Validation of the Korean Quick Dementia Rating System (K-QDRS)

Abstract

Background:

The Quick Dementia Rating System (QDRS) is a brief and rapid dementia staging tool that does not require a trained rater.

Objective:

The purpose of this study is to demonstrate the validity, reliability, and diagnostic usefulness of the Korean version of the QDRS (K-QDRS).

Methods:

We collected a total of 411 subject-informant dyads including cognitively unimpaired (CU, n = 22), mild cognitive impairment (MCI, n = 198), and dementia (n = 191). The Clinical Dementia Rating (CDR) scale, Korean version of the Mini-Mental State Examination (K-MMSE), Korean version of instrumental activity of daily living (K-IADL), Short Form of the Geriatric Depression Scale, Caregiver-Administered Neuropsychiatric Inventory (CGA-NPI), and detailed neuropsychological tests were administered as gold standards of dementia staging, cognition, function, mood, and behavior.

Results:

Internal consistency of the K-QDRS was excellent with Cronbach’s alpha of 0.933. Concurrent validity was also satisfactory, with the K-QDRS correlating highly with the CDR Sum of Boxes (Pearson’s r = 0.791), K-MMSE (Pearson’s r = –0.518), K-IADL (Pearson’s r = 0.727), and CGA-NPI (Pearson’s r = 0.700). The K-QDRS was highly correlated with the global CDR, K-IADL, and CGA-NPI. We suggested two types of comparisons (for initial diagnosis and for follow-up evaluation). The cutoff scores for follow-up were 1.0 for MCI, 3.5 for very mild dementia, 6.5 for mild dementia, and 11.0 for moderate dementia.

Conclusion:

The K-QDRS is a valid and reliable dementia rating questionnaire and can be used, briefly and rapidly, in various settings like clinical practices, longitudinal cohort studies, and community primary care.

Keywords

Activities of daily living clinical dementia rating cognition dementia neurocognitive disorders screening

INTRODUCTION

The world’s population is aging at an accelerating rate in part due to the development of advanced medical technologies, pharmacotherapeutics and medical services; however, as the population ages, so does the risk of dementia increase [1]. In addition to the need for methods of early detection, brief and valid methods of tracking dementia progression are essential for monitoring stage of dementia and determining appropriate therapeutic interventions and referrals for services. Therefore, in selecting an evaluation tool for long-term follow-up, observation in diverse clinical settings, accuracy and ease of administration, and scoring become equally important. A comprehensive dementia evaluation essentially includes assessments of cognition, function, behavior and personality changes, and mood [2]. The evaluation of cognition includes the assessment of multiple domains such as attention and concentration, memory, language, visuospatial ability, and executive function [3].

One of the most widely used dementia evaluation tools in research and specialty centers is the Clinical Dementia Rating (CDR) [4]. The CDR is a semi-structured interview with patients and their caregivers by an experienced clinician to determine the presence of dementia, and if present, determine the stage of dementia. It takes at least 30 min for accurate implementation, even by experienced clinicians or raters, making it less practical for use in primary care and non-specialty settings. The CDR has 6 domains (e.g., memory, orientation, judgment, and problem-solving) that expertly captures cognitive decline associated with Alzheimer’s disease [4]. However, the CDR does not include domains that may be important for non-Alzheimer dementias such as language, behavior, mood, or attention [5]. Language is included in the Diagnostic and Statistical Manual of Mental Disorders as a major cognitive domain [2]. Mood disturbance or behavioral problems and personality changes should also be assessed for dementia evaluation, independently of the CDR. Another commonly used tool for staging the degree of dementia is the Global Deterioration Scale (GDS) [6]. The GDS has been validated in Korea [7] and is widely used along with the CDR as a global measure for dementia staging. The GDS consists of a 7-point scale and provides specific examples for each stage but may place greater focus on more advanced stages of dementia. However, to be used as a brief staging tool, the major limitations are that the CDR and GDS both require a trained rater and time, thereby limiting their widespread use.

There are several brief screening tools for dementia screening. Currently, the most widely used screening tool for dementia screening worldwide is the Mini-Mental State Examination (MMSE) [8]. The MMSE includes items on various cognitive domains such as orientation, calculation, word recall, language, and copy, with a total score of 30 points. The MMSE takes about 10 min to complete. Montreal Cognitive Assessment [9] is also widely used as a screening test to evaluate a wider range of cognitive domains than the MMSE. It is a 30-point scale as same as the MMSE and takes about 15 min, requiring more time than the MMSE. Mini-Cog [10] is a 5-point scale and consists of a 3-item recall (3 points) and a clock drawing (2 points). It is designed to be used in non-specialist settings in a very short time of less than 3 min. These screening tools generally have the advantage of being able to screen dementia within a short time. However, the examiner must directly test the patient, functional decline is not considered, and dementia staging is not a direct objective. Among the screening tools in the form of a questionnaire, the AD8 [11] is representative. In Korea, the Korean version of the AD8 (K-AD8) [12] is widely used. The AD8 consists of 8 short questions, and for each question, informants or patients are asked to choose one of “YES, A change”, “No, No change”, or “N/A, Don’t know”. The score of AD8 is calculated by the number of items that select “YES, A change”. In Korea, along with the K-AD8, the Korean Dementia Screening Questionnaire-Cognition (KDSQ-C) [13] and the Subjective Memory Complaints Questionnaire (SMCQ) [14] are used as screening tools in the form of an informant questionnaire. The KDSQ-C consists of 5 items related to memory, 5 items related to other cognitive functions including language ability, and 5 items related to complex cognitive functions performing complex activities. For each question, the severity of symptoms is evaluated on a 3-point scale, with a total score of 30 points. The SMCQ consists of a total of 14 items, including 4 items measuring global memory and 10 items measuring everyday memory. In a study comparing dementia screening ability for the K-AD8, the KDSQ-C, and the SMCQ, all of them showed good discriminability for dementia [15]. In particular, the K-AD8 showed the highest value of area under the curve (AUC) analysis in both the questionnaires reported by informants and patients (0.93 for informants, 0.8 for patients respectively). However, although these questionnaire-type screening tools are useful for dementia detection, they are not designed for dementia staging. Therefore, the need for a screening tool that can perform dementia detection and staging at the same time is highly demanded.

