A mobile interactive cognitive self-assessment scale for screening Alzheimer’s disease

doi:10.21203/rs.3.rs-3938896/v1

Download PDF

Research Article

A mobile interactive cognitive self-assessment scale for screening Alzheimer’s disease

https://doi.org/10.21203/rs.3.rs-3938896/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

A feasible self-administered cognition scale with rigorous validation and high diagnostic accuracy is lacking for screening for cognitive impairment due to Alzheimer’s disease (AD).

Methods

An interactive cognitive self-assessment scale (CogSAS) was designed through the Delphi process, and 518 participants were subjected to item optimization. The scale was validated in 358 cognitively unimpaired and 396 cognitively impaired participants for reliability, validity, and diagnostic accuracy. Specificity and sensitivity were tested for 38 participants with cognitive impairment and 45 participants with cognitive impairment due to AD.

Results

The interactive CogSAS relies on speech recognition to achieve elderly friendly results. The internal consistency was 0.81, and the test-retest reliability was 0.82. The construct validity was 0.74, and the criterion validity was 0.77. The sensitivity and specificity for clinically diagnosed participants were 0.90 and 0.67, respectively. For cognitive impairment due to AD, the sensitivity and specificity were 1.00 and 0.78, respectively.

Conclusions

The CogSAS has good reliability, validity, and feasibility. It also showed high sensitivity and specificity in identifying cognitive impairment due to AD.

Alzheimer’s disease

diagnostic accuracy

mobile cognitive scale

screening measure

validation

Alzheimer’s disease (AD) is the leading cause of dementia in the general community.[1] Thus, early diagnosis of cognitive impairment due to AD, including mild cognitive impairment (MCI) and dementia, can improve functional preservation and primary care.[2] To achieve early diagnosis through community screenings, an efficient and self-administered cognitive scale would be optimal. However, current digital scales cannot provide satisfactory diagnostic accuracy and feasibility in screening for cognitive impairment due to AD with rigorous validity in large populations and are therefore impractical for community screening.

Numerous attempts have been made to develop a novel digital cognitive assessment method to solve this problem.[3] To address the inconvenience of labor-consuming traditional pen and paper tests, some have been converted to digital versions.[4] However, these computerized scales, such as the Cogstate and Montreal Cognitive Assessment (MoCA),[5, 6] have shown weak validity and inconsistent sensitivity in discriminating cognitive impairment due to AD. Most of the scales still require supervision and explanation by neurologists.[7, 8] These problems have affected the credibility and feasibility of revised scales for screening for cognitive impairment due to AD. To improve their feasibility and credibility, novel digital cognition tests using mobile applications have also been developed.[8–10] However, the validation of existing scales for large populations of patients with dementia has been poor. In addition, the feasibility of using existing mobile health scales was still influenced by education level. In China’s aging population, 29.6% of them are illiterate, and 41.5% of them are primary school graduates.[11] Low education levels limit the use of mobile applications, while existing scales require certain abilities to use mobile phones. Therefore, a novel cognitive assessment scale with improved performance in discriminating cognitive impairment due to AD, which is rigorously validated and elderly friendly, is urgently needed.

Thus, we developed an interactive cognitive self-assessment scale (CogSAS) that covers important domains of impairment in AD patients and was designed for assessing cognitive impairment through AD screening. This scale needs to be rigorously validated and made more feasible through speech recognition and other human-computer interaction techniques. To test the ability of the scale to distinguish cognitive impairment from AD pathology, it must also be both sensitive and specific within clinically diagnosed participants and those with cognitive impairment due to AD.

The flowchart in Fig. 1 shows a summary of the three phases in which the CogSAS was developed and validated. The study was conducted from 2017 to 2023, from scale development to validation analysis, and included six medical centres in Beijing, China.

2.1 Procedures

2.1.1 Scale development

2.1.1.1 Design of the initial CogSAS version

Our purpose was to create a cognitive assessment scale according to the characteristic domains of cognitive impairment in AD patients, including the assessment of both memory and executive functions. The scale was designed as a mini-application for touch panels, smartphones, and personal computers that works on both the WeChat app and the website. The scale was designed to be used alone or accompanied by caregivers, including social workers and family members. Immediate feedback of the results was designed to be presented at the end of the test. The scale was developed through the administration of cognitive tests, feedback assessment reports, and automatic data upload to a server. To increase the feasibility of this scale for aging participants, the scale was designed to be particularly age friendly. First, the scale was adapted for the touchscreen to make it easier to use. Second, to increase the convenience for elderly participants, we applied speech recognition to the scale to provide instructions and for the memory task procedure. By using voice instead of words, the influence of education level on this scale was reduced compared to traditional pen and paper tests. Two human–computer interaction technicians were recruited to provide technical support.

