown dementia syndrome. The benefits are Singapore and China Hong Kong) show that the reliability and
minimal once the disease progression is underway since available validity of MoCA in screening for MCI is superior to that of the
treatments are unable to reverse disease progression and restore MMSE. However, to account for the cultural differences in these
individuals to their premorbid level (Luis, Keegan, & Mullan, 2009). studies the recommended cut-off values were different to those in
the original paper (Dong et al., 2010; Fujiwara et al., 2010; Lee et al.,
2008; Wong et al., 2009).
* Corresponding author. Tel.: +86 571 56723001; fax: +86 571 56723001.
The primary purpose of our study was to evaluate the
E-mail address: yixu1961@yahoo.com.cn (Y. Xu). psychometric properties of the Chinese version of the MoCA

0167-4943/$ – see front matter ! 2012 Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.archger.2012.05.008

Please cite this article in press as: Hu, J.-b., et al., Cross-cultural difference and validation of the Chinese version of Montreal Cognitive
Assessment in older adults residing in Eastern China: Preliminary findings. Arch. Gerontol. Geriatr. (2012), http://dx.doi.org/10.1016/
j.archger.2012.05.008
G Model
AGG-2719; No. of Pages 6

e2 J.-b. Hu et al. / Archives of Gerontology and Geriatrics xxx (2012) xxx–xxx

(MoCA-C) in a community-based cohort residing in the Eastern 2.3. Translation and cultural modifications of the MoCA
China. We also used the MoCA-C to investigate cross-cultural
differences between Western and Eastern countries. The MoCA is a instrument that evaluates seven cognitive
domains on a single page. The domains are: visuospatial/executive
2. Methods functions, naming, verbal memory registration and learning,
attention, abstraction, 5-min delayed verbal memory, and
2.1. Participants orientation. Scores of the MoCA range from 0 to 30. The Chinese
Beijing Version used in the study is available from http://
Seventy-two patients with AD patients were recruited from www.mocatest.org. Items in the MoCA-C were identical to the
consecutive referrals to our hospital. Each met Diagnostic and English version with the exception of the following four cultural
Statistical Manual of Mental Disorders, Fourth Edition, Text and linguistic modifications:
revision (DSM-IV-TR) criteria (American Psychiatric Association,
2000) and the National Institute of Neurological and Communica- In Item 1 (visuospatial/executive functions – Alternating Trail
tive Disorders and Stroke-Alzheimer’s Disease and Related Making), Chinese character sequences replaced Roman alpha-
Disorders Association (NINCDS-ADRDA) diagnostic criteria bets. The number of steps required for completion of task was
(McKhann et al., 1984). retained.
Case definitions for MCI were based on Petersen et al. (2001) In Item 4 (Attention-Auditory Vigilance), Arabic numerals were
criteria. Subjects in our study were required to have subjective used instead of English alphabet letters. The number and
memory complaints proven using an abbreviated memory positions of responses remained identical to those in the
inventory from the Chinese MMSE and confirmed using the English version.
Clinical Dementia Rating (CDR). Objective memory impairment In Item 5 (Language – Sentence repetition), the names were
was determined by a composite memory score derived from the changed to a more common Chinese names to reflect local
Alzheimer’s Disease Assessment Scale-cognitive subscale (ADAS- familiarity.
Cog) mean immediate recall and recognition test scores, and by a In Item 5 (Language – Verbal fluency), semantic fluency using
10-min delayed recall score. The conventional cut-off level for the animal category replaced phonemic letter fluency as there
memory performance of MCI patients was set at the !1.5 standard are no letter-equivalent linguistic units in the Chinese language.
deviations (SDs) below the mean for education-matched groups. In order to avoid the influence on the score of the naming task
None of the MCI subjects in our study showed evidence of contains figures of animals, the subjects were informed three
dementia; all had a CDR score of 0, together with close to normal animal names in the naming task cannot be mentioned again,
