Validation into Arabic versions of Dementia Rating Scales, Dementia Caregivers Scales, and Dementia Research Instruments.

INTRODUCTION: This study aimed to examine the validity and internal consistency of Arabic versions of the eight-item Alzheimer's Dementia, Alzheimer Questionnaire, and Clinical Dementia Rating scales and to assess the Arabic version of Katz Activities of Daily Living, and Neuropsychiatric Inventory.
METHODS: One hundred fifty participants were recruited from different settings; they underwent clinical interviews and filled the aforementioned scales.
RESULTS: In our sample, 56.8% of the sample suffered from dementia. The Arabic eight-item Alzheimer's Dementia had excellent psychometric properties, and the Arabic Alzheimer Questionnaire showed near-perfect properties with sensitivity and specificity reaching 100%. In addition, the Arabic Clinical Dementia Rating (A-CDR)-sum of boxes was superior to the regular A-CDR score in detecting dementia cases among the study sample. The A-CDR showed similar characteristics as the original version. The Katz scores demonstrated a strong negative correlation with eight-item Alzheimer's Dementia scores.
CONCLUSION: Based on this study, health professionals now have reliable and validated tools to be used in clinical and research settings among Arabic-speaking populations.

Introduction

The total number of Arabic speakers in 2014 was estimated to be 366 million living in the Arab world (countries in the League of Arab States) and 17.5 million living in countries outside the Arab League [1], with a percentage of adults aged 60 years and older estimated around 6.5% [2]. Given the very rapid demographic aging in Arab countries, most of these countries are expected to have at least 15% of their populations aged 60 years and older by 2050 [2]. Consequently, the burden of diseases or conditions, which is common among older adults, is expected to grow exponentially. This is significant because these countries lack the social and health care policies to respond to the new challenges that accompany this demographic change. One of those challenges is the expected rapid increase in the number of people suffering from dementia in the North Africa and Middle East region, which is estimated to double from 1.2 million people in 2010 to 2.6 million in 2030 reaching 6.2 million in 2050 [3]. Dementia is a severe clinical syndrome characterized by inevitably progressive deterioration in cognitive ability and capacity for independent living. Despite the high estimated numbers of people with dementia in North Africa and the Middle East region [4], to our knowledge, only one study was done in the Arab countries to assess the prevalence of dementia [3], [5]. Studies regarding dementia in the Arab world are scarce, with very limited numbers of Arabic-validated diagnostic instruments. Therefore, the first step that can be taken to address the challenges posed by the growing numbers of people with dementia in the Arab world is to validate the most commonly used diagnostic instruments of dementia in Arabic, which subsequently can be used to fill the knowledge gap in dementia prevalence, and associated risk factors within the Middle East and North Africa region. Cognitive impairment is the key diagnostic criteria for dementia, of which the behavioral and functional aspects of the clinical picture of dementia derive from it. The impaired cognition will ultimately lead to impairments in the ability to carry out activities of daily living (function) and to behavioral and psychological symptoms. Consequently, this study aimed to validate in Arabic the most commonly used scales of these three areas: cognition, function, and behavior, with the hope of improving the diagnosis, treatment, and well-being of individuals affected by dementia.

Methods

Instruments translated in Arabic

Among the commonly used informant-based screening tools is the eight-item Alzheimer's Dementia (AD8) screening test, which is a brief, valid sensitive instrument that reliably discriminates signs of normal aging from mild dementia. AD8 is recommended as a key diagnostic tool to use in primary care, and it has been argued to be more appropriate for routine use in primary care than the Mini–Mental State Examination (MMSE) [6], [7]. It was originally validated as an informant-based interview; however, it also can be completed by the patient as a self-rating tool [8]. A high sensitivity of 84% and a specificity of 93% have been reported for the AD8 test, with excellent abilities in detecting early cognitive changes [8]. The AD8 has been translated from the original English version and validated into several languages, including French, Spanish, Portuguese, Chinese, Korean, and Taiwanese [9], [10], [11], [12], [13], [14].

The Alzheimer's Questionnaire (AQ) is another brief, informant-based assessment for cognitive impairment. From a clinical perspective, the AQ provides more information on the underlying etiology of the cognitive impairments than other screening tools. The AQ validation study revealed that the AQ has an excellent sensitivity of 89% and specificity of 91% for detecting amnestic mild cognitive impairment, and 99% and 96%, respectively, for Alzheimer's disease [15].

