Abstract
Background:
To reduce the diagnostic gap of dementia, 3 strategies can be employed for case-finding of cognitive impairment in ambulatory-care-settings, namely using informant-report, brief-cognitive-test, or combination of informant-report and brief-cognitive-test. The right strategy to adopt–across different healthcare settings–remains unclear. This diagnostic study compared the performance of the 3 strategies for detecting dementia (primary aim), as well as for detecting both mild-cognitive-impairment(MCI) and dementia (secondary aim).
Methods:
Participants ≥65 years (n=11,057) were recruited from Alzheimer’s-Disease-Centers across USA. Participants provided data on an informant-report (Functional-Activities-Questionnaire, FAQ), a brief-cognitive-test (4-item short-variant of Montreal-Cognitive-Assessment, MoCA), and a combined measure with informant-report and brief-cognitive-test (sum of FAQ and MoCA short-variant). They also received standardized assessments (clinical history, physical examination and neuropsychological testing) to diagnose MCI and dementia. Areas-under-the-receiver-operating-characteristic-curve(AUC) of the 3 strategies were compared using the DeLong method, with AUC>90% indicating excellent performance.
Results:
All 3 strategies had excellent performance in detecting dementia, although informant-report (AUC=95.9%, 95%CI=95.4–96.3%) was significantly better than brief-cognitive-test (AUC=93.0%, 95%CI=92.4–93.6%), and combined measure had the best performance (AUC=97.0%, 95%CI=96.7–97.4%). However, to detect both MCI and dementia, only the combined measure had excellent performance (AUC=93.0%, 95%CI=92.5–93.4%), while stand-alone informant-report or brief-cognitive-test performed sub-optimally (AUC<90%). Performance of the 3 strategies was not affected by participants’ age, educational attainment, or underlying prevalence of MCI and dementia.
Conclusions:
For case-finding of dementia in ambulatory-care-settings, informant-reports would suffice as first-line measures, and brief-cognitive-tests may optionally be added-on–in services with available resources–to further improve the accuracy of detection. For case-finding of both MCI and dementia, combination of informant-reports and brief-cognitive-tests remain the most appropriate strategy.
Keywords: Dementia, case finding, diagnostic gap, Functional Activities Questionnaire, Montreal Cognitive Assessment
INTRODUCTION
Dementia is consistently underdiagnosed around the world, with a recent meta-analysis – based on 23 studies from various countries – showing that 61.7% of older persons with dementia never received any formal diagnosis of the condition [1]. Without formal diagnosis of dementia, the older persons may not have access to the critical elements of dementia management – such as the prescription of cognitive enhancers [2–6], management of behavioral issues [7–11], and support for their caregivers [7–11] – which can impact on their well-being [8, 9] and put them at high-risk of premature nursing home placement [2–7, 10, 11]. In the presence of cognitive impairment, they may also have difficulty collaborating with healthcare professionals in managing their chronic diseases and adhering to medical advice on medications and lifestyle changes. Yet with the cognitive impairment unbeknownst to their physicians, they may not receive the necessary support that are relevant to individuals with cognitive impairment, such as medication reminders, case management, and home visits. Consequently, they can be at higher risk of inappropriate healthcare utilization related to missed appointments, emergency room visits, hospitalizations, and longer hospital stay [12, 13]. To address the issue of undiagnosed dementia, various international bodies (including the Alzheimer’s Disease International [14], the International Association of Gerontology and Geriatrics [15] and the Gerontological Society of America [16]) have emphasized the importance of active case-finding among high-risk groups in ambulatory care settings. Notably, most international guidelines have contrasted active case-finding from universal screening (that intends to detect dementia in all patients regardless of their risk profile), and highlighted the relevance of active case-finding specifically among patients who are at high-risk of dementia. In particular, the International Association of Gerontology and Geriatrics recommends active case-finding among the older patients in current healthcare services, given that age is the strongest risk factor for dementia [15].
Conventionally, 3 strategies can be employed in the case-finding of dementia, that is, using informant reports alone (such as AD8 and Functional Activities Questionnaire) [17, 18], brief cognitive tests alone (such as Mini-Cog and short variants of Montreal Cognitive Assessment [MoCA]) [19–22] or combined measures that pair informant reports with brief cognitive tests [23–29]. Each case-finding strategy has inherent strengths and limitations, with the literature remaining unclear on the right strategy to adopt across different healthcare services with varying clinical resources and patient profiles. Although informant reports can be more convenient to administer in routine practice (requiring only self-administration by patients or family members), they are subjected to measurement errors related to informant’s recall and desirability biases. In contrast, brief cognitive tests are less prone to informant-related biases and can provide more accurate descriptions of current cognitive function. However, brief cognitive tests are more time-consuming, require the availability of trained personnel to administer, and can be prohibitive in services with high patient load and limited resources for test administration. Combined measures that pair informant reports with brief cognitive tests have been suggested in the literature to balance the limitations of stand-alone informant reports or brief cognitive tests [15], yet evidence on combined measures has been conflicting across different healthcare settings, with some studies demonstrating their better performance [23–26] while others reporting the contrary [27–29].
This study sought to compare the 3 strategies of case-finding (informant report, brief cognitive test, or combined informant report and brief cognitive test); and using a large sample from USA, aimed to provide more conclusive evidence on the right strategy to adopt across different healthcare services. As the study was based on the dataset from National Alzheimer’s Coordinating Center (NACC), it utilized the informant report and brief cognitive test that were available in the dataset – namely the Functional Activities Questionnaire (FAQ) [18] and a short-variant of MoCA [30] – and compared the performance in dementia detection across the following 3 strategies:
FAQ alone;
MoCA short-variant alone;
Summed scores of FAQ and MoCA short-variant.
As a secondary aim, this study also examined the performance of the 3 strategies in detecting cognitive impairment (that is, capturing both mild cognitive impairment and dementia), considering that some healthcare services may also be interested to capture mild cognitive impairment for clinical trials [31] and preventive interventions (such as physical exercise, cognitive training, and management of cardiovascular risk factors) [32–36]. Additionally, as different healthcare settings often differ by patient characteristics (particularly, age and educational attainment) as well as underlying prevalence of cognitive impairment, the study also examined whether these factors (age, educational attainment, and underlying prevalence) would modify the performance of the 3 strategies and consequently, affect the choice of case-finding strategy across different healthcare settings.