Galvin [16] has recognized the demands for a dementia staging tool with easy and brief characteristics and developed and validated the Quick Dementia Rating System (QDRS). The QDRS is a 10-item informant questionnaire and covers areas essential for dementia assessment, while it takes a brief period of time (3 to 5 min) to complete. The first 6 questions of the QDRS correspond to the 6 domains of the CDR and the last 4 questions tap into additional cognitive and behavioral domains (language, personality and behavior, mood, and attention). Each item in the QDRS has five choices assigned 0, 0.5, 1, 2, and 3 points providing a range of scores from 0–30 with higher scores signifying greater impairment. As the scoring structure of the QDRS and CDR are similar, a global CDR and its sum of boxes (CDR-SB) can be derived from the informant QDRS (QDRS-derived CDR global and QDRS-derived CDR-SB score) in a brief fashion and is highly correlated with the full global CDR and CDR-SB. The QDRS also revealed excellent data quality. There were few missing responses (ranging from 0%to 1.7%), which is a concern when conducting informant questionnaires. There was also no significant difference in the QDRS score depending on whether an informant was a spouse or not. The QDRS performed equally well regardless of dementia etiology including Alzheimer’s disease dementia (ADD) [17], vascular dementia (VaD) [18], dementia with Lewy bodies (DLB) [19], frontotemporal degeneration (FTD) [20, 21], and Parkinson’s disease dementia (PDD) [22]. As the CDR global and CDR-SB used as the gold standard for staging dementia increased, the QDRS total score also gradually increased. QDRS cutoff scores were 1.5, 6.0, 12.5, and 17.5 points respectively corresponding to CDR 0 versus CDR 0.5, CDR 0.5 versus CDR 1, CDR 1 versus CDR 2, and CDR 2 versus CDR 3 with high diagnostic accuracy. More recently, a patient version of the QDRS has also been developed [23].

The QDRS has also been validated in other countries. A validation study with the QDRS was conducted in Russia and included 2,357 subjects [24]. In the study, the MMSE was used as the gold standard, and the usefulness of the QDRS was confirmed in the screening of mild and severe dementia. In Iran, a validation study of the QDRS was performed discriminating PD-MCI and PDD among PD [25]. On the other hand, in a longitudinal cohort study, the CDR was performed only when the score of the CDR score derived from the QDRS was greater than 0. It has been reported that implementation of the CDR can be reduced by about 80%by using the QDRS for detection of CDR 0 [26].

The simplicity of implementation, psychometric properties, ability to generate QDRS-derived CDR scores, and validity in other languages are potentially of high utility in clinical settings that are used by unskilled raters or require rapid screening or in research settings for establishing inclusion/exclusion criteria. This purpose of this study is to translate and validate the Korean version of the QDRS (K-QDRS) and test its psychometric properties. We will then determine the optimal cutoff scores that can differentiate cognitively unimpaired (CU), MCI, and dementia in a sample of Korean older adults and examine the correspondence of the K-QDRS-derived CDR global and CDR-SB score to gold standard measurements.

MATERIALS AND METHODS

Study participants

Participants (n = 489) were selected patient-informant pairs from the outpatient clinics at the Department of Neurology in the Konkuk University Medical Center from May 2017 to September 2019. Among the K-QDRS submitted, 16%of forms (n = 78) contained missing items that prevented computation of the K-QDRS scores and were excluded from further analysis. The final sample size is 411 cases. The clinical assessment for dementia diagnosis was made by two neurologists, while four neuropsychologists administered detailed neuropsychological assessments.

The CU group was defined as normal cognitive performance (Z score -1 or higher) adjusted for age and education according to detailed standardized tests [27] and was without psychiatric disorders or medical conditions that could influence cognitive function. For clinical diagnosis of MCI, the Petersen criteria [28] were used, and the MCI group included both amnestic and non-amnestic forms. The clinical diagnosis for ADD, VaD, DLB, FTD, PDD, and other dementia [2] was based on published clinical criteria. In this study, CDR 0 corresponds to CU, CDR 0.5 to MCI or very mild dementia, CDR 1 to mild dementia, and CDR 2 to moderate dementia. No CDR 3 (severe dementia) cases were included.

One neuropsychologist, one language pathologist, and two neurologists translated the QDRS into Korean (Supplementary Table 1). The back-translation was performed by a bilingual Korean-American to confirm that the meanings of QDRS items were well conveyed, and reviewed by the original creator of the QDRS. Finally, a preliminary test was conducted on 10 informants, and instructions were locally adapted to increase comprehension and administration.

Identical to the original QDRS (Supplementary Table 2), the K-QDRS consists of 10 items, each having five possible answers to choose only one among them. Scores assigned to the five answer options are given increasing score values (0, 0.5, 1, 2, and 3) indicating greater severity. The possible range of K-QDRS total score is from 0 to 30. The first six items of the K-QDRS correspond to the six domains of the CDR. Therefore, scores corresponding to CDR global and CDR-SB can be derived through the K-QDRS. In this study, when the K-QDRS is implemented, three scores are generated: a K-QDRS total score, a K-QDRS-derived CDR global score, and a K-QDRS-derived CDR-SB score.

The Korean version of the Mini-Mental State Examination (K-MMSE) [29] was administered for the assessment of general cognition. The CDR, CDR-SB, and GDS were performed for dementia severity evaluation. The Korean version of the Instrumental Activity of Daily Living (K-IADL) [30] was used for ADL assessment. The Short form of the Geriatric Depression Scale (SGDS) [31] and the Caregiver-Administered Neuropsychiatric Inventory (CGA-NPI) [32] were used to assess mood and abnormal behavior, respectively. The Seoul Neuropsychological Screening Battery, second edition [27] was administered for comprehensive cognitive test. Specifically, the Digit Span Test (Forward and Backward), Seoul Verbal Learning Test (SVLT), RCFT copy, Boston Naming Test-15 item, Digit Symbol Coding), Animal fluency, phonemic fluency, Stroop test (Word Reading and Color Reading), and Trail Making Test (TMT) (A and B) were administrated for detailed cognitive evaluation.

Procedures

During detailed neuropsychological assessment of patients, the K-QDRS was provided along with other questionnaires to be filled out by informants. It was completed by informants without assistance. The K-QDRS was conducted prior to the CDR interview with informants. Neither the neuropsychologist conducting the evaluation nor the neurologist doing the clinical assessment reviewed the contents of the submitted K-QDRS until all evaluations were completed and scored. The relationship of an informant to a patient was classified into spouse, adult child, and others. Since the study is data-gathering research, the review exemption was approved by the Konkuk University Medical Center Institutional Review Board (IRB No. KUMC 2021-01-031).

Statistical analysis

To examine group differences, one-way analysis of variance (ANOVA) was used for age, gender, and education data. Chi-square analysis was used for categorical data. Analysis of covariance (ANCOVA) was performed on continuous data, controlling age as a covariate. In case the results of the ANOVA and the ANCOVA were significant, post-hoc comparisons were conducted using Bonferroni correction. To assess data completeness, missing or multiple selection ratios for each K-QDRS item and for each CDR score were calculated. To examine data features of the K-QDRS, the score range, mean, median, standard deviation, floor percentage, and ceiling percentage were calculated on the total K-QDRS and two K-QDRS-derived CDR scores. Internal consistency was assessed using Cronbach’s alpha coefficient. Internal correlations between three K-QDRS scores were calculated using Spearman correlation coefficient. For construct validity, exploratory factor analysis on K-QDRS items was examined using principal axis factoring method with Kaiser’s criterion. For concurrent validity, correlations of the K-QDRS with the CDR, CDR-SB, K-MMSE, K-IADL, SGDS, and CGA-NPI were assessed using Pearson correlation coefficient. For convergent validity, Spearman correlations between K-QDRS items and neuropsychological tests were examined. Additionally, correlations between the K-QDRS and the CDR-SB by characteristics of informants were compared using the Fisher’s z method [33]. For interscale reliability, intraclass correlation coefficients (ICC) between K-QDRS items and corresponding CDR domains were assessed. To assess the discrimination ability of the K-QDRS, means of the K-QDRS were compared for CDR stages. To examine group differences among CU, MCI, and dementia, ANCOVA was used controlling age as a covariate. Also, values of partial eta squared were calculated for effect size. Finally, to discriminate dementia stages through K-QDRS scores, receiver operating characteristic analysis was performed, and optimal staging cutoff scores were presented. The data were analyzed with IBM SPSS v25 statistical software.