The Clinical Dementia Rating (CDR) is the gold-standard informant-based assessment scale with established reliability and validity, which has been widely used as a severity-ranking scale in many studies throughout the world [16]. From a diagnostic standpoint, the CDR is congruent with the Diagnostic and Statistical Manual of Mental Disorders approach of dementia diagnosis. However, the length of the interview makes it more practical to be used for research purposes rather than clinical purposes.

The Katz Index of Independence in Activities of Daily Living is an instrument developed to assess the functional status of older adults, as a standardized quantitative measure for evaluating, prognosis, treatment, and functional changes [17]. It is a long established widely used index and has been reported to be the most appropriate scale to assess patient's ability to perform activities of daily living independently [18]. It measures the client's ability to perform activities of daily living in six functions: bathing, dressing, toileting, transferring, continence, and feeding. Given that cognitive impairment in demented patients will ultimately lead to functional impairments, the Katz index scores among dementia patients are reported to be negatively correlated with the MMSE scores [19].

The Neuropsychiatric Inventory (NPI) is an informant-based instrument developed to assess the psychopathology in demented patients [20]. It is the most widely adopted and recommended clinical instrument for evaluating the behavioral and psychological symptoms of dementia [21]. It evaluates 12 neuropsychiatric disturbances common in dementia: delusions, hallucinations, agitation, dysphoria, anxiety, apathy, irritability, euphoria, disinhibition, aberrant motor behavior, nighttime behavior disturbances, and appetite and eating abnormalities. The severity and frequency of each neuropsychiatric symptom are rated on the basis of scripted questions administered to the patient's caregiver. The NPI also assesses the amount of caregiver distress engendered by each of the neuropsychiatric disorders.

Participants

Participants were recruited from outpatient clinics, inpatient units, and nursing home residents. We attempted to recruit 150 participants for this study—75 individuals suffering from dementia and 75 matched controls. Controls were matched according to sex, age (plus or minus 3 years), and educational background.

Every individual enrolled was interviewed by a clinical psychologist (J.C.D.) trained in performing a complete psychiatric examination or by a psychiatrist (N.S.S.). The diagnosis of dementia was based on the updated National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) [22]. In addition, people scoring less than 24 on the MMSE and not suffering from delirium were considered to suffer from memory impairment. Following the clinical interview, the translated scales were administered. The dementia rating scales were validated against this clinical interview.

Every individual was reevaluated by a blinded geriatric psychiatrist (G.E.K.) to establish a clinical diagnosis and control the quality of the interviews. Family members of all participants were also interviewed. This helped in further establishing the clinical diagnosis of dementia.

Translation

The scales were translated into Arabic by a bilingual professional translator and reviewed by four psychiatrists and three psychologists. Subsequently, the scales were back-translated into English, and any conflicts were resolved through consensus expert panel discussions. The translated questionnaires were modified until the back-translated versions were comparable with the original English versions.

Ethics

The study was approved by the institutional review board of the Balamand University/St. George Hospital University Medical Center, which is registered and certified in the United States. Before the participation of any individual, the study procedure was explained, and informed consents were obtained from both the participants and their caregivers. In the nondemented group, consent was obtained from the participant only.

Statistical analysis

Data analyses were performed using Stata/SE statistical software (version 11.1). All statistical tests were two tailed, and P values less than .05 were considered statistically significant. The scales were compared to each other and were compared to the gold-standard interview. The internal consistency of the scales (internal reliability) was computed using Cronbach's α coefficient. A receiver operating curve was derived for each scale, and the area under the curve (AUC) was used to find the best cutoff points for the scales. Accordingly, the sensitivity, specificity, the positive predictive values, and negative predictive values of the translated scales were calculated.

Results

The total sample included in the analysis was composed of 127 participants (60 recruited from outpatient clinics, 37 from inpatient units, and 30 from nursing home residents). It was extremely difficult to recruit people residing in nursing homes, who were not suffering from dementia, which lowered the final number included in the analysis. The average age of the participants at the time of assessment was 81.9 years with a standard deviation of 7.8 years, and it was lower in the inpatient group (Table 1). The sample was mainly formed by females (61.4%) and widowed (54.3%). Slightly more than half were retired (54%), and very few were working (7%). Gender, work status, marital status, and education level were not significantly different between participants recruited from the outpatient clinics, inpatient units, and nursing home residents (Table 1).