The Delphi technique was used to develop the initial version of the scale. The expert panel included eleven specialists in dementia, including three neuropsychologists, five neurologists, and three geriatricians (Additional file 1: Table S1). All the experts were invited on the basis of their academic credentials and experience in clinical practice specializing in dementia and geriatrics. All the experts were willing to work through the entire scale development process.

The expert panel was asked to develop the initial version of the CogSAS, which included memory and executive function tasks and three self-reported measures of psychological status and activities of daily living. Reviews of original research papers and literature reviews were also conducted to form the questionnaires for item selection. All the experts were sent to the online survey questionnaires. All the responses were assured of confidentiality throughout the whole process. The items were selected through three rounds of the Delphi process, with each round lasting more than 2 weeks. During round 1, the experts were asked to select the ideal items for the scale and suggest additional items for inclusion in round 2. Reasons and amendments were also requested from those who disagreed. During round 2, in addition to item selection, the experts were asked to present the scoring rules for each item. A further round (third round) was carried out for the final selections of items and the establishment of the scoring rules. In all three rounds of the Delphi process, the response rate of the expert panel was 100%. Items with more than 70% agreement were accepted. After all three Delphi rounds, offline discussions were held to review the results of the Delphi survey. The results of each Delphi process are presented (Additional file 3 and Additional file 4). All members of the expert panel participated in the discussion and decided on the initial version of the CogSAS. The initial version of the CogSAS was approved by all members of the expert panel.

2.1.1.2 Item optimization

To further improve the credibility of the scale, item optimization was conducted for the cohort after the Delphi process. The aim of item optimization was to select applicable, sensitive items from the initial CogSAS version, adjust the scoring scheme, and establish the final CogSAS version. Item optimization was conducted with both cognitively impaired and cognitively unimpaired (CU) participants, applying receiver operating characteristic (ROC) analysis.

Scale item optimization was conducted with 518 participants recruited from communities. Participants, all of whom were older than 55 years, were classified as CU or cognitively impaired, the latter of which included dementia and MCI. Dementia was diagnosed according to the DSM-V criteria,[12] and MCI was diagnosed according to the Peterson criteria.[13] The Clinical Dementia Rating (CDR) was also used in the diagnosis of MCI and dementia. The eligibility criteria for participants without cognitive impairment were no reports of cognitively impaired performance and a CDR score of 0. Participants who were unable to cooperate on the neuropsychological examinations and digital scale were excluded. Participants were interviewed by trained neuropsychologists who administered the Mini-Mental State Examination (MMSE) and MoCA within 24 hours. Participants were then asked to complete the initial CogSAS version without assistance via a touch panel, smartphone, or computer. All participants’ performances were recorded and evaluated. The demographic and cognitive function data are shown (Additional file 1: Table S2).

To select applicable items, sensitive items, operation times and scores were analysed. ROC analysis was used to optimize the scale. The item optimization criteria were as follows: 1) initial CogSAS version items were excluded if their area under the curve (AUC) was < 0.6, and 2) initial CogSAS items that were completed by < 70% of participants were also excluded. Additionally, we believe that the rather long operation time made the participants impatient. The cognition status of cognitively impaired participants may also impede their understanding of self-evaluation questions. Hence, the self-reports section was deleted, which decreased the operation time from 30 minutes to 15 minutes. The results of item optimization are shown. (Additional file 1: Table S3)

The scale score was also adjusted to increase item sensitivity. Regression and canonical discrimination analyses were also conducted to improve the screening methods used. Different assignments were designed to reach ideal validity and reliability. Other modifications, including adding time intervals between items, changing wording, and recording participants’ years of education, were also assessed.

2.1.2 Scale validation in participants clinically diagnosed with cognitive impairment

In phase 2, scale validation aimed to test the validity and reliability of the final CogSAS version. Validation was conducted with participants who were clinically diagnosed with cognitive impairment by neuropsychological assessment and those who were CU.

In addition to those who participated in item optimization, 754 participants were recruited from the communities and six medical centers (Xuanwu Hospital, Fuxing Hospital, Beijing Friendship Hospital affiliated with Capital Medical University, China-Japan Friendship Hospital, Peking University People’s Hospital, and Beijing Yangfangdian Hospital) and were enrolled in phase 2; the inclusion criteria, diagnostic criteria, exclusion criteria, clinical data collection, and neuropsychological assessments were the same as those described above for phase 1.