RESULTS

Sample characteristics

The average age of 411 participants was 74.91 years, 66.9%were female, the average educational level was 9.06 years, the average K-MMSE score was 22.95, and the average CDR-SB score was 3.36. Sample characteristics are presented in detail in Table 1. There were significant differences between groups in age among age, gender, and education level. The group difference was significant in the CGA-NPI, K-IADL, K-MMSE, CDR-SB, and GDS except the SGDS when age was controlled as a covariate (all ps < 0.001). In the CGA-NPI, the score increased as the CDR score increased. Follow-up post-hoc results showed significant differences among the CDR 0.5 or less group, the CDR 1 group, and the CDR 2 group. K-IADL scores also increased with increasing CDR scores, indicating that functional decline appeared as cognitive function decreased. Follow-up post-hoc analyses showed significant differences among non-dementia group and each dementia group. The K-MMSE score decreased significantly as cognitive function worsened. The CDR-SB score showed significant differences among all groups.

Table 1

Demographic characteristics of the participants

Variables	CDR 0 CU^aN = 22	CDR 0.5 MCI^bN = 198	CDR 0.5 very mild dementia^cN = 67	CDR 1 mild dementia^dN = 105	CDR 2 moderate dementia^eN = 19	F or χ²	p	Post-hoc (Bonferroni)
Age (y)	68.36±10.44	72.30±8.70	77.75±6.69	79.37±8.25	75.11±11.12	F = 16.96	< 0.001	a,b < c,d
Gender (female), No. (%)	19 (86.36)	135 (68.18)	40 (59.70)	68 (64.76)	13 (68.42)	χ² = 5.72	0.221
Education (y)	10.14±5.83	9.27±5.03	8.43±5.54	8.84±5.41	9.21±5.83	F = 0.59	0.672
SGDS	3.18±3.84	4.65±4.58	5.00±4.04	5.23±4.36	4.16±4.88	F = 1.66	0.158
CGA-NPI	1.73±3.23	6.11±8.18	10.82±9.72	19.96±15.70	32.63±27.22	F = 38.25	< 0.001	a < c; a,b,c < d < e
K-IADL	0.03±0.05	0.21±0.21	0.86±1.19	1.50±0.56	2.37±0.32	F = 114.59	< 0.001	a,b < c < d < e
K-MMSE	27.45±3.63	25.55±3.03	21.55±3.20	19.42±3.91	15.16±4.73	F = 79.23	< 0.001	a,b > c > d > e
CDR-SB	0.07±0.18	1.52±0.67	3.58±0.55	6.02±1.27	10.76±1.36	F = 804.88	< 0.001	a < b < c < d < e
GDS	1.95±0.21	3.00±0.18	3.63±0.49	4.35±0.48	5.21±042	F = 437.11	< 0.001	a < b < c < d < e

Values are presented as mean±standard deviation and controlled age for variables except gender and education. CDR, Clinical Dementia Rating; CU, cognitively unimpaired; MCI, mild cognitive impairment; SGDS, Short Form of the Geriatric Depression Scale; CGA-NPI, Caregiver-Administered Neuropsychiatric Inventory; K-IADL, Korean-Instrumental Activities of Daily Living; K-MMSE, Korean Mini-Mental State Examination; CDR-SB, Clinical Dementia Rating-sum of boxes; GDS, Global Deterioration Scale.

Data quality of the K-QDRS

Participant scores for the K-QDRS-derived CDR global and K-QDRS-derived CDR-SB showed the full range of possible scores. But in the K-QDRS total, no participant received the ceiling score of 30 points (Supplementary Table 3). This is likely because a severe dementia group corresponding to CDR 3 was not included in this study.

The percentage of average scores for each possible total of three K-QDRS scores ranged from 17.5%(K-QDRS total) to 20.3%(K-QDRS-derived CDR global). The floor effects of three K-QDRS scores ranged from 11.7%(K-QDRS total) to 24.1%(K-QDRS-derived CDR global). The ceiling effects of three K-QDRS scores ranged from 0.0%(K-QDRS total) to 1.5%(K-QDRS-derived CDR global) (Supplementary Table 3). The diagnostic groups of dementia diagnosed subjects were ADD (N = 139), VaD (N = 21), DLB (N = 9), FTD (N = 4), PDD (N = 5), and others (N = 13). There was a mean difference between diagnostic groups only in gender (p = 0.028) among age, gender, and education level. There was no difference between diagnostic groups in three K-QDRS scores (the K-QDRS total, K-QDRS-derived CDR global, and K-QDRS-derived CDR-SB) with gender as a covariate. In K-QDRS items, there was a significant difference between diagnostic groups in two items: “language and communication” (F[5, 184] = 2.794, p = 0.013) and “attention and concentration” (F[5, 184] = 2.408, p = 0.029). However, there was no difference between groups in follow-up post-hoc with Bonferroni correction for these two items.

We examined whether there were differences in correlations between the K-QDRS and CDR-SB according to the type of informant. Spouse group (Pearson’s r = 0.828) showed a higher correlation than adult child group (Pearson’s r = 0.770). However, there was no significant difference in Pearson’s r values between comparison groups (Spouse versus Adult child, Cohabiting versus Non-cohabiting, and Cohabiting adult child versus Non-cohabiting adult child) (Table 2).

Table 2

Comparison of correlations between the K-QDRS and the CDR-SB

Comparison	Group 1		Group 2		Fisher’s z method
	Subject (N)	Pearson’s r	Subject (N)	Pearson’s r	p
Spouse versus Adult child	Spouse (119)	0.828	Adult child (272)	0.770	0.147
Cohabiting versus Non-cohabiting	Cohabiting (217)	0.794	Non-cohabiting (194)	0.797	0.936
Cohabiting adult child versus	Cohabiting adult	0.731	Non-cohabiting	0.792	0.259
Non-cohabiting adult child	child (96)		adult child (176)

K-QDRS, Korean Quick Dementia Rating System; CDR-SB, Clinical Dementia Rating-sum of boxes.