Table 1

Demographic characteristics across recruitment settings

Characteristics	Outpatients	Inpatients	Nursing homes	Total	P value
Number	60 (45.5%)	37 (28.0%)	35 (26.5%)	132	—
Age	82.9 (±6.0)	77.0 (±7.8)	85.3 (±8.1)	81.9 (±7.8)	<.00∗
Gender
Male	26 (43.3%)	13 (35.1%)	12 (34.3%)	51 (38.6%)	.60
Female	34 (56.7%)	24 (64.9%)	23 (65.7%)	81 (61.4%)
Marital status
Married	28 (47.5%)	12 (34.3%)	3 (8.6%)	43 (33.3%)	.14
Single	1 (1.7%)	3 (8.6%)	9 (25.7%)	13 (10.1%)
Divorced	2 (3.4%)	0 (0.0%)	1 (2.9%)	3 (2.3%)
Widowed	28 (47.5%)	20 (57.1%)	22 (62.9%)	70 (54.3%)
Education level
Illiterate	11 (18.3%)	3 (8.1%)	5 (14.3%)	19 (14.4%)	.30
6 years or less	29 (48.3%)	13 (35.1%)	11 (40.1%)	53 (40.1%)
7–10 years	10 (15.7%)	7 (18.9%)	5 (14.3%)	22 (16.7%)
11–13 years	5 (8.3%)	7 (18.9%)	7 (20.0%)	19 (17.4%)
Started university	1 (1.7%)	0 (0.0%)	0 (0.0%)	1 (0.8%)
University graduate	4 (6.7%)	7 (18.9%)	7 (20.0%)	18 (13.6%)
Working status
Working	4 (6.9%)	4 (10.8%)	1 (2.9%)	9 (7.0%)	.55
Retired	30 (51.7%)	18 (48.7%)	22 (64.7%)	70 (54.3%)
Never worked	24 (41.4%)	15 (40.5%)	11 (32.4%)	50 (38.8%)

P value < .05.

Based on the NINCDS-ADRDA criteria (considered the gold standard in this study), 75 participants (56.8%) were suffering from dementia. There were no statistically significant differences in the distributions of the demographic variables between the demented and nondemented groups (Table 2). Using different cutoff points for the AD8 score (Table 3), the number of demented patients ranged from 85 patients with the cutoff of ≥2 to 65 patients for the cutoff of ≥5. Both the ≥3 and the ≥4 cutoff values gave excellent sensitivity (96% and 95%, respectively) and specificity (97% and 98%, respectively). The area under the receiver operating curve was calculated to be 0.98 (95% confidence interval [CI]: 0.96–1.00), indicating an excellent ability of the Arabic AD8 to discriminate cases from noncases of dementia in this population. The internal consistency reliability for the Arabic version of the AD8, as calculated by Cronbach's α coefficient, was found to be 0.90.

Table 2

Demographic characteristics across cases and controls

Characteristics	Cognitively normal	Demented	Total	P value
Number	57 (43.2%)	75 (56.8%)	132
Age	81.6 (±7.8)	82.0 (±7.8)	81.9 (±7.8)	.78
Gender
Male	20 (35.1%)	31 (41.3%)	51 (38.6%)	.47
Female	37 (64.9%)	44 (58.7%)	81 (61.4%)
Marital status
Married	22 (40.0%)	21 (28.4%)	43 (33.3%)	.14
Single	3 (5.5%)	10 (13.5%)	13 (10.1%)
Divorced	0 (0.0%)	3 (4.1%)	3 (2.3%)
Widowed	30 (54.6%)	40 (54.1%)	70 (54.3%)
Education level
Illiterate	7 (12.3%)	12 (16.0%)	19 (14.4%)	.53
6 years or less	28 (49.1%)	25 (33.3%)	53 (40.1%)
7–10 years	9 (15.8%)	13 (17.3%)	22 (16.7%)
11–13 years	7 (12.3%)	12 (16.0%)	19 (17.4%)
Started university	0 (0.0%)	1 (1.3%)	1 (0.8%)
University graduate	6 (10.5%)	12 (16.0%)	18 (13.6%)
Working status
Working	7 (12.7%)	2 (2.7%)	9 (7.0%)	.08
Retired	29 (52.7%)	41 (55.4%)	70 (54.3%)
Never worked	19 (34.6%)	31 (41.9%)	50 (38.8%)
Interview setting
Outpatients	33 (52.6%)	30 (40.0%)	60 (45.5%)	.12
Inpatients	17 (29.8%)	20 (26.7%)	37 (28.0%)
Nursing homes	10 (17.5%)	25 (33.3%)	35 (26.5%)

Table 3

Sensitivity and specificity for the A-AD8

Cutoff point	Sensitivity, %	Specificity, %	Correctly classified, %
≥0	100.00	0.00	56.82
≥1	98.67	52.63	78.79
≥2	97.33	78.95	89.39
≥3	96.00	96.49	96.21
≥4	94.67	98.25	96.21
≥5	86.67	100.00	92.42
≥6	78.67	100.00	87.88
≥7	65.33	100.00	80.30
≥8	42.67	100.00	67.42
>8	0.00	100.00	43.18

Abbreviation: A-AD8, Arabic eight-item Alzheimer's Dementia.

