Original Research Article Accepted: April 19, 2010 Published online: June 5, 2010
Dement Geriatr Cogn Disord 2010;29:516–522 DOI: 10.1159/000313981
Validity and Reliability of the Persian Language Version of the Neuropsychiatry Unit Cognitive Assessment Tool M. Barekatain a M. Walterfang b, c M. Behdad a M. Tavakkoli a J. Mahvari d M.R. Maracy e D. Velakoulis b, c
a
Department of Psychiatry and Behavioral Sciences Research Center, Isfahan University of Medical Sciences, Isfahan, Iran; b Neuropsychiatry Unit, Royal Melbourne Hospital, and c Melbourne Neuropsychiatry Centre, University of Melbourne, Parkville, Vic., Australia; d Department of Neurology, e Faculty of Health and Behavioral Sciences Research Center, Isfahan University of Medical Sciences, Isfahan, Iran
Key Words Cognition ⴢ Screening ⴢ Neurocognition ⴢ Mini-Mental State Examination
specificity of 100%) patients from controls. Conclusions: The Persian language NUCOG appears to perform strongly in an unselected population, reliably differentiating patients with dementia from controls, and detecting cognitive impairment in a range of clinical disorders.
Abstract Background/Aims: Only a limited number of cognitive screening tools are available for the Persian-speaking population, and we sought to translate and validate the Neuropsychiatry Unit Cognitive Assessment Tool (NUCOG), a multidimensional cognitive screening tool. Methods: We used multiple language specialists to translate and then back-translate the NUCOG, and administered the Persian language NUCOG and Mini-Mental State Examination (MMSE) to 184 individuals: 60 controls and 124 patients, 33 of whom had dementia, 30 non-dementing neurological disorders and 61 a psychiatric illness. Results: The NUCOG outperformed the MMSE in differentiating the patient groups and controls. The ‘profile’ across the 5 NUCOG domains differentiated dementia subgroups such as senile dementia of the Alzheimer type (SDAT), frontotemporal dementia and mild cognitive impairment (MCI). Psychiatric patients with psychosis and posttraumatic stress disorder were more impaired than patients with affective disorders. The NUCOG reliably differentiated controls from patients with MCI (at 86.5/100, sensitivity of 83.3% and specificity of 87.5%) and SDAT (at 75/100, sensitivity and
Copyright © 2010 S. Karger AG, Basel
© 2010 S. Karger AG, Basel 1420–8008/10/0296–0516$26.00/0 Fax +41 61 306 12 34 E-Mail
[email protected] www.karger.com
Accessible online at: www.karger.com/dem
Introduction
Many instruments are available for the clinician who wants to undertake a brief and rapid screening of cognitive status. Ideally, a cognitive tool balances breadth, depth and brevity, enabling the evaluation of most domains of cognition in the shortest possible time. The reliable assessment of cognition in psychiatric and neurological practice not only allows for the detection of significant cognitive impairment when it exists, but may add significantly to the diagnostic process, providing additional clinical information that allows for the differentiation of clinical neurocognitive disorders. The Mini-Mental State Examination (MMSE) [1] is the most widely used tool both by primary care physicians and specialists. The MMSE does not require specialized training or specific equipment, and its administration time is short. The MMSE is sensitive to cognitive impairDr. Mark Walterfang Level 2, John Cade Building Royal Melbourne Hospital, Grattan Street Parkville, Vic. 3050 (Australia) Tel. +61 3 9342 8750, Fax +61 3 9342 8483, E-Mail mark.walterfang @ mh.org.au
Table 1. The 5 different cognitive domains of the NUCOG, and individual items that make up each domain Attention
Visuoconstructional
Memory
Executive
Language
Orientation Digit span – forward Digit span – reverse Days of the week in reverse
Drawing reproduction Limb and orobuccal praxis Left/right orientation Visual/somatosensory neglect Calculation
Verbal memory – registration and recall Spatial recall Remote memory