Participants were asked to complete the final version of the CogSAS. Internal consistency, test-retest reliability, construct validity, and criterion validity were used to test the reliability and validity of the CogSAS. A total of 70 participants were retested in the clinic 2–4 weeks after the initial tests for test-retest reliability. The ROC curve was also used to assess the diagnostic ability for identifying participants who were cognitively impaired. A cut-off value for the total score was calculated based on the discrimination between participants who were CU and those who were cognitively impaired. The sensitivity, specificity, and likelihood ratio were also assessed.

2.1.3 Scale validation in participants with cognitive impairment due to AD

Given that the CogSAS was designed to screen for cognitive impairment due to AD, participants with cognitive impairment with AD pathology were also recruited to test its ability to discriminate this form of cognitive impairment.

Phase 2 participants who completed comprehensive clinical data collection, magnetic resonance imaging (MRI), and lumbar puncture were enrolled in phase 3. A total of 83 participants were enrolled with a diagnosis supported by the amyloid, tau, neurodegeneration (ATN) classification system.[14, 15] All participants were classified based on the Phase 2 inclusion criteria. For those who were classified as CU, ‘A-T-N-’ was tested; for those who were classified as cognitively impaired, ‘A + T + N+’ was tested.

After the MRI scan, a clinical diagnosis-blind MTA score was assigned independently by two experienced neurologists according to the standard criteria.[16, 17] A lumbar puncture was also performed within one week after the MRI. Cerebrospinal fluid Aβ and tau protein levels were assessed by a qualified laboratory. Detailed clinical data were collected, and the ATN classification system was used,[14] where ‘A+’ was defined as an Aβ42/40 ratio < 0.1, ‘T+’ was defined as a decreased phospho-tau value (< 35.84 pg/ml), and ‘N+’ was defined as an MTA score ≥ 1.

The sensitivity, specificity, and likelihood were also re-examined and compared with the diagnostic accuracy in phase 2.

2.2 Statistical analysis

The data were analysed using the Statistical Package for the Social Sciences (PC version 23.0; IBM, Inc., New York, USA) and GraphPad Prism software (version 8.3.0; GraphPad Software, Inc., La Jolla, CA). Between-group differences in age, years of education, and cognitive scores were assessed by Student’s t test or ANOVA with Tukey’s post hoc test. Sex differences were assessed using a chi-square test.

The reliability and validity of the scale were assessed. Reliability refers to consistency when the scale is repeated under identical conditions and is assessed by test-retest reliability and internal consistency.[18] Test-retest reliability was used to assess the consistency of sum scores across time using Pearson’s correlation coefficients.[19] The higher the correlation is, the greater the test-retest reliability.[20] Internal consistency was the degree to which the set of items varied relative to the sum of the scores and was measured using Cronbach’s alpha coefficients. The acceptable threshold of the alpha coefficient was 0.7 for reliability.[18]

Validity is used to test whether the instrument indeed measures the latent dimension or aimed-evaluated construct, which is assessed by construct validity and criterion validity.[19] We constructed validity tests to determine whether the scale measures the intended construct by applying the Kaiser–Meyer–Olkin (KMO) test and Bartlett’s test of factor analysis. The KMO test should be at least 0.5, and Bartlett’s test showed significant results.[21] Criterion validity measures the relationship between the sum score and the criterion, typically referred to as the gold standard, which is assessed by Pearson’s correlation coefficients. A higher correlation coefficient indicates greater validity.[19] ROC curves were plotted to visualize sensitivity and specificity and to determine cut-offs. The AUC was used to assess diagnostic accuracy. Youden’s index was used to select optimal cut-off values. p = 0.05 was considered to indicate statistical significance.

3.1 Scale development

The final CogSAS version included three tasks that could be completed within 15 minutes and had a total scale score ranging from 0–64 (Fig. 2). The final version was applicable to touch screens and used speech recognition techniques. The task details are as follows:

Task 1—The sentence memory task (SEN): The SEN is a sentence repetition task in which the participant is asked to repeat a sentence following a voice prompt. The sentence describes a short story containing essential verbal elements, each of which is worth 8 points. The participant was asked to repeat the sentence immediately after hearing it for the first time. After the first repetition, the participant listened to the sentence again and was asked to repeat the sentence a second time. Approximately 10 minutes later, at the end of the full scale, the participant was asked to recall the same sentence a third time. After each repetition, the participant pressed the stop button, and the speech was automatically uploaded to the speech recognition system. Each task item is scored according to the number of elements recalled, for a total range of 0–8 for each trial.