An additional analysis was performed on the proportion of responses without additional guidance on the questionnaire in the entire sample. Among the total 489 patients, 78 (16.0%) had one or more missing items or multiple selection errors for multiple options. Missing response rates ranged from 1.4%(“memory and recall”) to 5.3%(“activities outside the home”) while multiple selection rates ranged from 1.2%(“activities outside the home”) to 3.5%(“memory and recall”). Total missing or multiple selection rates were relatively high, ranging from 3.5%(“toileting and personal hygiene”) to 6.3%(“activities outside the home”). The proportion of missing or multiple selections increased with increasing CDR scores (0.0%for CDR 0, 12.4%for CDR 0.5, 21.1%for CDR 1, and 29.6%for CDR 2).

Reliability of the K-QDRS

Internal consistency was excellent, as Cronbach’s alphas for the K-QDRS total and two K-QDRS-derived CDR scores were 0.933 and 0.920 respectively. Correlations among the K-QDRS total, K-QDRS-derived CDR global, and K-QDRS-derived CDR-SB were high (Table 3). The first six items of the K-QDRS were compared with the corresponding CDR domains with ICC. The ICCs were moderate, ranging from 0.499 (“memory and recall”) to 0.695 (“toileting and personal hygiene”). The ICC for the K-QDRS-derived CDR-SB (ICC = 0.770) was higher than that for the K-QDRS-derived CDR global (ICC = 0.605) (Table 4).

Table 3

Internal consistency and internal correlations

K -QDRS	Number of Items	Cronbach alpha (95%CI)	Spearman r
			Total	Derived CDR global	Derived CDR-SB
Total	10	0.933 (0.923–0.943)	1
Derived CDR global	6	0.920 (0.907–0.931)	0.899	1
Derived CDR-SB	6	0.920 (0.907–0.931)	0.974	0.915	1

K-QDRS, Korean Quick Dementia Rating System, CDR, Clinical Dementia Rating, CDR-SB, Clinical Dementia Rating-sum of boxes.

Table 4

Interscale reliability between the K-QDRS and the CDR

K -QDRS item or derived CDR global and CDR-SB	CDR domain	ICC (95%CI)	p
Memory and recall	Memory	0.499 (0.423–0.568)	< 0.001
Orientation	Orientation	0.600 (0.534–0.658)	< 0.001
Decision making and problem-solving	Judgment &Problem Solving	0.528 (0.454–0.594)	< 0.001
Activities outside the home	Community Affairs	0.675 (0.619–0.724)	< 0.001
Function at home and hobby activities	Home and Hobbies	0.645 (0.584–0.698)	< 0.001
Toileting and personal hygiene	Personal Care	0.695 (0.642–0.742)	< 0.001
K-QDRS-derived CDR global	CDR global	0.605 (0.540–0.663)	< 0.001
K-QDRS-derived CDR-SB	CDR-SB	0.770 (0.727–0.806)	< 0.001

K-QDRS, Korean Quick Dementia Rating System; CDR, Clinical Dementia Rating; CDR-SB, Clinical Dementia Rating-sum of boxes; ICC, intraclass correlation coefficient.

Validity of the K-QDRS

The Kaiser-Meyer-Olkin’s measure of sampling adequacy was 0.946 and the Bartlett’s test of sphericity was significant (p < 0.001), indicating that factor analysis might be useful with K-QDRS data. The exploratory factor analysis for the K-QDRS retrieved one factor with an eigenvalue of 6.40 (the second largest was 0.72) from 10 items, accounting for 63.98%of the total variance of the K-QDRS. Factor loadings ranged from 0.681 (“mood”) to 0.871 (“decision making and problem-solving”). Correlations between the K-QDRS item scores and the K-QDRS total score ranged from 0.684 (“mood”) to 0.870 (“decision making and problem-solving”) (Supplementary Table 4). Correlations between the K-QDRS total score and the CDR global (r = 0.685, p < 0.001), CDR-SB (r = 0.791, p < 0.001), K-MMSE (r = 0.518, p < 0.001), K-IADL (r = 0.727, p < 0.001), and CGA-NPI (r = 0.700, p < 0.001) were significant, revealing that the K-QDRS is concurrently valid (Table 5).

Table 5

Correlation of the K-QDRS and the CDR, CDR-SB, K-MMSE, K-IADL, SGDS, and CGA-NPI

K -QDRS	CDR	CDR-SB	K-MMSE	K-IADL	SGDS	CGA-NPI
Total	0.685	0.791	–0.518	0.727	0.042	0.700
Derived CDR global	0.655	0.748	–0.501	0.689	–0.005	0.580
Derived CDR-SB	0.692	0.801	–0.536	0.750	0.006	0.634

Values are the Pearson’s correlation coefficients. All ps < 0.001. K-QDRS, Korean Quick Dementia Rating System; CDR, Clinical Dementia Rating; CDR-SB, Clinical Dementia Rating-sum of boxes; K-MMSE, Korean Mini-Mental State Examination; K-IADL, Korean-Instrumental Activities of Daily Living; SGDS, Short Form of the Geriatric Depression Scale; CGA-NPI, Caregiver-Administered Neuropsychiatric Inventory.

Spearman correlations between the K-QDRS items and neuropsychological tests were examined (Table 6). If the correlation was 0.4 or more, it was judged as convergent [11]. All items were highly correlated with measures reflecting the overall functional state such as the K-IADL, CDR, CDR-SB, and GDS (ranging from 0.778 to 0.457). The K-MMSE showed a high correlation with all items (item 1 to 5) except “toileting and personal hygiene” (item 6) among K-QDRS-derived CDR items. “Memory and recall” (item 1) showed a high correlation with the SVLT delayed recall that represents verbal memory. “Attention and concentration” (item 10) was highly correlated with the color reading of the Stroop test and the TMT-B that are related to executive functions. “Orientation” (item 2), “decision making and problem-solving” (item 3), “activities outside the home” (item 4), and “function at home and hobby activities” (item 5) were highly correlated with various tests measuring memory and executive functions. On the other hand, “toileting and personal hygiene” (item 6), “behavior and personality changes” (item 7), “language and communication” (item 8), and “mood” (item 9) did not show a high correlation with detailed cognitive tests.