scores for activities of daily living (ADL) that were above the level otherwise were not scored.
for clinical AD. Subtle changes in complex ADL did not exclude
subjects from the study. However, all subjects with MCI were 2.4. Statistical analysis
required to be able to perform basic self-care.
In order to include subjects with possible non-amnestic MCI, All statistical analyses were performed using SPSS version 13.0.
those with memory scores above the !1.5 SD criteria were Differences between groups in demographic variables, MoCA-C,
recruited into the MCI group provided there were two or more and MMSE scores were examined using one-way analysis of
non-memory domains of CDR rated as 0.5 or above. Patients with variance (ANOVA) or chi square (x2) analyses, depending on the
significant depression or other psychiatric disorders assessed by level of variable measurement. Statistically significant demo-
the (DSM-IV) or those with MRI evidence of cortical infarct or a graphic variables were used as covariates in examining group
history of hemorrhagic stroke were excluded. differences in the MoCA-C and MMSE.
The final determination of diagnoses at entry was based on a Cronbach’s alpha was used to assess the internal consistencies
consensus meeting involving a neurologist, psychiatrist, neuro- of MoCA-C. Test–retest and inter-rater reliability were evaluated
psychologist, and radiologist, when applicable. A study coordinator by calculating intra-class correlation coefficients (ICCs) for scores
collected demographic data and performed all screening measures at the baseline and at follow-up. A nonparametric ROC analysis
(MMSE and MoCA-C) prior to diagnostic work-up. appropriate for small sample sizes was utilized to assess the ability
Sixty cases were subjected to a retest 4 weeks after the initial of the MoCA-C and MMSE to differentiate MCI and AD from normal
visit to assess the test–retest reliability. cognition (Hanley and McNeil, 1983). This type of analysis was also
A control group comprised 146 subjects recruited from the used to identify the optimal balance between sensitivity and
community in Hangzhou who were independent in ADL, had no specificity. Area under the curve (AUC) was used to compare the
history of psychiatric or neurological disease and no memory diagnostic performance of each test. Values of P < 0.05 were
complaints. They performed in the normal range on standardized considered statistically significant.
neuropsychological tests.
The study was approved by the Domain Specific Review Board 3. Results
and Ethics Committee of the Healthcare Group of China. Written
informed consent was obtained from all participants or their Demographic information, MMSE and MoCA scores for the
legally acceptable representatives. three groups are shown in Table 1. The mean age and education
level for the total sample was 64.4 years and 9.4 years, respective-
2.2. Neuropsychological assessment ly. No differences were found between groups in terms of
education level, or gender, but on average patients with MCI
A battery of neuropsychological tests, conducted by trained (60.7 years) were significantly younger (P < 0.001) than those with
psychologists, consisted of the MMSE, CDR, MoCA-C, ADL scale, AD (68.4 years) or normal cognitive function (67.2 years). Gender
ADAS-Cog and the Auditory Verbal Learning Test (AVLT). There was was not related to MMSE or MoCA-C scores. However, age was
a break of about 5 min after each test to eliminate the possibility of found to correlate with performance on MMSE but not on MoCA-C.
interaction effects. To avoid experimental bias, the investigators In addition, education level impacted performance on both the
performing and scoring the cognitive tests were blind with respect MMSE and MoCA-C with individuals with 6 years of education or
to each participant’s background. less performing less well on both measures.

Please cite this article in press as: Hu, J.-b., et al., Cross-cultural difference and validation of the Chinese version of Montreal Cognitive
Assessment in older adults residing in Eastern China: Preliminary findings. Arch. Gerontol. Geriatr. (2012), http://dx.doi.org/10.1016/
j.archger.2012.05.008
G Model
AGG-2719; No. of Pages 6

J.-b. Hu et al. / Archives of Gerontology and Geriatrics xxx (2012) xxx–xxx e3

Table 1
Demographic characteristics.