Motor sequencing Categorical fluency Abstraction Managing interference
Verbal comprehension Repetition Object naming Sentence writing Written comprehension Word finding
ment in a range of disorders [2]. It has been translated and validated in several languages including Persian, the Iranian language, for clinical screening and research purposes in the Persian-speaking population [3]. The MMSE has a number of limitations as a screening tool in the broad range of patients who may present with cognitive impairment in a conventional clinical setting [4–6]. Most notably, executive functioning is not evaluated, and the MMSE has no assessment of delayed spatial recall [7]. Furthermore, it has a reduced sensitivity to early cognitive impairment (ceiling effect), in addition to a poor sensitivity to cognitive decline in patients that have severe cognitive impairment (floor effect) [5]. Finally, it results in a single unitary score, with a variable contribution to the final MMSE score by the domains of language, attention and memory, and it lacks graded scoring in a number of cognitive domains [7]. Although other valid and reliable tools such as the Neurobehavioral Cognitive Status Examination [8], Addenbrooke’s Cognitive Examination (ACE) [9, 10] and the modified MMSE [11] have been introduced to overcome some of the structural limitations of the MMSE, they too have strengths and weaknesses including limited sensitivity in some patient populations, lack of multidimensional scoring, and limited executive function testing. Walterfang et al. [12] attempted to address some of these limitations in the design and implementation of the Neuropsychiatry Unit Cognitive Assessment Tool (NUCOG). The NUCOG was designed to be a brief and portable bedside cognitive screening tool with a high degree of face validity, covering all major cognitive domains, generating graded scores for specific cognitive functions, and providing multidimensional scoring and the generation of a cognitive ‘profile’. The NUCOG tests cognitive functions in five main areas: attention, memory, executive functioning, language and visuoconstructional function, utilizing a number of tests in each domain (table 1). Each of the five domains has a maximum score of 20, yielding a maximum possible total score of 100. It showed
a high correlation with the MMSE and formal neuropsychological testing in a large mixed sample of neuropsychiatric patients and controls, discriminated broad patient groups, reliably differentiated between demented and non-demented patients, and showed superior specificity and sensitivity for the detection of significant cognitive impairment when compared with the MMSE [12, 13]. To date, the MMSE is the only cognitive screening instrument validated for use in Iran. We sought to translate the NUCOG into Persian and investigated the validity and reliability of the Persian version of the NUCOG in patients with dementia, major psychiatric disorders and neurological disorders compared to healthy control subjects. Additionally, we sought to compare the performance of the Persian NUCOG in this population with the Persian MMSE.
Validity and Reliability of the Persian NUCOG
Dement Geriatr Cogn Disord 2010;29:516–522
Methodology The protocol of this study was approved by the Research and Ethics Council of the Behavioral Sciences Research Center, Isfahan University of Medical Sciences, Isfahan, Iran. Participants The total sample for the study consisted of 184 subjects. The patient group had been selected randomly from adult inpatients and outpatients who were referred to Noor, Al-Zahra and Kashani Hospitals, affiliated to the Isfahan University of Medical Sciences, for psychiatric or neurological evaluation between November 2008 and May 2009. The patients (n = 124) consisted of 3 distinct groups: patients with major psychiatric disorders (n = 61), patients with dementia (n = 33), and non-demented neurological patients (n = 30). Of the psychiatric patients, the main diagnostic groups were primary psychotic disorders (n = 9), major depressive disorder (n = 22), bipolar affective disorder (n = 15), posttraumatic stress disorder (n = 9) and other psychiatric disorders (n = 6). The demented group was made up of senile dementia of the Alzheimer type (SDAT, diagnosed according to NINCDS-ADRDA criteria; n = 17), mild cognitive impairment (MCI, diagnosed as patients with pure amnestic memory impairment in the setting of otherwise normal functioning and activities of daily living; n = 8), fron-