Despite the excellent psychometric abilities shown for the Arabic AD8, the Arabic AQ revealed superior psychometric and discriminating properties. The Arabic version of the AQ with a cutoff value of ≥8 demonstrated exceptional sensitivity (100%), specificity (100%), and positive predictive value (100%). The ability of the Arabic AQ to discriminate cases from noncases of dementia in this population was highest, with an AUC equal to 1 (95% CI: 0.99–1.00). The Arabic AQ demonstrated high internal consistency with Cronbach's α equal to 0.94.

Per the Arabic Clinical Dementia Rating scale (A-CDR), 66 (51%) participants suffered from dementia, and 23 (18%) were questionable dementia cases. Among the 66 dementia cases, 30 were classified as mild, 20 as moderate, and 16 as severe cases. By using an A-CDR score of ≥0.5, the sensitivity of the A-CDR in relation to the diagnostic criteria (gold standard) was 99% for detecting both questionable and dementia cases. The specificity was 71%. For an A-CDR score of ≥1, the sensitivity drops to 87% for detection of dementia cases; however, the specificity increases to 96%. Although the A-CDR demonstrated excellent discriminating ability with an AUC of 0.97 (95% CI: 0.94–0.99), the sum of its boxes (A-CDR-SOB) was even better with an AUC equal to 0.98 (95% CI: 0.97–1). The A-CDR-SOB demonstrated excellent sensitivity (97%), specificity (93%), and positive predictive value (95%) at the cutoff value of ≥2 and was superior to the regular CDR score in detecting dementia cases among the study sample. The internal consistency for the A-CDR scale was excellent (alpha = 0.98).

As for the different items of the NPI, they were significantly associated with the presence of dementia, except for euphoria, sleep, and appetite. The most common neuropsychiatric abnormality recorded for demented patients was apathy (51%), followed by depression (45%) and anxiety (45%), and the least was euphoria (3%) (Table 4). The internal consistency between the different items of the NPI was estimated to be alpha = 0.76.

Table 4

Neuropsychiatric inventory across cases and controls

Characteristics	Cognitively normal (%)	Demented (%)	P value
Delusion	0 (0.0)	20 (26.7)	.000∗
Hallucination	0 (0.0)	8 (10.7)	.013∗
Agitation	3 (5.6)	21 (28.0)	.001∗
Depression	11 (20.4)	34 (45.3)	.003∗
Anxiety	11 (20.4)	34 (45.3)	.003∗
Euphoria	0 (0.0)	2 (2.7)	.227
Apathy	5 (9.3)	38 (50.7)	.000∗
Disinhibition	0 (0.0)	12 (16)	.002∗
Irritability	5 (9.26)	22 (29.3)	.006∗
Motor behavior	0 (0.0)	11 (14.7)	.003∗
Sleep	17 (31.5)	33 (44.0)	.150
Appetite	12 (22.6)	28 (37.3)	.077

P value < .05.

The Katz Index of Independence in Activities of Daily Living scale was significantly associated with the diagnosis of dementia and reported lower for demented participants when compared to nondemented participants (5.7 and 3.0, respectively, P value < .001). The average Katz scores decreased with the increasing severity of dementia reported by the CDR and demonstrated a strong negative correlation with CDR total score (r value −0.8, P value <.001). Similarly, Katz scores were negatively associated with AD8 scores (r value −0.7, P value <.001), and positively with MMSE scores (r value 0.7, P value <.001) (Table 5).