Mean $ SD

AD (n = 72) MCI (n = 84) NC (n = 146) F P-value

Age (years) 68.4 $ 4.3 60.7 $ 5.0 67.2 $ 5.3 32.074 0.001
Gender (M/F) 32/40 36/48 65/81 0.066c 0.967
Education level (years) 8.9 $ 3.1 9.82 $ 3.0 9.3 $ 2.6 2.056 0.130
MoCA-C score* 17.2 $ 5.5 24.5 $ 1.9 27.7 $ 1.3a 294.311 <0.001
MMSE score 20.1 $ 3.4b 27.3 $ 1.2b 28.5 $ 1.1b 468.747 <0.001

MMSE, Mini-Mental State Examination; MoCA-C, Chinese version of Montreal Cognitive Assessment; AD, Alzheimer’s disease; MCI, mild cognitive impairment; NC, normal
controls; SD, standard deviation.
a
AD < MCI < NC.
b
AD < NC and MCI.
c
The x2-test was used to test association among categorical variables, and ANOVA was used to compare the means of continuous variables.
*
Scores include education correction.

Diagnostic group differences were found by ANOVA for both the sensitivity (95%) for detecting MCI. Using a modified MMSE cut-off
MMSE (P < 0.001) and MoCA-C (P < 0.001; Table 1). Post hoc point of 28 resulted in good sensitivity (85%), but poor specificity
analysis revealed that MMSE scores were significantly lower in the (53%) for screening MCI.
AD group than in the MCI and control groups, neither of which No significant differences in the MoCA-C scores (P > 0.05) were
differed from each other. By contrast, the MoCA-C scores were observed when we used the recommended education level from
significantly different between all three groups (Table 1). the original MoCA test to divide the normal cognitive group into
The MoCA-C scores were highly and positively associated with two subgroups of "12 years (n = 114) or >12 years (n = 32).
MMSE scores (r = 0.83, P < 0.001), ADAS-Cog scores (r = 0.73, However formation of subgroups with education levels of <6 years
P < 0.001) and were negatively associated with the sum of CDR ars (n = 63) or #6 years (n = 83), resulted in statistically signifi-
scores (r = !0.79, P < 0.001). cantly higher MOCA-C scores in those with less education
The ICCs for test–retest and inter-rater reliability were 0.862 (P < 0.05).
(95%CI: 0.789; 0.918) and 0.945 (95%CI: 0.874; 0.997), respectively To compensate for the low education level in the local elderly
(both P < 0.001). The Cronbach’s alpha for the MoCA-C was 0.867, population1 point was added to the total score for subjects with
indicating a high level of internal consistency. <6 years education. Table 3 shows corrected MOCA-C scores in the
ROC curves were drawn for the MCI group and normal controls two educational level subgroups in subjects with normal cognitive
(Fig. 1) and for the AD group and normal controls (Fig. 2) to function.
determine the discriminatory validity of MoCA-C and MMSE. The The overall average correct scoring rate was 71.49%. Below
results demonstrate that the discriminatory validity of the MoCA-C average scoring was observed for a number of items (‘Trail B’, ‘Cube
for MCI was excellent. The AUC was 0.928 (95%CI: 0.921; 0.936) for coping’, ‘Clock drawing’, ‘Rhinoceros’, ‘Sentence repetition 1’,
MCI and 0.962 (95%CI: 0.955; 0.970) for AD. Corresponding values ‘Abstraction 2’, ‘Face’, ‘Velvet’, ‘Church’ and ‘Orientation’). As
for MMSE were 0.741 (95%CI: 0.728; 0.751) for MCI and 0.984 shown in Table 3, scores for ‘Trail B’, ‘Cube coping’, ‘Clock drawing’,
(95%CI: 0.981; 0.987) for AD. ‘Rhinoceros’, ‘Camel’, ‘Vigilance’, ‘Serials 7 s’, ‘Sentence repetition
The results in Table 2 show the sensitivity and specificity 1’, ‘Abstraction 2’, ‘Face’, ‘Velvet’ and ‘Church’ were all influenced
analyses for the MMSE and MoCA-C among diagnostic groups by the level of education.
using the recommended cut-off scores of 26 for the MoCA-C and 24 The MoCA-C was generally well accepted by elderly Chinese
for the MMSE. We found that the recommended cut-off of 26 subjects. The average administration time depended on the
provided a sensitivity of 92% and specificity of 85% for discrimi- education level and severity of cognitive symptoms. In patients
nating between the MCI and normal control groups. The cut off of with MCI the average was 14 min for the MoCA-C compared with
24 for the MMSE provided poor specificity (23%) but excellent 10 min for the MMSE (P < 0.01).