517
Table 2. Baseline demographic characteristics, MMSE and NUCOG total scores, and NUCOG subscale scores
Agea, years Educationa, years Maleb, n (%) MMSE totala NUCOG totala NUCOG attentiona NUCOG visuospatiala NUCOG memorya NUCOG executivea NUCOG languagea
Neurological (n = 30)
Psychiatric (n = 61)
Dementia (n = 33)
Control (n = 60)
49.57814.75 11.0783.10 17 (56.66) 25.7783.54 78.85811.07 13.3083.26 17.1382.17 14.7083.04 15.0583.14 18.6681.47
53.11813.80 10.6483.27 36 (59.01) 26.8482.59 80.95810.99 13.6782.94 17.5782.38 14.7183.27 15.9583.18 19.0381.42
59.5289.56 8.5283.34 20 (60.60) 23.5284.34 63.72816.85 9.9084.46 14.3984.13 10.8083.90 10.7184.02 17.9082.51
61.0888.69 10.9383.43 32 (53.33) 28.9381.31 90.5484.38 15.7581.25 18.9381.10 17.6081.48 18.5581.41 19.7080.56
A ll figures except male gender are presented as means 8 SD. a Kruskal-Wallis test, p < 0.003. b 2 test, p > 0.05.
totemporal dementia (FTD, diagnosed according to Neary’s consensus criteria; n = 6) and other dementing disorders (n = 2). The neurological group was made up of patients with central nervous system tumours (n = 6), multiple sclerosis (n = 5), epilepsy (n = 5), stroke (n = 4), Parkinson’s disease (n = 4), head trauma (n = 3) and hypoxic or metabolic brain injury (n = 3). The Mini-International Neuropsychiatric Interview (MINI) was used to confirm the diagnosis [14]. Healthy control subjects (n = 60) were randomly selected from relatives and carers of patients of neurology and psychiatry clinics who had never been diagnosed with a psychiatric and neurological illness, and who were negative when screened by the MINI [14], and by physical and neurological examination. All subjects had passed elementary school, had clear consciousness and normal sensory perception. Measures Mini-Mental State Examination. The MMSE is a well-established cognitive screening tool [1]. The Persian version of this tool has been demonstrated as reliable (Cronbach’s ␣ = 0.81) and valid in a dementia population [3]. Neuropsychiatry Unit Cognitive Assessment Tool. The NUCOG evaluates cognition in 5 domains (attention, visuoconstructional functioning, memory, executive functioning and language). NUCOG scores have been shown to correlate strongly with the MMSE (r = 0.894; p ! 0.001) and the internal consistency of the NUCOG is very high (Cronbach’s ␣ = 0.915) [13]. The NUCOG was translated into Persian by 2 experienced biological psychiatrists (M. Barekatain and M. Behdad) and tested by backtranslation into English. Mini-International Neuropsychiatric Interview. The MINI is a brief standardized psychiatric procedure, focusing predominantly on current axis I DSM-IV disorders using an efficient series of probe questions [14]. Translation Procedure Two native Pesian-speaking specialists translated the NUCOG independently from English to Persian. The differences in the initial translation were resolved by consensus in a session with a third interpreter. Then, back-translation of the translated Per-
518
Dement Geriatr Cogn Disord 2010;29:516–522