Table 5

Neuropsychiatric abnormalities across dementia patients

Characteristics	Questionable CDR 0.5 (%)	Mild CDR 1 (%)	Moderate CDR 2 (%)	Sever CDR 3	P value
N	23	30	20	16	-
Delusion	1 (4.4)	4 (13.8)	10 (50.0)	4 (25.0)	.000∗
Hallucination	0 (0.0)	3 (10.0)	2 (10.0)	3 (18.8)	.050
Agitation	2 (8.7)	12 (40.0)	6 (30.0)	1 (6.3)	.010∗
Depression	6 (26.1)	14 (46.7)	11 (55.0)	6 (37.5)	.017∗
Anxiety	4 (17.4)	15 (50.0)	9 (45.0)	6 (37.5)	.072
Euphoria	0 (0.0)	1 (3.3)	1 (5.0)	0 (0.0)	.377
Indifference	3 (13.0)	13 (43.3)	14 (70.0)	9 (56.3)	.000∗
Disinhibition	0 (0.0)	6 (20.0)	6 (30.0)	0 (0.0)	.001∗
Irritability	2 (8.7)	10 (33.3)	8 (40.0)	2 (12.5)	.113
Motor behavior	1 (4.4)	3 (10.0)	5 (25.0)	1 (6.3)	.021∗
Sleep	7 (30.4)	12 (40.0)	10 (50.0)	9 (56.3)	.236
Appetite	4 (17.4)	9 (30.0)	8 (40.0)	9 (56.3)	.030∗

Abbreviation: CDR, Clinical Dementia Rating.

P value < .05.

Discussion

The diagnosis of AD is usually based on the NINCDS-ADRDA. According to these guidelines, the diagnosis is classified as definite (clinical diagnosis with histologic confirmation), probable (typical clinical syndrome without histologic confirmation), or possible (atypical clinical features but no alternative diagnosis apparent; no histologic confirmation). The currently accepted criteria support a probabilistic diagnosis of AD within a clinical context where there is no definitive diagnostic biomarker. A definite diagnosis of AD is only made according to the NINCDS-ADRDA criteria when there is a histopathologic confirmation of the clinical diagnosis [22]. Autopsy results support the “probable” clinical diagnosis in 86% to 90% of cases [23]. Well-trained interviewers can accurately diagnose 90% of patients suffering from dementia. However, most health workers do not have the appropriate training or the time to conduct a comprehensive interview based on the NINCDS-ADRDA. Based on that, user-friendly scales such as the AD8 and AQ are essential. In addition, NPI is essential for the psychopathology in demented patients. Based on that, we undertook this study to translate and validate the aforementioned scales into Arabic. We also translated and validated the CDR to stimulate more Arabic-speaking researchers into the field of dementia.

Based on our findings, just like the original English version, the Arabic AD8 has an excellent ability to discriminate cases from noncases of dementia. The original version has a sensitivity of >84% and a specificity of >80% for a cutoff value of ≥2, whereas the Arabic version has a sensitivity of 96% and a specificity of 97% for a cutoff value of ≥3.

In addition, AQ has a sensitivity of 99% and a specificity of 96%. In the Arabic version, both sensitivity and specificity were at 100%. This may be explained by the fact that our sample had probably more advanced dementia compared to the original English sample.

By demonstrating that the A-CDR has excellent sensitivity, specificity, and AUC, researchers may be able to use this tool after training and certification on the original English language. A-CDR showed similar characteristics as the original version. In addition, A-CDR-SOB was superior to the regular A-CDR score in detecting dementia cases among the study sample. Owing to the increased range of values, the A-CDR-SOB score offers several advantages over the global score, including increased utility in tracking changes within and between stages of dementia severity [24].

As for the Arabic version of the NPI, the translation was easily understood and proved to be useful in identifying patients with behavioral disturbances. However, the lack of gold standard for behaviors, such as apathy, irritability, delusions, and others, made the criterion validation of the NPI in this study not possible. Therefore, a group of experts were asked to participate in a panel for content validation of the NPI.

Finally, the translated Katz Index of Independence in Activities of Daily Living was again easily understood and as expected had a very strong correlation with dementia severity (based on the MMSE, CDR, and AD8).

In conclusion, Arabic-speaking health workers and researchers now have validated tools helping them in the study and the diagnosis of patients suffering from dementia. Some limitations however should be considered. First, the final sample size was smaller than the initially planned sample size. This is due to the fact that it was very difficult to find people not suffering from dementia residing in assisted living arrangements, which lead to a smaller control group. In addition, the lack of gold standard for behavioral disturbance made the interpretation of the NPI more difficult.

Systematic review: In this study, we aimed to examine the validity and internal consistency of Arabic versions of the AD8 (A-AD8), Alzheimer Questionnaire (A-AQ), and Clinical Dementia Rating (CDR) scales, and to assess the Arabic version of Katz ADL, and Neuropsychiatric Inventory.

Interpretation: All the translated scales into Arabic had a strong validity and internal consistency. These validated versions can be relied on similar to the to the original English versions.

Future directions: Now that validated scales are available in Arabic, this will facilitate the diagnosis and related symptoms of dementia among Arabic speaking doctors and patients.