Fig. 1. ROC curves for MoCA-C and MMSE in the early detection of MCI (P < 0.01). Fig. 2. ROC curves for MoCA-C and MMSE in the early detection of AD (P > 0.01).

Please cite this article in press as: Hu, J.-b., et al., Cross-cultural difference and validation of the Chinese version of Montreal Cognitive
Assessment in older adults residing in Eastern China: Preliminary findings. Arch. Gerontol. Geriatr. (2012), http://dx.doi.org/10.1016/
j.archger.2012.05.008
G Model
AGG-2719; No. of Pages 6

e4 J.-b. Hu et al. / Archives of Gerontology and Geriatrics xxx (2012) xxx–xxx

Table 2 4. Discussion
The sensitivity and specificity of MoCA-C and MMSE for predicting MCI and AD.

MCI AD Our results demonstrate that the MoCA-C is effective in

Sensitivity Specificity Sensitivity Specificity distinguishing patients with MCI patients from normal controls
subjects, and is associated with high test–retest, inter-rater
MoCA-C (cut-off)
19/20 0.01 1.00 0.65 1.00
reliability and internal consistency. We also found that the clinical
20/21 0.02 1.00 0.67 1.00 use of the MoCA-C was generally good in elderly Chinese elderly
21/22 0.06 1.00 0.74 1.00 with predominately low levels of education.
22/23 0.17 1.00 0.83 1.00 Our results demonstrate patients with MCI were significantly
23/24 0.20 1.00 0.88 1.00
younger than those with AD or normal cognitive function. Age is a
24/25 0.40 0.97 0.89 0.98
25/26 0.70 0.96 0.92a 0.96a risk factor for cognitive impairment. With the age increasing, the
26/27 0.92a 0.85a 0.96 0.85 prevalence of MCI also increased, MCI progress to AD annual
27/28 0.96 0.58 0.99 0.72 conversion rate was about 8% follow-up after 4–6 years (Hansson
28/29 0.99 0.31 0.99 0.47 et al., 2006). The results of our study also found that the age
29/30 1.00 0.06 1.00 0.21
influenced on the results, this is a limitation of the study. We
MMSE (cut-off)
19/20 0.01 1.00 0.57 1.00 should recruit the three groups whose mean ages were not
20/21 0.01 1.00 0.67 1.00 significantly different in the further study.
21/22 0.04 1.00 0.76 1.00 The moderately high level of correlation between the MMSE
22/23 0.17 0.98 0.88 1.00
and the MoCA-C in our study was similar to that reported by other
23/24 0.21 0.95 0.89 1.00
24/25 0.23 0.95 0.92a 1.00a researchers (Freitas, Simões, Alves, & Santana, 2011; Larner, 2011;
25/26 0.32 0.90 0.95 0.95 Sweet et al., 2011), This was in part expected, given the large
26/27 0.53 0.77 0.99 0.77 overlap of cognitive domains covered by both tests. There are,
27/28 0.76 0.59 1.00 0.53 however, differences in the structure and content of the MMSE and
28/29 0.85a 0.53a 1.00 0.38
MoCA-C. The MMSE is less capable than the MoCA-C in testing for
29/30 0.99 0.19 1.00 0.19
complex cognitive impairments in visuospatial, executive function
AD, Alzheimer’s disease; MCI, mild cognitive impairment; MoCA-C, Chinese version
and abstract reasoning domains. In addition, the MMSE sub-tests
of the Montreal Cognitive Assessment.
a
The optimal cut-off score.
for Attention and Delayed Recall contain test items which are less
challenging than those in the MoCA-C. For example, the only MMSE
test for attention is the Serial 7 s test while the MoCA-C includes

Table 3
MoCA-C subtest scores in NC group with different education levels.