sian version was undertaken by 2 English-speaking linguists who were blinded to the original English language NUCOG and the aim of this study. The differences were resolved via consensus amongst the 5 translators. Language-Based Alterations The following changes were made due to differing linguistic properties of the 2 languages: as the days of the week in Persian are ordered by number (‘yek-shanbe, do-shanbe, se-shanb’, etc., translating into ‘one-day, two-day, three-day’, etc.), we replaced reverse-naming of days of the week with months of the year based on the Iranian calendar (‘Esfand, Bahman, Day, Azar, Aban, Mehr, Shahrivar’). Regarding the vocabulary, we replaced ‘guitar’ with ‘setar’, a popular musical instrument similar to the guitar. We also replaced ‘banana’ with ‘ananas’ (pineapple), ‘artillery’ with ‘baarbary’ (handling), and ‘constitutional’ with ‘farsishenasi’ (the knowledge of Persian language syntax and semantics). To choose suitable synonymous sentences for ‘the orchestra played and the audience applauded’ and ‘no ifs, ands, or buts’ from the NUCOG, a list of possible surrogates were tested through a pilot study and the most appropriate ones were chosen. Statistical Analysis Demographic and cognitive data were compared using the Kruskal-Wallis test, and the 2 test was used for gender proportion. Criterion validity was assessed by receiver operator characteristic (ROC) curves. The influence of demographic variables on the NUCOG scores was calculated using Pearson’s correlation coefficient. Convergent validity was evaluated using Spearman’s correlation coefficient. Internal consistency was measured by Cronbach’s ␣. Comparison of the performance of the MMSE, NUCOG and its subscales across patient group samples was undertaken using one-way analysis of covariance (ANCOVA) with post hoc comparisons undertaken by the Games-Howell procedure. Repeated-measures ANCOVA, controlling for age and years of education, was used for group-by-subscale analyses, with interaction effects tested using the Greenhouse-Geisser test.
Barekatain/Walterfang/Behdad/ Tavakkoli/Mahvari/Maracy/Velakoulis
Results 20
The demographic characteristics and cognitive performance of the sample are outlined in table 2.
Relationship between Demographic and Cognitive Variables In the control group, MMSE scores were significantly correlated with age (r = –0.488; p ! 0.0001) and years of education (r = 0.564; p ! 0.0001). Although the MMSE scores did not significantly correlate with age across the patient group as a whole (r = –0.129; p 1 0.05), there was a significant correlation between years of education and MMSE scores (r = 0.203; p ! 0.05). The MMSE score showed a significant difference between the 3 patient groups (p ! 0.003; dementia ! neurologic = psychiatric ! controls). Similarly, the total NUCOG scores in the control group revealed significant correlations with age (r = –0.500; p ! 0.0001) and years of education (r = 0.691; p ! 0.0001). There was a significant correlation between the total NUCOG score and years of education (r = 0.364; p ! 0.0001) but not age (r = 0.101; p 1 0.05) in the entire patient group. The total NUCOG score showed significant differences in scoring between the 3 patient groups (F2, 123 = 20.387; p ! 0.0001; dementia ! neurologic = psychiatric ! control). Distinguishing Patient Groups by Test An interaction of test by diagnostic group was found in favor of the NUCOG across the 4 study groups (F3, 180 = 56.67; p ! 0.0001) and the 3 patient groups (F2, 123 = 17.04; p ! 0.0001). However, the NUCOG showed no advantage over the MMSE in differentiating psychiatric from neurological patients (F1, 90 = 0.584; p = 0.447). Performance of NUCOG Subscales There was a significant subscale-by-group effect (F3, 183 = 319.10; p ! 0.0001), suggesting that different subject groups showed distinctly different ‘cognitive profiles’ (fig. 1). The NUCOG profile differentiated dementia subgroups, with the FTD and SDAT groups being most impaired compared to other patients, particularly in executive functioning (F3, 32 = 2.733; p ! 0.005) (fig. 2), although Validity and Reliability of the Persian NUCOG
15
Scale score
Demographic Data There were significant differences between groups in age (p ! 0.0001; neurologic ! psychiatric ! dementia ! controls) and years of education (p ! 0.005; dementia ! psychiatric ! neurologic ! controls). Gender was not significantly different across groups (p 1 0.05).