METHOD
Participants and procedures
This cross-sectional diagnostic-study involved participants recruited from approximately 39 Alzheimer’s Disease Centers across USA (as available in NACC dataset between March 2015 and August 2019) [37]. It was developed in accordance with the Standards for Reporting of Diagnostic Accuracy Studies in dementia and cognitive impairment (STARDdem) reporting guideline [38]. The study included those who fulfilled the following criteria: (1) age ≥65 years; (2) provided data on FAQ; and (3) provided data on MoCA. Of note, although only those with complete data on FAQ and MoCA were included in the primary analysis of this study, multiple imputation was conducted in subsequent sensitivity analyses to examine the robustness of the results to missing data. The Alzheimer’s Disease Centers obtained informed consent from participants, as well as received approval by their local institutional review boards.
Participants were accompanied by informants who knew them well (usually their spouse, child or friend) and took part in standardized assessments which included clinical history, physical examination and detailed neuropsychological testing (involving the domains of immediate memory, visuospatial abilities, delayed memory, language, attention, processing speed and executive function) [31, 39]. The clinical diagnoses were made based on the standardized assessments, with majority (83.6%) made via consensus conference and remainder made by single clinicians. Mild cognitive impairment (MCI) was diagnosed using modified Petersen criteria [40]; while dementia was diagnosed with McKhann (2011) criteria [41], with further classification into the primary etiologies of Alzheimer’s dementia [41], vascular dementia [42], dementia with Lewy Bodies [43, 44], frontotemporal lobar degeneration [45–48], or other etiologies.
Measures
FAQ is an informant measure of the instrumental activities of daily living that can be affected in people with dementia (such as managing personal finances, shopping for groceries, and traveling in the neighborhood) [18]. Its 10 items were rated by informants on 4-point Likert scales, and could be summed [18] to generate a maximum score of 30. FAQ was used in this study as it was the only informant measure that was available in the NACC dataset. To align FAQ score in the same direction with that of MoCA short-variant, FAQ was reverse-coded in this study with higher scores indicating better functional abilities.
MoCA [49] is a widely-used cognitive test that assesses seven cognitive domains: Visuospatial/Executive, Naming, Attention, Language, Abstraction, Delayed recall and Orientation. It has a maximum score of 30, with higher scores corresponding to better cognition. Although MoCA is widely-regarded as a useful cognitive test, it often requires 10–15 minutes to administer [20] and may not be well-suited for case-finding efforts where large number of patients are involved. Consequently, various short-variants of MoCA have been developed in the literature to reduce its administration time. In a recent comparative study [22], the short-variant by Cecato (2016) [30] has been suggested to have better performance than the other short-variants of MoCA in distinguishing dementia from non-dementia, and hence were selected as the exemplar of brief cognitive test in this study. The short-variant by Cecato (2016) comprises 4 items (Clock-drawing, Naming–rhinoceros, Delayed-recall and Orientation), and has a maximum score of 15 [30]. It was previously developed by identifying MoCA items that are most discriminative of participants with dementia [30]. Of note, although the short-variant by Cecato (2016) was used in the primary analysis of this study, an alternative short-variant of MoCA [21] as well as the full version of the MoCA were also tested in the subsequent sensitivity analyses to examine the robustness of the results to the choice of brief cognitive test.
Statistical analyses
Area-under-the-receiver-operating-characteristic-curve (AUC) was computed and used as the primary metric to examine the performance of the 3 strategies (FAQ, MoCA short-variant, or summed scores of FAQ and MoCA short-variant) in distinguishing dementia from non-dementia, with the analyses further stratified by educational attainment (≤12 years of education; and >12 years of education). AUC estimates the probability of correctly distinguishing dementia from non-dementia when a test is presented with pairs of dementia and non-dementia. It provides a convenient approach to compare the 3 strategies in this study, as AUC incorporates information on test sensitivity and specificity across all possible cut-off scores and circumvents the need to select arbitrary cut-off scores for comparison purposes. The use of AUC can also be particularly relevant in this study as there have not been well-established cut-off scores for FAQ or the combined measure based on FAQ and MoCA short-variant, and hence direct comparisons based on specific cut-off scores may not be possible across the three strategies. In general, AUC>0.9 is deemed as excellent performance [50], and indicates that a strategy has low error-rate (<10%) and is suitable for case-finding purposes. AUC were also compared among the 3 strategies via a non-parametric approach proposed by DeLong et al [20, 21, 31, 51–53], with statistically-significant differences defined as p≤0.05; and clinically-significant differences defined as AUC differences of ≥1% (when a test is used for case-finding in a large population of say one million people, 1% difference in AUC would have translated into an additional 10,000 people being misclassified) [21].
For the secondary aim, the analyses were repeated to examine the performance of the 3 strategies in distinguishing MCI and dementia from normal cognition. Additionally, sensitivity analyses were conducted to assess robustness of the results:
among the younger participants between the age of 65 to 74 years;
among the older participants ≥75 years;
if the prevalence of MCI and dementia was artificially halved (this was done by randomly selecting and including only a subset of participants with MCI and dementia, so that the proportion of MCI and dementia was lowered by half in comparison to that in the present dataset);
if the brief cognitive test was based on an alternative short-variant of MoCA by Liew (2019) [21], instead of the short-variant by Cecato (2016). The MoCA short-variant by Liew (2019) comprises 4 items (Clock-drawing, Tap-at-letter-A, Orientation and Delayed-recall), and has a maximum score of 15 [21]. This short-variant was previously derived from MoCA using a computationally-intensive technique, which selected MoCA items that best distinguished dementia from non-dementia [21]. It was shown to have comparable performance to MoCA in dementia detection [21]. It also demonstrated better performance in dementia detection than the other short variants of MoCA, as well as 3 other commonly-used brief cognitive tests (Memory Impairment Screen [54], Mini-Cog [19] and GPCOG [55]) [21];
if the brief cognitive test was based on the original, full-version of MoCA (instead of the short-variants of MoCA);
If multiple imputation was used to address the missing data on FAQ and MoCA in NACC dataset (instead of excluding participants with such missing data). Details on the conduct of multiple imputation are separately described in Supplementary Material 1.
All analyses were conducted in Stata (version 16).