Sub-test Number (%) subjects

Accuracya (n = 146) <6 years (n = 63) #6 years (n = 83) x2 P-value

Visuo-executive functions
Trail B 88 (60.27%) 29 (46.03%) 59 (71.08%) 9.388 0.002
Cube copying 74 (50.68%) 18 (28.57%) 56 (67.47%) 21.681 0.000
Clock drawing
3 score 92 (63.01%) 19 (30.16%) 74 (89.16%)
2 score 41 (28.08%) 32 (50.79%) 8 (9.64%)
1 score 13 (8.90%) 12 (19.05%) 1 (1.20%) 54.518 <0.001
Naming
Lion 137 (93.84%) 56 (88.89%) 81 (97.59%) 3.304 0.069
Rhinoceros 85 (58.22%) 19 (30.16%) 56 (67.47%) 19.959 <0.001
Camel 120 (82.19%) 45 (71.43%) 75 (90.36%) 8.771 0.003
Attention
Forward order 146 (100.00%) 63 (100.00%) 83 (100.00%)
Backward order 140 (95.89%) 58 (92.06%) 82 (98.80%) 2.587 0.108
Vigilance 128 (87.67%) 50 (79.37%) 78 (93.98%) 7.074 0.008
Serials 7 s
3 score 120 (82.19%) 43 (68.25%) 77 (92.77%)
2 score 15 (10.27%) 12(19.05%) 3 (3.61%)
1 score 11 (7.53%) 8(12.70%) 3 (3.61%) 14.845 0.001
Language
Sentence repetition 1 79 (54.11%) 32 (50.79%) 47 (56.63%) 0.491 0.484
Sentence repetition 2 112 (76.71%) 46 (73.02%) 66 (79.52%) 0.848 0.357
Verbal fluency 128 (87.67%) 55 (87.30%) 73 (87.95%) 0.014 0.906
Abstraction
Abstraction 1 120 (82.19%) 51 (80.95%) 69 (83.13%) 0.116 0.733
Abstraction 2 61 (41.78%) 17 (26.98%) 44 (53.01%) 15.151 <0.001
Recall
Face 97 (66.44%) 20 (31.75%) 77 (92.77%) 59.816 <0.001
Velvet 59 (40.41%) 11 (17.46%) 48 (57.83%) 24.24 <0.001
Church 81 (55.48%) 23 (36.51%) 58 (69.88%) 16.148 <0.001
Daisy 115 (78.77%) 45 (71.43%) 70 (84.34%) 3.568 0.059
Red 130 (89.04%) 55 (87.30%) 75 (90.36%) 0.344 0.558
Orientation
6 score 80 (54.79%) 30 (47.62%) 50 (60.24%)
5 score 52 (35.62%) 23 (36.51%) 29 (34.94%)
4 score 14 (9.59%) 10 (15.87%) 4 (4.82%) 5.63 0.060
a
Overall number of correct responses.

Please cite this article in press as: Hu, J.-b., et al., Cross-cultural difference and validation of the Chinese version of Montreal Cognitive
Assessment in older adults residing in Eastern China: Preliminary findings. Arch. Gerontol. Geriatr. (2012), http://dx.doi.org/10.1016/
j.archger.2012.05.008
G Model
AGG-2719; No. of Pages 6