10
5
Group Psychiatric Neurologic Dementia Control
0 Attention Visuoconstructional
Memory
Executive
Language
Fig. 1. NUCOG profile across 4 main subject groups.
the SDAT group performed more poorly on memory and visuoconstructional function than the FTD group. When the psychiatric group was divided into its main constituent diagnostic groups, the cognitive profile again differentiated psychiatric subgroups, with psychotic disorder and posttraumatic stress disorder patients being the most impaired (F4, 60 = 1.741; p ! 0.05) (fig. 3). As in the initial NUCOG validation study, most patient groups performed at a higher level on the language subscale than on the other 4 subscales. Criterion Validity In the patient group, the sensitivity and specificity of the NUCOG in detecting dementia was evaluated using ROC methods. In comparing demented with non-demented patients, the NUCOG more robustly differentiated the patient groups, with the areas under the ROC curve being 0.878 for the NUCOG and 0.810 for the MMSE (p ! 0.001) (fig. 4). An optimum cutoff for dementia on the NUCOG of 77.5 gave a sensitivity of 79.5% and specificity of 81.8% for the detection of dementia, whereas the optimum cutoff in this sample on the MMSE was a score of 26.5, with a sensitivity of 71.5% and specificity of 72.7%. When the comparison was restricted to the SDAT group versus controls, the NUCOG perfectly sepaDement Geriatr Cogn Disord 2010;29:516–522
519
20
15
15
Scale score
Scale score
20
10
10
Diagnosis 5
Dementia group SDAT FTD MCI Other
0 Attention Visuoconstructional
Memory
Executive
Language
5 Psychotic disorder Depressive disorder Bipolar disorder Posttraumatic stress disorder Other psychiatric disorder
0 Attention Visuoconstructional
Domain
Memory
Executive
Language
Domain
Fig. 2. NUCOG profile across dementia patient subgroups.
Fig. 3. NUCOG profile across psychiatric patient subgroups.
rated the 2 groups (p ! 0.0001); a cutoff score of 75 gave a sensitivity and specificity of 100% (with the MMSE also differentiating the groups, with a cutoff of 26.5 showing a sensitivity of 95% and specificity of 94.1%). When the MCI group was compared with controls, the NUCOG robustly differentiated the groups (p ! 0.0001); a cutoff of 86.5 gave a sensitivity of 83.3% and specificity of 87.5%, whereas an MMSE cutoff of 28.5 provided a sensitivity of 68.3% and specificity of 87.5% (p ! 0.001).
COG was analysed (p = 0.81), and no site-by-group interaction (p = 0.56), suggesting that the validity of the tool was maintained after translation. However, when MMSE scores were analysed, an overall effect of site was shown (p ! 0.01), with the Melbourne group overall demonstrating a lower mean MMSE score (25.84 vs. 26.75). Furthermore, a site-by-group interaction (p = 0.005) was shown, with a lower mean MMSE score in the Melbourne dementia group (20.22 vs. 23.52). As for the Melbourne dataset, we generated a cognitive ‘profile’ that displays the mean and up to 3 SD from this mean (fig. 5).
Convergent Validity The NUCOG and MMSE were strongly correlated across the entire sample (r = 0.922; p ! 0.0001), which remained significant when controlling for age and years of education using Pearson’s partial correlation test (r = 0.918; p ! 0.0001). Reliability Using Cronbach’s ␣, the internal consistency of the NUCOG across the whole sample was 0.924. Comparison with Original Dataset When the dataset was compared with the original NUCOG validation dataset (n = 342), ANCOVA showed no main effect of site (Melbourne vs. Iran) when the NU520
Dement Geriatr Cogn Disord 2010;29:516–522
Discussion
A number of cognitive screening tools are in widespread use in clinical practice to aid in the detection and accurate diagnosis of cognitive impairment. Many of these tools have clinical limitations such as ceiling and floor effects or a lack of breadth in cognitive testing. As the MMSE is the most widely available cognitive screening tool translated and validated in Persian, we sought to translate and validate a tool with a number of attractive psychometric properties including a greater sensitivity, greater breadth in cognitive testing, and the additional Barekatain/Walterfang/Behdad/ Tavakkoli/Mahvari/Maracy/Velakoulis
1.0
20
Mean 1 SD 2 SD 3 SD
0.9
18 0.8
16 14
0.6
Scale score
Sensitivity
0.7
0.5 0.4
12 10 8 6
0.3
Source of the curve 4
0.2
NUCOG MMSE Reference line
0.1
2 0
0 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1 – specificity
Attention
Visuoconstructional
Memory
Executive
Language
Domain
Fig. 4. ROC curves comparing the NUCOG and MMSE in their
Fig. 5. Cognitive profile of the control group, showing mean
differentiation of demented from non-demented patients.