RESULTS
Flow diagram related to participant selection is presented in Supplementary Material 2. Among the participants who were ≥65 years (n=14,320), 3,263 participants could not be included in the primary analysis due to missing data in FAQ or MoCA, resulting in the final sample size of 11,057 participants. A comparison of demographic information between those with and without missing data is shown in Supplementary Material 3. Those with missing data were more likely to be older, female, and non-White. In addition, those with missing FAQ were more likely to have normal cognition, while those with missing MoCA were more likely to have dementia. In the final sample that was included in the primary analysis (n=11,057), 53.9% had normal cognition, 22.7% had MCI and 23.4% had dementia. Among those with dementia (n=2,589), 70.9% had the primary etiology of Alzheimer’s dementia, 1.3% had vascular dementia, 7.3% mixed Alzheimer’s/vascular dementia, 6.7% dementia with Lewy Bodies, 11.6% frontotemporal lobar degeneration, and 2.2% due to other or unknown etiology. The participant characteristics of the final sample are further shown in Table 1.
Table 1.
Characteristics of the study participants at baseline (n=11,057)
| Variable | Overall sample (n=11,057) | Normal cognition (n=5,958) | MCI (n=2,510) | Dementia (n=2,589) |
|---|---|---|---|---|
| Age, median (IQR) | 75 (70–80) | 74 (69–79) | 76 (71–81) | 76 (71–82) |
| Female sex, n (%) | 6,381 (57.7) | 3,862 (64.8) | 1,262 (50.3) | 1,257 (48.6) |
| Years of education, median (IQR) | 16 (14–18) | 16 (14–18) | 16 (14–18) | 16 (13–18) |
| Ethnicity, n (%) | ||||
| White | 9,007 (81.5) | 4,775 (80.1) | 2,020 (80.5) | 2,212 (85.4) |
| African American | 1,300 (11.8) | 770 (12.9) | 309 (12.3) | 221 (8.5) |
| Others/Unknown | 750 (6.8) | 413 (6.9) | 181 (7.2) | 156 (6.0) |
| FAQ total score, median (IQR)a | 30 (24–30) | 30 (30–30) | 29 (25–30) | 13 (7–20) |
| MoCA total score, median (IQR) | 24 (20–27) | 26 (24–28) | 22 (20–25) | 15 (11–19) |
IQR, interquartile range; MCI, mild cognitive impairment; FAQ, Functional Activities Questionnaire; MoCA, Monteral Cognitive Assessment.
The original FAQ was reverse-scored to align its total score in the same direction with that of MoCA short-variant. Higher scores indicate better functional abilities. As the participants may find some FAQ items inapplicable to them, those who provided data on at least 5 out of the 10 items in FAQ were included in this study, with the total score rescaled to match the maximum score of 30 (the total score was divided by the number of items completed, and multiplied by 10).
All 3 strategies had excellent performance (AUC>90%) in detecting dementia (Table 2). However, informant report was significantly better than brief cognitive test (AUC 95.9% versus 92.7%), while combined measure had the best performance (AUC 97.0%). The results remained consistent across education subgroups. The sensitivity and specificity statistics are presented in Table 3. Informant report had 90.9% sensitivity and 88.6% specificity at the cut-off score of <27; while brief cognitive test had lower sensitivity and specificity (84.5% and 87.3%, respectively) at the cut-off score of <10; and combined measure had the best sensitivity and specificity (91.7% and 90.8%, respectively) at the cut-off score of <36.
Table 2.
Comparing the performance of the 3 case-finding strategies in distinguishing dementia from non-dementia in: (a) overall sample; (b) ≤12 years of education; and (c) >12 years of education.
| Case-finding strategy | AUC, % (95% CI) | P-valuea |
|---|---|---|
| (a) Overall sample (n=11,057) | ||
| Informant report (FAQ)b | 95.9 (95.4–96.3) | Reference |
| Brief cognitive test (MoCA short-variant)c | 92.7 (92.1–93.3) | <0.001 |
| Informant report and Brief cognitive test (FAQ + MoCA short-variant)d | 97.0 (96.6–97.3) | <0.001 |
| (b) ≤12 years of education (n=1,848) | ||
| Informant report (FAQ)b | 95.2 (94.2–96.2) | Reference |
| Brief cognitive test (MoCA short-variant)c | 92.6 (91.2–93.9) | <0.001 |
| Informant report and Brief cognitive test (FAQ + MoCA short-variant)d | 96.6 (95.9–97.3) | <0.001 |
| (c) >12 years of education (n=9,209) | ||
| Informant report (FAQ)b | 95.9 (95.4–96.5) | Reference |
| Brief cognitive test (MoCA short-variant)c | 92.6 (91.9–93.3) | <0.001 |
| Informant report and Brief cognitive test (FAQ + MoCA short-variant)d | 97.0 (96.5–97.4) | <0.001 |
FAQ, Functional Activities Questionnaire; MoCA, Montreal Cognitive Assessment; AUC, area under the receiver operating characteristics curve; CI, confidence interval.
Bonferroni-adjusted p-values, based on comparisons of AUC with that of FAQ (reference).
The original FAQ was reverse-scored to align its total score in the same direction with that of MoCA short-variant.
Based on the MoCA short-variant by Cecato (2016), comprising Clock-drawing, Naming–rhinoceros, Delayed-recall and Orientation.
The sum of FAQ (reverse-scored) and MoCA short-variant.
Table 3.
The sensitivity and specificity of the 3 case-finding strategies, at various cut-off scores, in distinguishing dementia from non-dementia.
| Cut-off | Overall sample a | ≤12 years of education a | >12 years of education a | |||
|---|---|---|---|---|---|---|
| Se, % | Sp, % | Se, % | Sp, % | Se, % | Sp, % | |
| Informant report (FAQ) b | ||||||
| <23 | 82.1 | 95.4 | 82.0 | 94.1 | 82.1 | 95.6 |
| <24 | 83.9 | 94.4 | 83.6 | 92.6 | 84.0 | 94.7 |
| <25 | 86.5 | 92.8 | 85.6 | 90.6 | 86.8 | 93.2 |
| <26 | 88.3 | 91.3 | 87.3 | 88.6 | 88.6 | 91.7 |
| <27 | 90.9 | 88.6 | 90.3 | 84.9 | 91.0 | 89.2 |
| <28 | 93.7 | 85.2 | 93.9 | 80.7 | 93.6 | 85.9 |
| <29 | 95.7 | 81.0 | 96.3 | 75.8 | 95.5 | 81.9 |
| Brief cognitive test (MoCA short-variant) c | ||||||
| <8 | 61.7 | 97.6 | 71.3 | 95.2 | 58.5 | 98.0 |
| <9 | 74.9 | 93.8 | 84.1 | 88.5 | 71.9 | 94.7 |
| <10 | 84.5 | 87.3 | 90.3 | 78.3 | 82.6 | 88.8 |
| <11 | 90.7 | 76.9 | 93.9 | 62.9 | 89.7 | 79.3 |
| <12 | 93.9 | 66.0 | 95.9 | 50.8 | 93.2 | 68.5 |
| Informant report and Brief cognitive test (FAQ + MoCA short-variant) d | ||||||
| <31 | 80.7 | 96.7 | 82.0 | 95.5 | 80.3 | 96.9 |
| <32 | 82.8 | 95.9 | 83.6 | 94.2 | 82.5 | 96.2 |
| <33 | 85.6 | 94.9 | 86.1 | 92.3 | 85.4 | 95.3 |
| <34 | 87.8 | 93.8 | 87.2 | 90.0 | 88.0 | 94.4 |
| <35 | 89.8 | 92.4 | 90.3 | 88.3 | 89.7 | 93.0 |
| <36 | 91.7 | 90.8 | 93.0 | 86.5 | 91.3 | 91.5 |
| <37 | 93.4 | 89.0 | 95.2 | 83.4 | 92.9 | 90.0 |
| <38 | 94.5 | 86.8 | 95.9 | 80.5 | 94.0 | 87.8 |
| <39 | 96.1 | 82.9 | 98.1 | 75.0 | 95.4 | 84.2 |
FAQ, Functional Activities Questionnaire; MoCA, Montreal Cognitive Assessment; Se, sensitivity; Sp, specificity.