J.-b. Hu et al. / Archives of Gerontology and Geriatrics xxx (2012) xxx–xxx e5

two additional tests (Digit Span and Vigilance). Similarly, the 3- There are several limitations to our study. The small popula-
item Delayed Recall in the MMSE is less difficult than 5-item tions were all selected from urban areas. It is, therefore, planned to
Delayed Recall in the MoCA-C. undertake a follow up study in elderly subjects from rural areas. It
The optimal cut-off of 26 in our study was similar to that used is also necessary to redefine cut-off rates based on age and
by others (Damian, Jacobson, & Hentz, 2011; Nasreddine et al., education-specific normative values so that individual test items
2005). The diagnostic accuracy of MoCA-C (92% sensitivity and 85% can be modified appropriately, enabling better content validity.
specificity) was well balanced when a cut-off of 26 was applied to This may eventually result in the development of a new screening
differentiate MCI from normal controls, and in this respect it tool specific to Chinese culture.
compared favorably compared with the MMSE (85% sensitivity and
53% specificity). Furthermore, the diagnostic accuracy of the Conclusions
MoCA-C in our study was almost the same as that of the original
version of MoCA for MCI (90% sensitivity and 87% specificity) and We have demonstrated that the current version of the MoCA-C
AD (100% sensitivity and 87% specificity). is superior to the MMSE in screening for MCI. The test can be
In terms of detecting AD from NC at the optimal cut-off points, recommended in primary clinical setting and geriatric health
the diagnostic accuracy of the MoCA-C (92% sensitivity and 96% screening in the community. However, due to the differences in
specificity) was similar to that seen with the MMSE (94% Western culture, further revision is required to tailor the contents
sensitivity and 95% specificity). more accurately to local populations.
Some patients with MCI and AD overlap on the MoCA-C, and in
these cases the MoCA-C, as well as the original version of MoCA, is Conflict of interest statement
unable to distinguish MCI from AD. Other difficulties that can
hamper early detection of MCI, include unrecognized symptoms or None.
misattribution of symptoms to the normal aging process, denial or
unawareness of symptoms, and allowing inadequate time for Role of the funding source
conducting appropriate evaluations. These factors all lead to errors
that can delay the early diagnosis of MCI. For these reasons, real life The design and conduct of the study, collection, management,
testing in the setting of a memory disorders clinic, a lower analysis, and interpretation of the data, and preparation, review,
threshold may be associated with better predictive value. and approval of the manuscript was jointly supported by a grant
Consistent with previous reports (Lee et al., 2008; Wong et al., 2008B091 from Medical scientific research foundation project
2009), we found that performance on the MoCA-C was affected by of Zhejiang Province, grant 2010R50049 from Department of
education level, but not by age or gender. As stated by Wong et al. Science and Technology Foundation of Zhejiang Province, grant
(2009), elderly persons in many Asia countries received much less Y201120182 from Department of Education Foundation of
education than their counterparts in Western countries. The Zhejiang Province.
average education level reported by LEE et al. in a Korean sample
and also in our study was 7–8 years which is much lower than the Acknowledgments
12 years of education reported in Caucasian studies (Luis et al.,
2009; Nasreddine et al., 2005). In our study 43.2% of the subjects We acknowledge our Research Consultant Dr John Snowdon
fell below this educational level. An education cut-off at 12 years as from the Department of Psychological Medicine, University of
proposed by the original MoCA study was not relevant to our study Sydney for comments on the research and for help in writing the