scores across the 5 NUCOG domains (thick line) and profiles of 1–3 SD below the mean.
utility afforded by the use of multidimensional scoring, resulting in a ‘cognitive profile’. We have shown that in an unselected, mixed sample of psychiatric, neurological and dementia patients, the Persian translation of the NUCOG showed greater sensitivity and specificity in diagnosing dementia when compared to the MMSE, and was better at differentiating most patient groups. Additionally, the cognitive profile differed significantly amongst the psychiatric and dementia subgroups, suggesting that the NUCOG may offer clinicians additional utility in the diagnostic process. Besides the MMSE, the only other cognitive assessment tool available in Persian is the ACE, which has recently also been validated in a Persian-speaking population although this study used a smaller sample size (n = 135) and only included controls, individuals diagnosed with MCI and patients with SDAT, with the aim of distinguishing these three groups from each other [15]. This study showed that the ACE strongly differentiated MCI and SDAT from the healthy control group, and that the ACE significantly outperformed the MMSE with regard to sensitivity and specificity in discriminating SDAT and MCI from healthy individuals. In our study, we demonstrated that the NUCOG also outperformed the MMSE
although our study aimed to be more ecologically valid in that our comparator group were non-demented individuals (including non-demented patients, not just healthy controls), which is more likely to be representative of an unselected sample that would be seen in inpatient and outpatient clinics where tools like the NUCOG are most used. When we examined the capacity of the NUCOG to differentiate between MCI, SDAT and control subjects, the NUCOG performed similarly. The lack of an effect of site (itself potentially a proxy for language) when the groups were compared between the original Melbourne validation sample and the Iran sample suggests that this translated version of the NUCOG appears to perform similarly to the English language version of the NUCOG across healthy individuals and patients with dementia, neurological illness and psychiatric disorders. The use of a cognitive ‘profile’ across cognitive domains is not unique to the NUCOG, but being able to plot an individual patient’s cognitive profile against the mean profile of a control group, and to determine how many SD below the mean an individual’s subscale score might fall, allows a clinician administering the NUCOG to rely on not just an overall score but a pattern across cognitive
Validity and Reliability of the Persian NUCOG
Dement Geriatr Cogn Disord 2010;29:516–522
521
domains to aid in his/her diagnosis. Unlike a unitary score – such as provided by the MMSE – a cognitive profile allows patients with different cognitive disorders (e.g. FTD and SDAT) to be more readily differentiated than they would otherwise be by a unitary score alone. Thus, a key advantage of a multidimensional tool such as the NUCOG may be that it provides significantly richer diagnostic information to a clinician faced with a patient with a cognitive disorder, and thus facilitates a more rapid and precise diagnosis. The brevity and portability of the NUCOG and other cognitive screening instruments are essential to their deployment across a wide range of health settings in a range of socioeconomic and cultural groups and in a variety of health systems. Very few health services worldwide, including the Iranian health system, have access to formal neuropsychological testing, and the availability of lin-
guistically and culturally appropriate cognitive screening tools like the NUCOG is essential to ensure the feasibility of a reliable assessment of patients presenting with primary cognitive disorders, or with neurological or psychiatric illnesses that may have cognitive sequelae. By expanding the spectrum of cognitive screening tools available to clinicians who assess and treat Persianspeaking patients, this patient group may have access to improved cognitive assessment, and ultimately to improved treatment of disorders associated with significant cognitive impairment. The Persian version of the NUCOG appears to be a reliable and clinically useful tool for use in the differentiation of patients with dementia from the healthy population, and also appears to detect the more subtle cognitive impairments seen in psychiatric and neurological illness in Persian as well as it does in its original English language version.