The optimal cut-off score is bold-faced and based on a balance between sensitivity and specificity.
The original FAQ was reverse-scored to align its total score in the same direction with that of MoCA short-variant.
Based on the MoCA short-variant by Cecato (2016), comprising Clock-drawing, Naming–rhinoceros, Delayed-recall and Orientation.
The sum of FAQ (reverse-scored) and MoCA short-variant.
In contrast, for the detection of MCI and dementia, stand-alone informant report or brief cognitive test had poorer performance (AUC<90%), and combined measure was the only strategy that had excellent performance (AUC 92.9%) (Table 4). At the cut-off score of <41, combined measure had 85.2% sensitivity and 85.7% specificity in detecting MCI and dementia. Detailed results on sensitivity and specificity statistics are presented in Table 5.
Table 4.
Comparing the performance of the 3 case-finding strategies in distinguishing mild cognitive impairment and dementia from normal cognition in: (a) overall sample; (b) ≤12 years of education; and (c) >12 years of education.
| Case-finding strategy | AUC, % (95% CI) | P-valuea |
|---|---|---|
| (a) Overall sample (n=11,057) | ||
| Informant report (FAQ)b | 87.7 (87.1–88.3) | Reference |
| Brief cognitive test (MoCA short-variant)c | 88.5 (87.9–89.1) | 0.076 |
| Informant report and Brief cognitive test (FAQ + MoCA short-variant)d | 92.9 (92.4–93.4) | <0.001 |
| (b) ≤12 years of education (n=1,848) | ||
| Informant report (FAQ)b | 87.7 (86.3–89.1) | Reference |
| Brief cognitive test (MoCA short-variant)c | 87.6 (86.0–89.1) | 1.00 |
| Informant report and Brief cognitive test (FAQ + MoCA short-variant)d | 91.5 (90.2–92.7) | <0.001 |
| (c) >12 years of education (n=9,209) | ||
| Informant report (FAQ)b | 87.6 (86.9–88.3) | Reference |
| Brief cognitive test (MoCA short-variant)c | 88.4 (87.7–89.1) | 0.105 |
| Informant report and Brief cognitive test (FAQ + MoCA short-variant)d | 93.0 (92.5–93.6) | <0.001 |
FAQ, Functional Activities Questionnaire; MoCA, Montreal Cognitive Assessment; AUC, area under the receiver operating characteristics curve; CI, confidence interval.
Bonferroni-adjusted p-values, based on comparisons of AUC with that of FAQ (reference).
The original FAQ was reverse-scored to align its total score in the same direction with that of MoCA short-variant.
Based on the MoCA short-variant by Cecato (2016), comprising Clock-drawing, Naming–rhinoceros, Delayed-recall and Orientation.
The sum of FAQ (reverse-scored) and MoCA short-variant.
Table 5.
The sensitivity and specificity of the 3 case-finding strategies, at various cut-off scores, in distinguishing mild cognitive impairment and dementia from normal cognition.
| Cut-off | Overall sample a | ≤12 years of education a | >12 years of education a | |||
|---|---|---|---|---|---|---|
| Se, % | Sp, % | Se, % | Sp, % | Se, % | Sp, % | |
| Informant report (FAQ) b | ||||||
| <27 | 62.5 | 97.7 | 66.1 | 95.6 | 61.5 | 98.0 |
| <28 | 67.9 | 96.3 | 71.8 | 94.2 | 66.8 | 96.6 |
| <29 | 73.1 | 94.0 | 76.2 | 90.9 | 72.3 | 94.4 |
| <30 | 80.0 | 89.0 | 81.4 | 86.5 | 79.6 | 89.3 |
| Brief cognitive test (MoCA short-variant) c | ||||||
| <9 | 46.8 | 98.7 | 59.4 | 96.9 | 43.3 | 99.0 |
| <10 | 59.5 | 96.2 | 69.8 | 90.5 | 56.7 | 97.0 |
| <11 | 72.0 | 89.4 | 80.2 | 77.9 | 69.7 | 91.0 |
| <12 | 80.8 | 80.0 | 86.2 | 65.4 | 79.3 | 82.1 |
| <13 | 89.1 | 63.8 | 92.1 | 46.1 | 88.2 | 66.3 |
| Informant report and Brief cognitive test (FAQ + MoCA short-variant) d | ||||||
| <38 | 67.8 | 98.1 | 74.1 | 95.6 | 66.0 | 98.5 |
| <39 | 72.9 | 96.4 | 78.9 | 91.8 | 71.2 | 97.0 |
| <40 | 78.9 | 92.9 | 83.2 | 85.4 | 77.8 | 94.0 |
| <41 | 85.2 | 85.7 | 88.0 | 73.2 | 84.5 | 87.5 |
| <42 | 90.1 | 75.6 | 92.5 | 60.1 | 89.4 | 77.8 |
| <43 | 94.5 | 60.0 | 95.7 | 42.0 | 94.2 | 62.5 |
FAQ, Functional Activities Questionnaire; MoCA, Montreal Cognitive Assessment; Se, sensitivity; Sp, specificity.
The optimal cut-off score is bold-faced and based on a balance between sensitivity and specificity.