as only 21.9% of our subjects were above this cut-off. We therefore paper. We also thank Yi Shen for statistical advice and express our
adjusted the educational level subgroups around a cut-off of gratitude to the community social workers and subjects who
6 years, which translates to the completion of primary education in participated in the study.
the local Chinese context.
Although we used a Chinese version of MOCA (Wen, Zhang, Niu, References
& Lim, 2008), in our study we still found several cultural
Cumming, T. B., Bernhardt, J., & Linden, T. (2011). The Montreal cognitive assessment:
differences in practice. Subjects with lower levels of education
Short cognitive evaluation in a large stroke trial. Stroke, 42, 2642–2644.
do not understand how to perform the ‘Alternating Trail Making’ Damian, A. M., Jacobson, S. A., & Hentz, J. G. (2011). The Montreal Cognitive Assessment
test. This has also been reported in a study in subjects from Hong and the mini-mental state examination as screening instruments for cognitive
Kong (Wong et al., 2009). We therefore propose a conversion to the impairment: Item analyses and threshold scores. Dementia and Geriatric Cognitive
Disorders, 31, 126–131.
color changes, rather than characters sequence, similar to that used Dong, Y., Sharma, V. K., Chan, B. P., Venketasubramanian, N., Teoh, H. L., Seet, R. C., et al.
in the ‘Color Trails Test 2’ (Elia, Satz, Uchiyama, & White, 1996). In (2010). The Montreal Cognitive Assessment (MoCA) is superior to the Mini-Mental
the ‘Naming tests’, 58.2% of our subjects were unable to name the State Examination (MMSE) for the detection of vascular cognitive impairment after
acute stroke. Journal of the Neurological Sciences, 299, 15–18.
‘rhinoceros’ and ‘camel’ since the animal is not within the realm of American Psychiatric Association. (2000). Diagnostic and Statistical Manual of Mental
their general understanding or vocabulary. We therefore propose Disorders, IV-TR. Washington, DC: American Psychiatric Association.
replacing these with ‘tiger’ and ‘panda’ both of which are Elia, L. F. D., Satz, P., Uchiyama, C. L., & White, T. (1996). Color Trails Test Professional
Manual. Florida: Psychological Assessment Resources Inc.
indigenous to Chinese, used with medium frequency in the Freitas, S., Simões, M. R., Alves, L., & Santana, I. (2011). Montreal Cognitive Assessment
Chinese language and distinctly identifiable in pictures. In the (MoCA): Normative study for the Portuguese population. Journal of Clinical and
word memory test, the original words are not easy to memorize for Experimental Neuropsychology, 33, 989–996.
Fujiwara, Y., Suzuki, H., Yasunaga, M., Sugiyama, M., Ijuin, M., Sakuma, N., et al. (2010).
Chinese people especially the words ‘velvet’ and ‘church’, which
Brief screening tool for mild cognitive impairment in older Japanese: Validation of
fall outside Chinese cultural background. We recommend that the Japanese version of the Montreal Cognitive Assessment. Geriatrics & Gerontol-
these two terms are replaced by ‘silk’ and ‘temple’ which are both ogy International, 10, 225–232.
Gauthier, S., Reisberg, B., Zaudig, M., Petersen, R. C., Ritchie, K., Broich, K., et al. (2006).
more culturally appropriate. In the ‘Abstract thinking’ test the
Mild cognitive impairment. Lancet, 367, 1262–1270.
‘watch’ and ‘ruler’ were only correctly remembered by 41.8% of Gaviria, M., Pliskin, N., & Kney, A. (2011). Cognitive impairment in patients with
subjects, and even among those of a higher level of education they advanced heart failure and its implications on decision-making capacity. Conges-
were only correctly identified 50% of the time. These observations tive Heart Failure, 17, 175–179.
Gill, D. J., Freshman, A., Blender, J. A., & Ravina, B. (2008). The Montreal cognitive
highlight the difficulty in separating the influence of culture from assessment as a screening tool for cognitive impairment in Parkinson’s disease.
education level in the current MoCA-C test. Movement Disorders, 23, 1043–1046.