Table 6

Construct validity of the K-QDRS

K -QDRS items	Tests (Spearman r > 0.4) with descending order
1. Memory and recall	K-IADL (0.712), CDR-SB (0.692), GDS (0.627), CDR (0.590), CGA-NPI (0.502), SVLT-DR (–0.438), K-MMSE (–0.433)
2. Orientation	K-IADL (0.729), CDR-SB (0.714), GDS (0.658), CDR (0.611), K-MMSE (–0.544), CGA-NPI (0.472), SVLT-DR (–0.469), TMT-B (0.460), Stroop-CR (–0.445), SVLT-IR (–0.406)
3. Decision making and problem-solving	K-IADL (0.765), CDR-SB (0.714), GDS (0.669), CDR (0.622), CGA-NPI (0.529), K-MMSE (–0.512), TMT-B (0.484), Stroop-CR (–0.426), DSC (–0.420), TMT-A (0.407), SVLT-IR (–0.405), Stroop-WR (–0.401)
4. Activities outside the home	K-IADL (0.800), CDR-SB (0.750), GDS (0.716), CDR (0.700), CGA-NPI (0.561), K-MMSE (–0.492), SVLT-IR (–0.491), TMT-B (0.475), TMT-A (–0.455), Stroop-WR (–0.453), DSC (–0.450), SVLT-DR (–0.443), Stroop-CR (–0.438), COWAT (–0.402)
5. Function at home and hobby activities	K-IADL (0.778), CDR-SB (0.314), GDS (0.666), CDR (0.661), CGA-NPI (0.562), TMT-B (0.507), K-MMSE (–0.477), DSC (–0.468), Stroop-CR (–0.464), COWAT (–0.452), Stroop-WR (–0.451), TMT-A (0.446), Animal (–0.437)
6. Toileting and personal hygiene	CDR-SB (0.638), K-IADL (0.630), CDR (0.614), GDS (0.599), CGA-NPI (0.491)
7. Behavior and personality changes	CGA-NPI (0.647), CDR-SB (0.614), K-IADL (0.610), GDS (0.564), CDR (0.546)
8. Language and communication	K-IADL (0.575), CDR-SB (0.535), GDS (0.535), CDR (0.509), CGA-NPI (0.434)
9. Mood	CGA-NPI (0.636), K-IADL (0.488), CDR-SB (0.471), GDS (0.460), CDR (0.457)
10. Attention and concentration	CDR-SB (0.629), K-IADL (0.613), CGA-NPI (0.561), GDS (0.548), CDR (0.540), TMT-B (0.411), Stroop-CR (–0.407)

All ps < 0.001. K-QDRS, Korean Quick Dementia Rating System; K-IADL, Korean-Instrumental Activities of Daily Living; CDR-SB, Clinical Dementia Rating-sum of boxes; GDS, Global Deterioration Scale; CDR, Clinical Dementia Rating; CGA-NPI, Caregiver-Administered Neuropsychiatric Inventory; SVLT-DR, Seoul Verbal Learning Test delayed recall; K-MMSE, Korean Mini-Mental State Examination; TMT-B, Trail Making Test-B; Stroop-CR, Stroop Color and Word Test 60 seconds-Color Reading; SVLT-IR, Seoul Verbal Learning Test-Immediate Recall; DSC, Digit Symbol Coding; TMT-A, Trail Making Test-A; Stroop-WR, Stroop Color and Word Test 60 seconds-Word Reading; COWAT, Controlled Oral Word Association Test; Animal, Animal fluency.

The ANCOVA with age as a covariate revealed a significant difference among CDR groups on the K-QDRS three scores (all ps < 0.001). Follow-up post-hoc analysis with Bonferroni correction showed that the mean scores of CDR groups were all different significantly on the K-QDRS three scores (CDR 0 < CDR 0.5 < CDR 1 < CDR 2). Group differences among CU, MCI, and dementia with respect to K-QDRS three scores were all significant. Follow-up post-hoc analysis with Bonferroni correction showed that CU and MCI groups were significantly different from the dementia group (Table 7).

Table 7

Discrimination ability of the K-QDRS by diagnostic stages

Variables	CU ^aN = 22	MCI ^bN = 198	Very mild dementia ^cN = 67	Mild dementia^dN = 105	Moderate dementia^eN = 19	F	p	Partial eta	Post-hoc (Bonferroni)
K-QDRS Total	0.32±0.52	2.31±2.17	5.16±2.77	10.20±4.92	14.29±6.80	114.77	< 0.001	0.531	a,b < c < d < e
K-QDRS-derived CDR global	0.02±0.11	0.32±0.27	0.60±0.28	1.09±0.60	1.66±0.90	81.67	< 0.001	0.446	a < b < c < d < e
K-QDRS-derived CDR-SB	0.11±0.26	1.36±1.38	3.37±2.02	6.92±3.35	9.71±4.52	123.14	< 0.001	0.549	a,b < c < d < e

Values are presented as mean±standard deviation and controlled age for variables. K-QDRS, Korean Quick Dementia Rating System; CU, cognitively unimpaired; MCI, mild cognitive impairment; CDR, Clinical Dementia Rating; CDR-SB, Clinical Dementia Rating-sum of boxes.

Staging cutoff scores of the K-QDRS

We examined the discrimination ability of the K-QDRS to distinguish different diagnostic groups. For the purpose of initial diagnosis, the discrimination ability was at an excellent level, with AUC values ranging from 0.829 to 0.927. In this case, the cutoff scores were 1.0 for MCI, 3.0 for very mild dementia, 5.0 for mild dementia, and 7.0 for moderate dementia. For the purpose of follow-up evaluation, the discrimination ability was moderate or good, with the AUCs ranging from 0.680 to 0.839. In this case, the cutoff scores were 1.0 for MCI, 3.5 for very mild dementia, 6.5 for mild dementia, and 11.0 for moderate dementia. The sensitivity and specificity of the cutoff points are presented in Table 8.

Table 8

Optimal staging cutoff scores for the K-QDRS

Comparison		Optimal cutoff score	AUC (95%CI)	Sensitivity (%)	Specificity (%)
Contrast group	Detection group
For initial diagnosis
CU	MCI	1.0	0.839 (0.776–0.903)	0.722	0.818
CU + MCI	Very mild dementia	3.0	0.829 (0.779–0.882)	0.791	0.691
CU + MCI + Very mild dementia	Mild dementia	5.0	0.927 (0.901–0.953)	0.876	0.801
CU + MCI + Very mild dementia + Mild dementia	Moderate dementia	7.0	0.896 (0.848–0.944)	0.947	0.750
For follow-up evaluation
CU	MCI	1.0	0.839 (0.776–0.903)	0.722	0.818
MCI	Very mild dementia	3.5	0.812 (0.755–0.869)	0.746	0.742
Very mild dementia	Mild dementia	6.5	0.819 (0.757–0.881)	0.790	0.672
Mild dementia	Moderate dementia	11.0	0.680 (0.549–0.811)	0.632	0.610

All ps < 0.001. K-QDRS, Korean Quick Dementia Rating System; AUC, Area of under the curve; CU, cognitively unimpaired; MCI, mild cognitive impairment; CDR, Clinical Dementia Rating; CDR-SB, Clinical Dementia Rating-sum of boxes; Very mild dementia, dementia with CDR 0.5; Mild dementia, dementia with CDR 1; Moderate dementia, dementia with CDR 2.