The original FAQ was reverse-scored to align its total score in the same direction with that of MoCA short-variant.
Based on the MoCA short-variant by Cecato (2016), comprising Clock-drawing, Naming–rhinoceros, Delayed-recall and Orientation.
The sum of FAQ (reverse-scored) and MoCA short-variant.
The results remained robust in the sensitivity analyses even when the performance of the 3 strategies were re-examined across the various scenarios, namely among younger participants aged 65–74 years (Supplementary Material 4–5), among older participants aged ≥75 years (Supplementary Material 6–7), when the prevalence of MCI and dementia was artificially halved (Supplementary Material 8–9), when an alternative MoCA short-variants was used (Supplementary Material 10–11), when the full-version of MoCA was used (Supplementary Material 12–13), and when the missing data on FAQ and MoCA were accounted for using multiple imputation (Supplementary Material 14–15).
DISCUSSION
This study sought to provide more conclusive evidence on the utility of 3 case-finding strategies (informant report, brief cognitive test, or combined informant report and brief cognitive test) for detecting cognitive impairment in ambulatory care settings. All 3 strategies had excellent performance in detecting dementia (AUC>90%), although informant report was better than brief cognitive test, and combined measure had the best performance. However, for detecting MCI and dementia, only the combined measure had excellent performance while stand-alone informant report or brief cognitive test performed sub-optimally (AUC<90%). Notably, performance of the case-finding strategies was not affected by participants’ age, educational attainment, underlying prevalence of MCI and dementia, the different variants of a brief cognitive test, or the presence of missing data.
The findings on the performance of brief cognitive test are consistent with prior meta-analyses of brief cognitive tests [56, 57], where brief cognitive tests were generally shown to have better performance in detecting dementia than MCI. For example, in a recent meta-analysis focusing on dementia detection [56], the AUC for Addenbrooke’s Cognitive Examination–Revised (ACE-R) was 96%, Mini-Cog 95%, and Mini-Mental State Examination 92%. In contrast, in a separate meta-analysis focusing on MCI detection [57], the AUC for ACE-R was 84%, Consortium to Establish a Registry for Alzheimer’s Disease Battery (CERAD) 86%, MoCA 85%, and Mini-Mental State Examination 76%. Similarly, the findings on the better performance of combined measure are also consistent with some of the available evidence [23–25]. They are understandable, because informant reports and brief cognitive tests respectively capture the two criteria that are critical to the definition of MCI and dementia, namely subjective reports of cognitive decline and objective evidence of cognitive deficits [41, 58]. When used in combination, the two measures provide complementary information on individuals’ cognitive status and hence would have better performance in detecting cognitive impairment. In particular, combined measure offers better performance in capturing MCI (on top of dementia), when stand-alone informant reports or brief cognitive tests would not have been sensitive enough in detecting earlier and more subtle cognitive impairment.
Contrary to the call for blanket use of combined measures to detect dementia (as recommended by some researchers in the field) [15], this study provided more clarity on the appropriate indications to use combined measures and potentially supports a more nuanced approach in case-finding efforts. For the purpose of detecting dementia in ambulatory care settings, a workflow similar to that presented in Figure 1 is prudent (especially among older patients who have reliable informants), where informant reports would have sufficed as first-line measures, given its excellent performance (AUC>90%), its better performance than brief cognitive tests, as well as its ease-of-administration (which can be especially critical in services with high patient load and limited resources). Optionally, in services with available resources, brief cognitive tests may still be added-on (as part of combined measures) to further improve the accuracy of dementia detection. Notwithstanding the above, brief cognitive tests can still be used as first-line measures if no suitable informants are available, given the similarly excellent performance of brief cognitive tests (AUC>90%) as shown in this study. Moreover, if the intention is to detect cognitive impairment (that is, capturing both MCI and dementia), combined measures with both informant reports and brief cognitive tests would remain the most appropriate option as seen in this study.
Figure 1.
A proposed workflow to facilitate the active detection of dementia in ambulatory care settings, among older patients who have reliable informants.
FAQ, Functional Activities Questionnaire; MoCA, Montreal Cognitive Assessment.
aThe cut-off age, as well as the frequency, of case finding depends on the available resources and the risk profile of the patients in the respective healthcare services. b The original FAQ was reverse-scored to align its total score in the same direction with that of MoCA short-variant.
The generalizability of the results may potentially come into question, as there can be genuine doubts on whether findings based on NACC dataset may apply to other healthcare settings that may have different age profiles, lower educational attainment (than the participants in NACC), or much lower prevalence of cognitive impairment (such as in primary care settings). This issue is pertinent, given the suggestions from recent literature that age [19], educational attainment [15, 19, 59, 60], and underlying prevalence of cognitive impairment [61] may impact on the performance of case-finding instruments and consequently, the choice of case-finding strategy. This concern was addressed in the sensitivity analyses, by stratifying the results by age and education subgroups, and by examining the effects of underlying prevalence of cognitive impairment. Of which, the study demonstrated the inconsequential effects of age, educational attainment and underlying prevalence on the choice of case-finding strategy, and suggests the potential applicability of the findings even across different healthcare services with varying patient profiles.
The findings should be considered in the light of the strengths and limitations of this study. The study has the strengths of well-characterized samples (with standardized assessments, and rigorous diagnostic evaluations of cognitive impairment), as well as a large number of participants that were recruited from across USA. At the same time, the study also has several limitations that should be noted. First, the participants in the study involved those who volunteered at the Alzheimer’s Disease Centers, which may have explained the higher prevalence of dementia (23.4%) compared to those commonly reported in community-settings. Taking this limitation into consideration, the findings from this study is probably less generalizable to patients in community settings, and mostly applicable to patients in ambulatory care settings, where patients often present voluntarily to healthcare services (not unlike those who volunteered in this study) [62, 63]. Second, majority of the participants with dementia (70.9%) had the primary etiology of Alzheimer’s dementia. Although such large proportion of Alzheimer’s dementia is consistent with what is expected of the older population with dementia, the findings may vary in other populations with a different composition of dementia etiologies. Third, FAQ and MoCA short-variant may not necessarily be the most commonly used measures in many clinical practices, and were only used in this study as exemplars of informant reports and brief cognitive tests. Although they provided estimations on the relative performance across the 3 strategies, the results may plausibly vary when alternative measures are used. Fourth, the primary analysis in this study excluded participants with missing data in FAQ or MoCA, which may arguably bias the findings. However, further sensitivity analyses – using multiple imputation – showed that the results remained consistent even when these missing data could be accounted for. Although there can be other approaches to address missing data in diagnostic studies [64], multiple imputation can be a particularly useful approach to address the missing data in this study, given that there were many auxiliary variables that differed significantly between those with and without missing data (Supplementary Material 3), and hence these auxiliary variables could be included in multiple imputation to predict the missing data.