Please cite this article in press as: Hu, J.-b., et al., Cross-cultural difference and validation of the Chinese version of Montreal Cognitive
Assessment in older adults residing in Eastern China: Preliminary findings. Arch. Gerontol. Geriatr. (2012), http://dx.doi.org/10.1016/
j.archger.2012.05.008
G Model
AGG-2719; No. of Pages 6

e6 J.-b. Hu et al. / Archives of Gerontology and Geriatrics xxx (2012) xxx–xxx

Hanley, J., & McNeil, B. (1983). A method of comparing the areas under receiver with ‘‘normal’’ global cognition according to mini-mental state examination score.
operating characteristic curves derived from the same cases. Radiology, 148, Journal of the American Geriatrics Society, 57, 304–308.
839–843. Olson, R. A., Chhanabhai, T., & McKenzie, M. (2008). Feasibility study of the Montreal
Hansson, O., Zetterberg, H., Buchhave, P., Londos, E., Blennow, K., & Minthon, L. (2006). Cognitive Assessment (MoCA) in patients with brain metastases. Supportive Care in
Association between CSF biomarkers and incipient Alzheimer’s disease in patients Cancer, 16, 1273–1278.
with mild cognitive impairment: A follow-up study. Lancet Neurology, 5, 228–234. Petersen, R. C., Stevens, J. C., Ganguli, M., Tangalos, E. G., Cummings, J. L., & DeKosky, S. T.
Larner, A. J. (October 2011). Screening utility of the Montreal Cognitive Assessment (2001). Practice parameter: Early detection of dementia: Mild cognitive im-
(MoCA): In place of – oras well as – the MMSE? International Psychogeriatrics, 3, pairment (an evidence based review). Report of the Quality Standards Subcom-
1–6. mittee of the American Academy of Neurology. Neurology, 56, 1133–1142.
Lee, J. Y., Dong, W. L., Cho, S. J., Na, D. L., Hong, J. J., & Maeng, J. C. (2008). Brief screening Popovic, I. M., Seric, V., & Demarin, V. (2007). Mild cognitive impairment in symptom-
for mild cognitive impairment in elderly outpatient clinic: Validation of the Korean atic and asymptomatic cerebrovascular disease. Journal of the Neurological Sciences,
version of the Montreal Cognitive Assessment. Journal of Geriatric Psychiatry and 257, 185–193.
Neurology, 21, 104–110. Ritchie, K., Artero, S., & Touchon, J. (2001). Classification criteria for mild cognitive
Levey, A., Lah, J., Goldstein, F., Steenland, K., & Bliwise, D. (2006). Mild cognitive impairment: A population-based validation study. Neurology, 56, 37–42.
impairment: An opportunity to identify patients at high risk for progression to Sweet, L., Van Adel, M., Metcalf, V., Wright, L., Harley, A., Leiva, R., et al. (2011). The
Alzheimer’s disease. Clinical Therapeutics, 28, 991–1001. Montreal Cognitive Assessment (MoCA) in geriatric rehabilitation: Psychometric
Luis, Cheryl A., Keegan, Andrew P., & Mullan, Michael. (2009). Cross validation of the properties and association with rehabilitation outcomes. International Psychoger-
Montreal Cognitive Assessment in community dwelling older adults residing in the iatrics, 23, 1582–1591.
Southeastern US. International Journal of Geriatric Psychiatry, 24, 197–201. Videnovic, A., Bernard, B., Fan, W., Jaglin, J., Leurgans, S., & Shannon, K. M. (2010). The
McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D., & Stadlan, E. M. (1984). Montreal Cognitive Assessment as a screening tool for cognitive dysfunction in
Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA Work Huntington’s disease. Movement Disorders, 25, 401–404.
Group under the auspices of Department of Health and Human Services Task Force Wen, H. B., Zhang, Z. X., Niu, F. S., & Li, L. (2008). The application of Montreal cognitive
on Alzheimer’s Disease. Neurology, 34, 939–944. assessment in urban Chinese residents of Beijing. Zhonghua Nei Ke Za Zhi, 47,
Nasreddine, Z. S., Phillips, N. A., Bédirian, V., Charbonneau, S., Whitehead, V., Collin, I., 36–39.
et al. (2005). The Montreal Cognitive Assessment, MoCA: A brief screening tool for Wong, A., Xiong, Y. Y., Kwan, P. W., Chan, A. Y., Lam, W. W., Wang, K., et al. (2009). The
mild cognitive impairment. Journal of the American Geriatrics Society, 53, 695–699. validity, reliability and clinical utility of the Hong Kong Montreal Cognitive
Nazem, S., Siderowf, A. D., Duda, J. E., Have, T. T., Colcher, A., & Weintraub, D. (2009). Assessment (HK-MoCA) in patients with cerebral small vessel disease. Dementia
Montreal Cognitive Assessment performance in patients with Parkinson’s disease and Geriatric Cognitive Disorders, 28, 81–87.

Please cite this article in press as: Hu, J.-b., et al., Cross-cultural difference and validation of the Chinese version of Montreal Cognitive
Assessment in older adults residing in Eastern China: Preliminary findings. Arch. Gerontol. Geriatr. (2012), http://dx.doi.org/10.1016/
j.archger.2012.05.008