DISCUSSION

The Korean version of the QDRS showed an excellent internal consistency with Cronbach’s alpha of 0.933, reflecting the same underlying construct. Our results revealed good concurrent validity with the Pearson’s correlation coefficient of 0.791 related to the CDR-SB. Regardless of the informant type (Spouse or Adult child; Cohabiting adult child or Non-Cohabiting adult child), the K-QDRS showed a high correlation with the CDR-SB. The K-QDRS-derived CDR global and K-QDRS-derived CDR-SB showed a high correlation with the original CDR and CDR-SB. In addition, the correlation between K-QDRS items and corresponding CDR domains was also excellent. As a result of checking the correlation between the individual items of the K-QDRS and detailed tests, all items were significantly related to tests representing global functions related to cognition, daily activity function, and behavior, such as the CDR, CDR-SB, GDS, K-IADL, and CGA-NPI. This indicates that all items of the K-QDRS represent the contents important for dementia detection and staging. Item 1, 2, 3, 4, 5, and 10 were highly correlated with detailed cognitive tests indicating a strong association with memory and executive function. In particular, “memory and recall” (item 1) showed a high correlation only with delayed recall representing memory. Therefore, item 1 was thought to be an item that directly reflects memory impairments. The CGA-NPI, which indicates the degree of behavioral and psychological symptoms of dementia, showed the highest correlations with “behavior and personality changes” (item 7) and “mood” (item 9), which are the most relevant in terms of content, among K-QDRS items. Item 10 (“attention and concentration”) was most related to executive functions among detailed cognitive tests. When the K-QDRS was applied to various dementia etiology groups, differences between groups were confirmed in two items of “language and communication” (F [5, 184] = 2.794, p = 0.013) and “attention and concentration” (F [5, 184] = 2.408, p = 0.029) among K-QDRS items. Post-hoc analysis did not show any significant difference, but VaD and FTD groups tended to respond more to these two items, although due to the small sample sizes, these conclusions should be interpreted with caution. Although the largest subgroup in the cohort was ADD and the numbers of other dementia etiologies are small, our data suggest that the K-QDRS could be applied to ADD and other dementia etiologies (e.g., VaD, DLB, FTD, and PDD). The K-QDRS was found to be a useful screening tool for dementia staging, as it showed an excellent discriminative ability for the purpose of initial diagnosis and a moderate-to-good discriminative ability for the purpose of follow-up. These results support that the K-QDRS is a valid and reliable dementia rating questionnaire, regardless of dementia etiology.

The original QDRS reported there was no significant difference between spouse and non-spouse informants in QDRS scoring [16]. Similarly, in this study, the correlation between the K-QDRS and the CDR-SB was high in both spouse informants (p = 0.828) and adult child informants (p = 0.770). Reporting reliability of the K-QDRS was acceptable regardless of the type of informant or cohabitation with the patient. The correlation between K-QDRS and CDR-SB was highest in the spouse group. Also, the Pearson correlation coefficients of the spouse group and the cohabiting adult child group were compared statistically through the Fisher’s z method and showed a value (p = 0.072) close to the significance level of 0.05. This suggests that spouse reports are likely to be more reliable than cohabiting adult child [34]. Therefore, it seems advantageous for spouses to participate as informants, if possible.

In the original QDRS [16], two factors (cognitive and behavioral) were extracted, whereas in a recent patient version of the QDRS [23], one factor was extracted. In our study, the result of the factor analysis showed a good construct validity as all factor loadings were greater than 0.681 with one factor extracted. This indicates that changes in various areas related to dementia, such as cognition, behavior, and mood, are being explained as one factor, the dementia severity. Therefore, the cognitive and behavioral factors from the original QDRS were not further analyzed in this study. Regarding individual items, our analyses showed that the accuracy of informants’ rating of “memory and recall” may be the most challenging. The agreement between the K-QDRS “memory and recall” (item 1) and the CDR memory domain was statistically significant (ICC 0.499, p < 0.001) but showed relatively low compared to other items (ICC 0.528 ∼0.695). This suggests that it is more difficult for informants to accurately grasp subjects’ state with respect to memory than other cognitive areas, mood, and behavioral aspects.

One of the main purposes of this study is to provide cutoff scores that can discriminate diagnostic groups (CU, MCI, very mild dementia, mild dementia, and moderate dementia) through the K-QDRS. The selection of a detection group and a contrast group may differ depending on the purpose of use. If the disease has progressive stages, cutoff scores can be provided by comparing two adjacent groups [35]. On the other hand, several types of possible comparisons could be presented. For example, when detecting dementia, possible comparisons (e.g., CU versus dementia, MCI versus dementia, and CU/MCI versus dementia) can be provided. When various comparisons are provided, there is an advantage that one can select one suitable for the setting to which the tool is applied. On the other hand, it can be difficult to determine what is desirable to apply. Therefore, in this study, two types of comparisons were presented, one for initial diagnosis and one for follow-up evaluation. In the comparison for initial diagnosis, the entire groups below a specific detection group were set as a contrast group. By using a method of minimizing the detection group and maximizing the contrast group, cutoff scores are relatively lower than those of other methods, and this was intended to enable the early detection of disease. In this study, the suggested cutoff scores of the K-QDRS for initial diagnosis were 1.0 for MCI, 3.0 for very mild dementia, 5.0 for mild dementia, and 7.0 for moderate dementia. The AUC values ranged from 0.829 to 0.927. This shows that the K-QDRS is an excellent tool in discriminating cognitive impairment groups for initial diagnosis [36]. The cutoff scores for the purpose of follow-up were obtained through the comparison of two adjacent groups. The proposed cutoff scores for follow-up were 1.0 for MCI, 3.5 for very mild dementia, 6.5 for mild dementia, and 11.0 for moderate dementia. The AUC values for follow-up ranged from a minimum of 0.680 (moderate dementia) and a maximum of 0.839 (MCI). It shows that the discriminative ability for the moderate dementia group is rather low, while the discriminative abilities for the other groups are excellent. In the original QDRS, CDR groups were analyzed for staging, whereas, in this study, diagnostic groups were analyzed, so direct comparison of cutoff scores is difficult. For example, the CDR 0.5 group of the original QDRS was divided into two groups in this study: MCI and very mild dementia groups. The cutoff score for the CDR 0.5 group in the original QDRS was 1.5 and that for the MCI group in the K-QDRS was 1.0. Therefore, it would be judged that at least the onset cutoff score of cognitive and functional impairment is presented almost similarly. Comparing the cutoff scores for initial and follow-up evaluation of the K-QDRS, the cutoff score applied to MCI is the same as 1.0. However, for very mild dementia (3.0 versus 3.5), mild dementia (5.0 versus 6.5), and moderate dementia (7.0 versus 11.0), it can be seen that the cutoff score for initial diagnosis was lower than that for follow-up in each diagnosis group. This shows the relationship between cutoff scores for follow-up and those for initial diagnosis. In conclusion, our results show that the K-QDRS can be applied as a useful tool for dementia staging and can be estimated with a considerable degree of accuracy.

In the validation of the K-QDRS, some limitations should be mentioned. In this study, there was no severe dementia group, so cutoff scores were not provided for this group. This study does not provide inter-rater reliability by two different caregivers and no test-retest reliability. Since the number of subjects according to dementia types other than ADD is small, the characteristics of the K-QDRS that can appear in various types of dementia have not been sufficiently reviewed. This should be considered to be a part that requires further research.