CONCLUSIONS
This study sought to identify the right strategy to employ for case-finding of cognitive impairment – across different ambulatory care settings with varying clinical resources and patient profiles – and potentially provide a more nuanced approach in case-finding efforts. As shown in this study, informant reports would suffice for case-finding of dementia, although brief cognitive tests may optionally be added-on – in services with available resources – to further improve the accuracy of dementia detection. For case-finding of both MCI and dementia, combined measures are preferred, as stand-alone informant reports or brief cognitive tests perform sub-optimally. Notably, the choice of case-finding strategy is not dependent on age profile, educational attainment, or prevalence of cognitive impairment of the clinical population.
Supplementary Material
ACKNOWLEDGEMENT
The NACC database is funded by NIA/NIH Grant U01 AG016976. NACC data are contributed by the NIA-funded ADCs: P30 AG019610 (PI Eric Reiman, MD), P30 AG013846 (PI Neil Kowall, MD), P50 AG008702 (PI Scott Small, MD), P50 AG025688 (PI Allan Levey, MD, PhD), P50 AG047266 (PI Todd Golde, MD, PhD), P30 AG010133 (PI Andrew Saykin, PsyD), P50 AG005146 (PI Marilyn Albert, PhD), P50 AG005134 (PI Bradley Hyman, MD, PhD), P50 AG016574 (PI Ronald Petersen, MD, PhD), P50 AG005138 (PI Mary Sano, PhD), P30 AG008051 (PI Thomas Wisniewski, MD), P30 AG013854 (PI M. Marsel Mesulam, MD), P30 AG008017 (PI Jeffrey Kaye, MD), P30 AG010161 (PI David Bennett, MD), P50 AG047366 (PI Victor Henderson, MD, MS), P30 AG010129 (PI Charles DeCarli, MD), P50 AG016573 (PI Frank LaFerla, PhD), P50 AG005131 (PI James Brewer, MD, PhD), P50 AG023501 (PI Bruce Miller, MD), P30 AG035982 (PI Russell Swerdlow, MD), P30 AG028383 (PI Linda Van Eldik, PhD), P30 AG053760 (PI Henry Paulson, MD, PhD), P30 AG010124 (PI John Trojanowski, MD, PhD), P50 AG005133 (PI Oscar Lopez, MD), P50 AG005142 (PI Helena Chui, MD), P30 AG012300 (PI Roger Rosenberg, MD), P30 AG049638 (PI Suzanne Craft, PhD), P50 AG005136 (PI Thomas Grabowski, MD), P50 AG033514 (PI Sanjay Asthana, MD, FRCP), P50 AG005681 (PI John Morris, MD), P50 AG047270 (PI Stephen Strittmatter, MD, PhD).
Footnotes
DISCLOSURE OF CONFLICTS OF INTEREST
The author declares no financial or other conflicts of interest.
DATA SHARING AND DATA ACCESSIBILITY
The data were obtained from the National Alzheimer’s Coordinating Center (NACC). For further information on access to the database, please contact NACC (contact details can be found at https://www.alz.washington.edu/WEB/researcher_home.html).
REFERENCES
- [1].Lang L, Clifford A, Wei L, et al. Prevalence and determinants of undetected dementia in the community: a systematic literature review and a meta-analysis. BMJ Open. 2017. 7: e011146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Beusterien KM, Thomas SK, Gause D, Kimel M, Arcona S, Mirski D. Impact of rivastigmine use on the risk of nursing home placement in a US sample. CNS Drugs. 2004. 18: 1143–1148. [DOI] [PubMed] [Google Scholar]
- [3].Geldmacher DS, Provenzano G, McRae T, Mastey V, Ieni JR. Donepezil is associated with delayed nursing home placement in patients with Alzheimer’s disease. J Am Geriatr Soc. 2003. 51: 937–944. [DOI] [PubMed] [Google Scholar]
- [4].Wattmo C, Wallin AK, Londos E, Minthon L. Risk factors for nursing home placement in Alzheimer’s disease: a longitudinal study of cognition, ADL, service utilization, and cholinesterase inhibitor treatment. Gerontologist. 2011. 51: 17–27. [DOI] [PubMed] [Google Scholar]
- [5].Becker M, Andel R, Rohrer L, Banks SM. The effect of cholinesterase inhibitors on risk of nursing home placement among medicaid beneficiaries with dementia. Alzheimer Dis Assoc Disord. 2006. 20: 147–152. [DOI] [PubMed] [Google Scholar]
- [6].Howard R, McShane R, Lindesay J, et al. Nursing home placement in the Donepezil and Memantine in Moderate to Severe Alzheimer’s Disease (DOMINO-AD) trial: secondary and post-hoc analyses. Lancet Neurol. 2015. 14: 1171–1181. [DOI] [PubMed] [Google Scholar]
- [7].Cepoiu-Martin M, Tam-Tham H, Patten S, Maxwell CJ, Hogan DB. Predictors of long-term care placement in persons with dementia: a systematic review and meta-analysis. Int J Geriatr Psychiatry. 2016. 31: 1151–1171. [DOI] [PubMed] [Google Scholar]
- [8].Thyrian JR, Hertel J, Wucherer D, et al. Effectiveness and Safety of Dementia Care Management in Primary Care: A Randomized Clinical Trial. JAMA Psychiatry. 2017. 74: 996–1004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Vickrey BG, Mittman BS, Connor KI, et al. The effect of a disease management intervention on quality and outcomes of dementia care: a randomized, controlled trial. Ann Intern Med. 2006. 145: 713–726. [DOI] [PubMed] [Google Scholar]
- [10].Jennings LA, Laffan AM, Schlissel AC, et al. Health Care Utilization and Cost Outcomes of a Comprehensive Dementia Care Program for Medicare Beneficiaries. JAMA Intern Med. 2019. 179: 161–166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Spijker A, Vernooij-Dassen M, Vasse E, et al. Effectiveness of nonpharmacological interventions in delaying the institutionalization of patients with dementia: a meta-analysis. J Am Geriatr Soc. 2008. 56: 1116–1128. [DOI] [PubMed] [Google Scholar]
- [12].Lee SJ, Larson EB, Dublin S, Walker R, Marcum Z, Barnes D. A Cohort Study of Healthcare Utilization in Older Adults with Undiagnosed Dementia. J Gen Intern Med. 2018. 33: 13–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Wray LO, Wade M, Beehler GP, Hershey LA, Vair CL. A program to improve detection of undiagnosed dementia in primary care and its association with healthcare utilization. Am J Geriatr Psychiatry. 2014. 22: 1282–1291. [DOI] [PubMed] [Google Scholar]