References 1 Folstein MF, Folstein SE, McHugh PR: ‘Mini-Mental State’: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;2:189–198. 2 Ramirez M, Teresi JA, Holmes D, Gurland B, Lantigua R: Differential item functioning (DIF) and the Mini-Mental State Examination (MMSE): overview, sample, and issues of translation. Med Care 2006;44:S95–S106. 3 Seyedian M, Fallah M, Norozian M, Nejat S, Delavar A, Ghasemzadeh H: Validity of the Farsi (Persian) Version of the Mini-Mental State Examination. J Med Counc Iran 2007; 25:408–414. 4 Ihl R, Frölich TD, Martin EM, Maurer K: Differential validity of psychometric tests in dementia of Alzheimer type. Psychiatry Res 1992;44:93–106. 5 Tombaugh TN, McIntyre NJ: The MiniMental State Examination: a comprehensive review. J Am Geriatr Soc 1992;40:922–935. 6 Wind AW, Schellevis FG, van Staveren G, et al: Limitations of the Mini-Mental State Examination in diagnosing dementia in general practice. Int J Geriatr Psychiatry 1997; 12:101–108.
522
7 Nadler J, Richardson E, Malloy P: Detection of impairment with the Mini-Mental State Examination. Neuropsychiatry Neuropsychol Behav Neurol 1994;7:109–114. 8 Kieman R, Mueller J, Langston J, van Dyke C: The Neurobehavioral Cognitive Status Examination: a brief but differentiated approach to cognitive assessment. Ann Intern Med 1987;107:481–485. 9 Mathuranath PS, Nestor PJ, Berrios GE, Rakowicz W, Hodges JR: A brief cognitive test battery to differentiate Alzheimer’s disease and frontotemporal dementia. Neurology 2000;55:1613–1620. 10 Dudas R, Berrios GE, Hodges J: The Addenbrooke’s Cognitive Examination (ACE) in the differential diagnosis of early dementias versus affective disorder. Am J Geriatr Psychiatry 2005;13:218–226. 11 Teng EL, Chui HC: The Modified MiniMental State (3MS) Examination. J Clin Psychiatry 1987;48:314–318.
Dement Geriatr Cogn Disord 2010;29:516–522
12 Walterfang M, Velakoulis D, Gibbs A, Lloyd J: The NUCOG: construction and piloting of a cognitive screening instrument in a neuropsychiatric unit. Aust Psychiatry 2003; 11: 325–329. 13 Walterfang M, Siu R, Velakoulis D: The NUCOG: validity and reliability of a brief cognitive screening tool in neuropsychiatric patients. Aust NZ J Psychiatry 2006; 40: 995–1002. 14 Sheehan DV, Lecrubier Y, Sheehan K, Amorim P, Janavs J, Weiller E, Hergueta T, Baker R, Dunbar GC: The Mini-International Neuropsychiatric Interview (MINI): the development and validation of a structured diagnostic interview for DSM-IV and ICD-10. J Clin Psychiatry 1998;59(suppl 20):22–33. 15 Pouretemad HR, Khatibi A, Ganjavi A, Shams J, Zarei M: Validation of Addenbrook’s Cognitive Examination (ACE) in a Persian-speaking population. Dement Geriatr Cogn Disord 2009;28:343–347.
Barekatain/Walterfang/Behdad/ Tavakkoli/Mahvari/Maracy/Velakoulis