In general, when conducting a questionnaire, it is common to guide the respondent how to fill out and go through the process of checking the contents after filling out the questionnaire. However, in this study, by intentionally omitting these processes, we tried to find out how accurately the questionnaire was answered. In the case of the original QDRS, it was reported that there were no multiple selections, and the missing items were very low, ranging from 0%to 1.7%. On the other hand, in this study, the missing or multiple selections were relatively high, ranging from 3.5%to 6.3%. This was the case in 78 participants, 16.0%of the total 489 participants. The proportion of missing or multiple selections increased as CDR scores increased. There was no error data in the group with CDR 0, while it was very high at 29.6%for CDR 2. As can be seen from the results, it seems necessary to guide for filling out before implementation and check the contents after that in order to increase the accuracy of responses. With another approach, if the questionnaire is conducted through a device such as a tablet rather than paper, it will be possible to fundamentally prevent missing or multiple selection errors.

In the CDR evaluation, experienced clinicians evaluate a patient’s stage through various questions for each domain. The QDRS has a fundamental limitation in that it evaluates a patient’s stage with a single question for each domain. In addition, if the caregivers themselves have depression or anxiety due to patient care, the reliability of the evaluation may be lowered [37]. However, one of the main characteristics of informant screening questionnaires is to check the condition of the patient from the informant in a short time [38]. Moreover, in each sentence of the QDRS, various examples representing the key characteristics of each stage are presented. For example, item 4 refers to various aspects of outside activities, such as vocational activities, purchases, community activities, religious activities, volunteering, and social group activities. Therefore, a single sentence for each stage looks like one question in a formal way, but it contains several exemplars in its content. In this way, an informant is designed to select a stage that is considered to be closest to a patient’s current condition. In this study, the informant’s ability to distinguish stages was not separately evaluated. However, by confirming the high correlation between the K-QDRS and the CDR, the reliability of informant ratings was evaluated.

Through this study, we confirmed that the K-QDRS is suitable for use as a screening tool for dementia staging. Assessment of dementia staging is essential for dementia management and follow-up care. The K-QDRS, like the original QDRS, was not designed to place the CDR but provide a brief and valid dementia detection and staging instrument. In settings, where a gold standard such as the CDR cannot be completed, the K-QDRS will provide clinically actionable information. In other settings, it could be used as a dementia follow-up and monitoring tool to complement the CDR and provide additional domains not covered in the CDR. If the K-QDRS is performed first and then the CDR is evaluated in a way that further specifies the contents of the K-QDRS, the CDR could be evaluated more easily. Also, if the patient version of the K-QDRS is validated in the future, the usefulness of the K-QDRS will increase further. The patient version of the K-QDRS could be applied more broadly to subjects for whom no informant is available and the subjective cognitive decline group for early detection and follow-up.

In this study, we confirmed that the K-QDRS is a valid and reliable screening tool that can be useful in the detection of cognitive impairment and staging of dementia. The biggest strength of the K-QDRS is that it enables dementia screening within a short period of time and the ease of implementation that does not require well-skilled raters. Moreover, since the K-QDRS shows diagnostic power similar to that of the CDR, it will be useful in various scenes that require dementia screening such as primary care institutions or large-scale research.

DISCLOSURE STATEMENT

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/21-0584r2).

Footnotes

The supplementary material is available in the electronic version of this article: .

References

Prince

, Ali

, Guerchet

, Prina

, Albanese

, Wu

(2016) Recent global trends in the prevalence and incidence of dementia, and survival with dementia. Alzheimers Res Ther 8, 23.

Association.

(2013) Diagnostic and statistical manual of mental disorders (5th ed.), Author, Arlington, VA.

American Psychological A (2012) Guidelines for the evaluation of dementia and age-related cognitive change. Am Psychol 67, 1–9.

Morris

(1993) The Clinical Dementia Rating (CDR): Current version and scoring rules. Neurology 43, 2412–2414.

Knopman

, Kramer

, Boeve

, Caselli

, Graff-Radford

, Mendez

, Miller

, Mercaldo

(2008) Development of methodology for conducting clinical trials in frontotemporal lobar degeneration. Brain 131, 2957–2968.

Reisberg

, Ferris

, de Leon

, Crook

(1982) The Global Deterioration Scale for assessment of primary degenerative dementia. Am J Psychiatry 139, 1136–1139.

Choi

, Na

, Lee

, Hahm

, Jeong

, Koo

, Ha

, Ahn

, Korean Dementia Research Group (2002) The validation of the Korean version of Global Deterioration Scale. J Korean Neurol Assoc 20, 612–617.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. Apractical method for grading the cognitive state of patients for theclinician. J Psychiatr Res 12, 189–198.

Lam

, Middleton

, Masellis

, Stuss

, Harry

, Kiss

, Black

(2013) Criterion and convergent validity of the Montreal cognitive assessment with screening and standardized neuropsychological testing. J Am Geriatr Soc 61, 2181–2185.

10.

Borson

, Scanlan

, Brush

, Vitaliano

, Dokmak

(2000) The mini-cog: A cognitive ‘vital signs’ measure for dementia screening in multi-lingual elderly. Int J Geriatr Psychiatry 15, 1021–1027.

11.

Galvin

, Roe

, Powlishta

, Coats

, Muich

, Grant

, Miller

, Storandt

, Morris

(2005) The AD8: A brief informant interview to detect dementia. Neurology 65, 559–564.

12.

Ryu

, Kim

, Han

(2009) Validity and reliability of the Korean version of the AD8 informant interview (K-AD8) in dementia. Alzheimer Dis Assoc Disord 23, 371–376.

13.

Yang

, Cho

, Chey

, Kim

(2002) The development and validation of Korean Dementia Screening Questionnaire (KDSQ). J Korean Neurol Assoc 20, 1–8.

14.

Youn

, Kim

, Lee

, Jhoo

, Lee

, Park

, Choi

, Choe

, Jeong

, Choo

, Woo

(2009) Development of the Subjective Memory Complaints Questionnaire. Dement Geriatr Cogn Disord 27, 310–317.

15.

Kim

, Kim

, Park

, Youn

, Lee

, Jeong

, Choi

, Han

, Kim

, Na

, Park

, Yim

, Yang

(2019) A comparative evaluation of the KDSQ-C, AD8, and SMCQ as a cognitive screening test to be used in national medical check-ups in Korea. J Korean Med Sci 34, e111.

16.

Galvin

(2015) The Quick Dementia Rating System (QDRS): A rapid dementia staging tool. Alzheimers Dement (Amst) 1, 249–259.

17.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

Jr. , Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

18.