- [14].Prince M, Bryce R, Ferri C. World Alzheimer report 2011: the benefits of early diagnosis and intervention. 2011. [Google Scholar]
- [15].Morley JE, Morris JC, Berg-Weger M, et al. Brain health: the importance of recognizing cognitive impairment: an IAGG consensus conference. J Am Med Dir Assoc. 2015. 16: 731–739. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Gerontological Society of America. A 4-Step Process to Detecting Cognitive Impairment and Earlier Diagnosis of Dementia: Approaches and Tools for Primary Care Providers. 2017. [Google Scholar]
- [17].Galvin JE, Roe CM, Powlishta KK, et al. The AD8: a brief informant interview to detect dementia. Neurology. 2005. 65: 559–564. [DOI] [PubMed] [Google Scholar]
- [18].Pfeffer RI, Kurosaki TT, Harrah CH Jr., Chance JM, Filos S Measurement of functional activities in older adults in the community. J Gerontol. 1982. 37: 323–329. [DOI] [PubMed] [Google Scholar]
- [19].Borson S, Scanlan J, Brush M, Vitaliano P, Dokmak A. The Mini-Cog: a cognitive ‘vital signs’ measure for dementia screening in multi-lingual elderly. International Journal of Geriatric Psychiatry. 2000. 15: 1021–1027. [DOI] [PubMed] [Google Scholar]
- [20].Liew TM. The Optimal Short Version of Montreal Cognitive Assessment in Diagnosing Mild Cognitive Impairment and Dementia. J Am Med Dir Assoc. 2019. 20: 1055 e1051–1055 e1058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Liew TM. A 4-Item Case-Finding Tool to Detect Dementia in Older Persons. J Am Med Dir Assoc. 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].McDicken JA, Elliott E, Blayney G, et al. Accuracy of the short-form Montreal Cognitive Assessment: Systematic review and validation. Int J Geriatr Psychiatry. 2019. 34: 1515–1525. [DOI] [PubMed] [Google Scholar]
- [23].Galvin JE, Roe CM, Morris JC. Evaluation of cognitive impairment in older adults: combining brief informant and performance measures. Arch Neurol. 2007. 64: 718–724. [DOI] [PubMed] [Google Scholar]
- [24].Tew CW, Ng TP, Cheong CY, Yap P. A Brief Dementia Test with Subjective and Objective Measures. Dement Geriatr Cogn Dis Extra. 2015. 5: 341–349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Chan QL, Shaik MA, Xu J, Xu X, Chen CL, Dong Y. The Combined Utility of a Brief Functional Measure and Performance-Based Screening Test for Case Finding of Cognitive Impairment in Primary Healthcare. J Am Med Dir Assoc. 2016. 17: 372.e379–311. [DOI] [PubMed] [Google Scholar]
- [26].Roalf DR, Moberg PJ, Xie SX, Wolk DA, Moelter ST, Arnold SE. Comparative accuracies of two common screening instruments for classification of Alzheimer’s disease, mild cognitive impairment, and healthy aging. Alzheimers Dement. 2013. 9: 529–537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Ramlall S, Chipps J, Bhigjee AI, Pillay BJ. The sensitivity and specificity of subjective memory complaints and the subjective memory rating scale, deterioration cognitive observee, mini-mental state examination, six-item screener and clock drawing test in dementia screening. Dement Geriatr Cogn Disord. 2013. 36: 119–135. [DOI] [PubMed] [Google Scholar]
- [28].Larner AJ. AD8 Informant Questionnaire for Cognitive Impairment: Pragmatic Diagnostic Test Accuracy Study. J Geriatr Psychiatry Neurol. 2015. 28: 198–202. [DOI] [PubMed] [Google Scholar]
- [29].Kan CN, Zhang L, Cheng CY, et al. The Informant AD8 Can Discriminate Patients with Dementia From Healthy Control Participants in an Asian Older Cohort. J Am Med Dir Assoc. 2019. 20: 775–779. [DOI] [PubMed] [Google Scholar]
- [30].Cecato JF, Martinelli JE, Izbicki R, Yassuda MS, Aprahamian I. A subtest analysis of the Montreal cognitive assessment (MoCA): which subtests can best discriminate between healthy controls, mild cognitive impairment and Alzheimer’s disease? International Psychogeriatrics. 2016. 28: 825–832. [DOI] [PubMed] [Google Scholar]
- [31].Liew TM. Developing a Brief Neuropsychological Battery for Early Diagnosis of Cognitive Impairment. J Am Med Dir Assoc. 2019. 20: 1054.e1011–1054.e1020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Wang S, Yin H, Wang X, et al. Efficacy of different types of exercises on global cognition in adults with mild cognitive impairment: a network meta-analysis. Aging Clin Exp Res. 2019. 31: 1391–1400. [DOI] [PubMed] [Google Scholar]
- [33].Loprinzi PD, Blough J, Ryu S, Kang M. Experimental effects of exercise on memory function among mild cognitive impairment: systematic review and meta-analysis. Phys Sportsmed. 2019. 47: 21–26. [DOI] [PubMed] [Google Scholar]
- [34].Zhang H, Huntley J, Bhome R, et al. Effect of computerised cognitive training on cognitive outcomes in mild cognitive impairment: a systematic review and meta-analysis. BMJ Open. 2019. 9: e027062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [35].Sherman DS, Mauser J, Nuno M, Sherzai D. The Efficacy of Cognitive Intervention in Mild Cognitive Impairment (MCI): a Meta-Analysis of Outcomes on Neuropsychological Measures. Neuropsychol Rev. 2017. 27: 440–484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [36].Langa KM, Levine DA. The Diagnosis and Management of Mild Cognitive Impairment: A Clinical Review. Jama. 2014. 312: 2551–2561. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [37].Beekly DL, Ramos EM, van Belle G, et al. The National Alzheimer’s Coordinating Center (NACC) Database: an Alzheimer disease database. Alzheimer Dis Assoc Disord. 2004. 18: 270–277. [PubMed] [Google Scholar]