Sachdev

, Kalaria

, O’Brien

, Skoog

, Alladi

, Black

, Blacker

, Blazer

, Chen

, Chui

, Ganguli

, Jellinger

, Jeste

, Pasquier

, Paulsen

, Prins

, Rockwood

, Roman

, Scheltens

, Internationlal Society for Vascular Behavioral and Cognitive Disorders (2014) Diagnostic criteria for vascular cognitive disorders: A VASCOG statement. Alzheimer Dis Assoc Disord 28, 206–218.

19.

McKeith

, Boeve

, Dickson

, Halliday

, Taylor

, Weintraub

, Aarsland

, Galvin

, Attems

, Ballard

, Bayston

, Beach

, Blanc

, Bohnen

, Bonanni

, Bras

, Brundin

, Burn

, Chen-Plotkin

, Duda

, El-Agnaf

, Feldman

, Ferman

, Ffytche

, Fujishiro

, Galasko

, Goldman

, Gomperts

, Graff-Radford

, Honig

, Iranzo

, Kantarci

, Kaufer

, Kukull

, Lee

VMY

, Leverenz

, Lewis

, Lippa

, Lunde

, Masellis

, Masliah

, McLean

, Mollenhauer

, Montine

, Moreno

, Mori

, Murray

, O’Brien

, Orimo

, Postuma

, Ramaswamy

, Ross

, Salmon

, Singleton

, Taylor

, Thomas

, Tiraboschi

, Toledo

, Trojanowski

, Tsuang

, Walker

, Yamada

, Kosaka

(2017) Diagnosis and management of dementia with Lewy bodies: Fourth consensus report of the DLB Consortium. Neurology 89, 88–100.

20.

Gorno-Tempini

, Hillis

, Weintraub

, Kertesz

, Mendez

, Cappa

, Ogar

, Rohrer

, Black

, Boeve

, Manes

, Dronkers

, Vandenberghe

, Rascovsky

, Patterson

, Miller

, Knopman

, Hodges

, Mesulam

, Grossman

(2011) Classification of primary progressive aphasia and its variants. Neurology 76, 1006–1014.

21.

Rascovsky

, Hodges

, Knopman

, Mendez

, Kramer

, Neuhaus

, van Swieten

, Seelaar

, Dopper

, Onyike

, Hillis

, Josephs

, Boeve

, Kertesz

, Seeley

, Rankin

, Johnson

, Gorno-Tempini

, Rosen

, Prioleau-Latham

, Lee

, Kipps

, Lillo

, Piguet

, Rohrer

, Rossor

, Warren

, Fox

, Galasko

, Salmon

, Black

, Mesulam

, Weintraub

, Dickerson

, Diehl-Schmid

, Pasquier

, Deramecourt

, Lebert

, Pijnenburg

, Chow

, Manes

, Grafman

, Cappa

, Freedman

, Grossman

, Miller

(2011) Sensitivity of revised diagnostic criteria for the behavioural variant of frontotemporal dementia. Brain 134, 2456–2477.

22.

Emre

, Aarsland

, Brown

, Burn

, Duyckaerts

, Mizuno

, Broe

, Cummings

, Dickson

, Gauthier

, Goldman

, Goetz

, Korczyn

, Lees

, Levy

, Litvan

, McKeith

, Olanow

, Poewe

, Quinn

, Sampaio

, Tolosa

, Dubois

(2007) Clinical diagnostic criteria for dementia associated with Parkinson’s disease. Mov Disord 22, 1689–1707; quiz 1837.

23.

Galvin

, Tolea

, Chrisphonte

(2020) Using a patient-reported outcome to improve detection of cognitive impairment and dementia: The patient version of the Quick Dementia Rating System (QDRS). . PLoS One 15, e0240422.

24.

Levin

, Vasenina

, Sonin

, Tayupova

, Zatsepina

, Ostapenko

(2018) The feasibility of using Quick Dementia Rating System in clinical practice: Results of a multi-center prospective validation study. Zh Nevrol Psikhiatr Im S S Korsakova 118, 40–47.

25.

Mahmoudi Asl

, Mehdizadeh

, Raeesi Roudbari

, Mehdizadeh

, Habibi

, Niazi Khatoon

, Taghizadeh

(2020) Staging of cognitive impairment in Parkinson’s disease: Validity of Quick Dementia Rating System. Disabil Rehabil, doi: 10.1080/09638288.2020.1755728.

26.

Berman

, Koscik

, Clark

, Mueller

, Bluder

, Galvin

, Johnson

(2017) Use of the Quick Dementia Rating System (QDRS) as an initial screening measure in a longitudinal cohort at risk for Alzheimer’s disease. J Alzheimers Dis Rep 1, 9–13.

27.

Kang

, Jhang

, Na

(2012) Seoul Neuropsychological Screening Battery. Second Edition (SNSB-II): Professional Manual. Human Brain Research and Consulting, Incheon.

28.

Petersen

, Doody

, Kurz

, Mohs

, Morris

, Rabins

, Ritchie

, Rossor

, Thal

, Winblad

(2001) Current concepts in mild cognitive impairment. Arch Neurol 58, 1985–1992.

29.

Kang

(2006) A normative study of the Korean-Mini Mental State Examination (K-MMSE) in the elderly. Korean J Psychol Gen 25, 1–12.

30.

Kang

, Choi

, Lee

, Kwon

, Na

, Han

(2002) The reliability and validity of the Korean Instrumental Activities of Daily Living (K-IADL). J Korean Neurol Assoc 20, 8–14.

31.

Bae

, Cho

(2004) Development of the Korean version of the Geriatric Depression Scale and its short form among elderly psychiatric patients. J Psychosom Res 57, 297–305.

32.

Kang

, Choi

, Lee

, Jeong

, Hahm

, Han

, Cummings

, Na

(2004) Caregiver-Administered Neuropsychiatric Inventory (CGA-NPI). J Geriatr Psychiatry Neurol 17, 32–35.

33.

Fisher

(1915) Frequency distribution of the values of the correlation coefficient in samples of an indefinitely large population. Biometrika 10, 507–521.

34.

McLoughlin

, Cooney

, Holmes

, Levy

(1996) Carer informants for dementia sufferers: Carer awareness of cognitive impairment in an elderly community-resident sample. Age Ageing 25, 367–371.

35.

O’Bryant

, Waring

, Cullum

, Hall

, Lacritz

, Massman

, Lupo

, Reisch

, Doody

, Texas Alzheimer’s Research Consortium (2008) Staging dementia using Clinical Dementia Rating Scale Sum of Boxes scores: A Texas Alzheimer’s research consortium study. Arch Neurol 65, 1091–1095.

36.

Hosmer

DW LS

(2000) Applied Logistic Regression, 2nd Ed. Chapter 5, John Wiley and Sons, New York, NY.

37.

Jorm

(2004) The Informant Questionnaire on cognitive decline in the elderly (IQCODE): A review. Int Psychogeriatr 16, 275–293.

38.

Taylor-Rowan

, Nafisi

, Patel

, Burton

, Quinn

(2020) Informant-based screening tools for diagnosis of dementia, an overview of systematic reviews of test accuracy studies protocol. Syst Rev 9, 271.