- [38].Noel-Storr AH, McCleery JM, Richard E, et al. Reporting standards for studies of diagnostic test accuracy in dementia: The STARDdem Initiative. Neurology. 2014. 83: 364–373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [39].Weintraub S, Besser L, Dodge HH, et al. Version 3 of the Alzheimer Disease Centers’ Neuropsychological Test Battery in the Uniform Data Set (UDS). Alzheimer Dis Assoc Disord. 2018. 32: 10–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [40].Petersen RC, Morris JC. Mild cognitive impairment as a clinical entity and treatment target. Arch Neurol. 2005. 62: 1160–1163; discussion 1167. [DOI] [PubMed] [Google Scholar]
- [41].McKhann GM, Knopman DS, Chertkow H, et al. 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. Alzheimer’s & dementia : the journal of the Alzheimer’s Association. 2011. 7: 263–269. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [42].Román GC, Tatemichi TK, Erkinjuntti T, et al. Vascular dementia. Diagnostic criteria for research studies: Report of the NINDS‐AIREN International Workshop*. 1993. 43: 250–250. [DOI] [PubMed] [Google Scholar]
- [43].McKeith IG, Boeve BF, Dickson DW, et al. Diagnosis and management of dementia with Lewy bodies: Fourth consensus report of the DLB Consortium. Neurology. 2017. 89: 88–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [44].Litvan I, Bhatia KP, Burn DJ, et al. Movement Disorders Society Scientific Issues Committee report: SIC Task Force appraisal of clinical diagnostic criteria for Parkinsonian disorders. Mov Disord. 2003. 18: 467–486. [DOI] [PubMed] [Google Scholar]
- [45].Bensimon G, Ludolph A, Agid Y, Vidailhet M, Payan C, Leigh PN. Riluzole treatment, survival and diagnostic criteria in Parkinson plus disorders: the NNIPPS study. Brain. 2009. 132: 156–171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [46].Armstrong MJ, Litvan I, Lang AE, et al. Criteria for the diagnosis of corticobasal degeneration. Neurology. 2013. 80: 496–503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [47].Brooks BR, Miller RG, Swash M, Munsat TL. El Escorial revisited: revised criteria for the diagnosis of amyotrophic lateral sclerosis. Amyotroph Lateral Scler Other Motor Neuron Disord. 2000. 1: 293–299. [DOI] [PubMed] [Google Scholar]
- [48].Rascovsky K, Hodges JR, Knopman D, et al. Sensitivity of revised diagnostic criteria for the behavioural variant of frontotemporal dementia. Brain : a journal of neurology. 2011. 134: 2456–2477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [49].Nasreddine ZS, Phillips NA, Bedirian V, et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005. 53: 695–699. [DOI] [PubMed] [Google Scholar]
- [50].Hosmer DW, Lemeshow S, Sturdivant RX. Applied logistic regression. Third edition edn. New Jersey: Wiley, 2013. [Google Scholar]
- [51].DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988. 44: 837–845. [PubMed] [Google Scholar]
- [52].Yu J, Yap P, Liew TM. The optimal short version of the Zarit Burden Interview for dementia caregivers: diagnostic utility and externally validated cutoffs. Aging Ment Health. 2018: 1–5. [DOI] [PubMed] [Google Scholar]
- [53].Liew TM, Yap P. A 3-Item Screening Scale for Caregiver Burden in Dementia Caregiving: Scale Development and Score Mapping to the 22-Item Zarit Burden Interview. J Am Med Dir Assoc. 2019. 20: 629–633 e612. [DOI] [PubMed] [Google Scholar]
- [54].Buschke H, Kuslansky G, Katz M, et al. Screening for dementia with the memory impairment screen. Neurology. 1999. 52: 231–238. [DOI] [PubMed] [Google Scholar]
- [55].Brodaty H, Pond D, Kemp NM, et al. The GPCOG: a new screening test for dementia designed for general practice. J Am Geriatr Soc. 2002. 50: 530–534. [DOI] [PubMed] [Google Scholar]
- [56].Tsoi KK, Chan JY, Hirai HW, Wong SY, Kwok TC. Cognitive Tests to Detect Dementia: A Systematic Review and Meta-analysis. JAMA Intern Med. 2015. 175: 1450–1458. [DOI] [PubMed] [Google Scholar]
- [57].Breton A, Casey D, Arnaoutoglou NA. Cognitive tests for the detection of mild cognitive impairment (MCI), the prodromal stage of dementia: Meta-analysis of diagnostic accuracy studies. Int J Geriatr Psychiatry. 2019. 34: 233–242. [DOI] [PubMed] [Google Scholar]
- [58].American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders: DSM-5. Washington, DC: Amer Psychiatric Pub Incorporated, 2013. [Google Scholar]
- [59].Liew TM, Feng L, Gao Q, Ng TP, Yap P. Diagnostic utility of Montreal Cognitive Assessment in the Fifth Edition of Diagnostic and Statistical Manual of Mental Disorders: major and mild neurocognitive disorders. J Am Med Dir Assoc. 2015. 16: 144–148. [DOI] [PubMed] [Google Scholar]
- [60].Liew TM, Yap P, Feng L, Gao Q, Ng TP. Response to the Letter from Lu et al, “Utility of Montreal Cognitive Assessment (Hong Kong Version) in the Diagnosis of Mild Neurocognitive Disorders (NCD): NCD Due to Alzheimer Disease (NCD-AD) and NCD Due to Vascular Disease (NCD-Vascular)”. J Am Med Dir Assoc. 2016. 17: 367. [DOI] [PubMed] [Google Scholar]
- [61].Leeflang MM, Bossuyt PM, Irwig L. Diagnostic test accuracy may vary with prevalence: implications for evidence-based diagnosis. J Clin Epidemiol. 2009. 62: 5–12. [DOI] [PubMed] [Google Scholar]
- [62].Liew TM. Neuropsychiatric symptoms in cognitively normal older persons, and the association with Alzheimer’s and non-Alzheimer’s dementia. Alzheimers Res Ther. 2020. 12: 35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [63].Liew TM. Depression, subjective cognitive decline, and the risk of neurocognitive disorders. Alzheimers Res Ther. 2019. 11: 70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [64].Lees RA, Hendry Ba K, Broomfield N, Stott D, Larner AJ, Quinn TJ. Cognitive assessment in stroke: feasibility and test properties using differing approaches to scoring of incomplete items. Int J Geriatr Psychiatry. 2017. 32: 1072–1078. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.