Task 2—Visual memory task (VIS): The VIS consists of 20 color pictures, 10 of which have symmetrical geometry, 5 of which involve insects or animals, and 5 of which involve daily activities (Additional file 4). All the pictures were chosen on the basis of the Chinese cultural background and familiarity with elderly people. These pictures (set 1) are shown consecutively on screen for 10 seconds each. The participant was asked to remember each picture (encoding phase). The recognition task was administered 5 minutes later, and a target picture and a distractor were included (set 2). The distractor is designed to resemble the target picture in format or color. The time limit for each item was 10 seconds. Each correct response is worth 1 point on a scale from 0 to 20, with higher scores indicating better recall.

Task 3—The executive function task (EXE): The EXE assesses responses to conflicting instructions. Four finger shapes are shown on the screen, each pointing in a different direction (response key). Then, finger-shaped instruction keys are presented for the four finger shapes. If the color of the instruction finger was white, the participant was asked to press the response key pointing in the same direction as the instruction finger. If the color of the instruction finger was yellow, the participant was asked to press the response key opposite the instruction finger. The white and yellow instruction fingers are presented in random order at 10-second intervals. Participants are given a practice round before the actual 90-second task is administered. Scores range from 0–20 based on the total number of correct responses.

3.2 Scale validity and reliability analyses

Validity and reliability analyses were also conducted for the participants in phase 2; 358 of whom were classified as CU, and 396 of whom were clinically diagnosed as having cognitive impairment. There were no significant age or sex differences between the cognitively impaired and CU participants. Those who were cognitively impaired had fewer years of education (p < 0.05), and both the MMSE and MoCA scores were lower among participants who were cognitively impaired (p < 0.05). (Table 1)

Table 1

Demographics of phase 2 participants with clinically diagnosed cognitive impairment and cognitively unimpaired.
	Cognitively unimpaired	Cognitive impairment	p value
Number	358	396
Age	68.49 ± 6.62	69.52 ± 7.89	0.052
Sex (Male/Female)	173/185	168/228	0.121
Years of education	12.90 ± 3.69	10.77 ± 4.27	0.000^*
MMSE	28.63 ± 1.38	22.84 ± 5.56	0.000^*
MoCA	25.67 ± 2.68	17.19 ± 5.67	0.000^*
^*p < 0.05

3.2.1 Reliability analysis

The internal consistency of the CogSAS was assessed using Cronbach’s alpha coefficient to determine reliability. The Cronbach’s alpha coefficient was 0.81, indicating the internal consistency of the scale.

Test-retest reliability was assessed using Pearson’s correlation, and the coefficient was 0.82 (p < 0.001) between the first test and the repeated test.

3.2.2 Validity analysis

Construct validity was assessed using the KMO (0.74) and Bartlett’s tests (0.00), which verified the validity of the scale.

Criterion validity was verified by comparing the scale with the MMSE and MoCA as the ‘gold standards’. The CogSAS total score was positively correlated with both the MMSE (R = 0.77, p < 0.001) and the MoCA (R = 0.76, p < 0.001).

3.3 Sensitivity and specificity analyses

Sensitivity and specificity were both tested by receiver operating characteristic (ROC) curves for patients in phase 2 and phase 3. The same participants from phase 2 who conducted the validity and reliability analyses were used in the sensitivity and specificity analyses. The demographic data of patients in phase 2 are shown above.

Of the participants in phase 3, 38 were classified as CU, and 45 were diagnosed with cognitive impairment due to AD based on the ATN classification system. No significant differences were found in age or sex between the cognitively impaired and CU participants. A lower level of education and lower MMSE and MoCA scores were found in the cognitively impaired participants (p < 0.05) (Table 2).

Table 2

Demographics of phase 3 participants with cognitive impairment due to AD and cognitively unimpaired
	Cognitively unimpaired	Cognitive impairment	p value
Number	38	45
Age	65.68 ± 6.30	66.36 ± 7.59	0.666
Sex (Male/Female)	18/20	19/26	0.664
Years of education	11.79 ± 4.21	9.67 ± 4.24	0.025^*
MMSE	28.42 ± 1.00	17.38 ± 6.70	0.000^*
MoCA	25.18 ± 2.25	13.20 ± 6.86	0.000^*
^*p < 0.05

3.3.1 Sensitivity and specificity in clinically diagnosed participants

Among clinically diagnosed participants in phase 2, the ROC curve showed an optimal cut-off point of 38.7 with an AUC of 0.86, a sensitivity of 0.90 and a specificity of 0.67. The likelihood ratio was 2.73, and the negative likelihood ratio was 0.15 (Table 3 and Fig. 3). Using ROC analysis, the optimal cut-off, sensitivity and specificity for each subtest were also evaluated (Additional file 1: Table S4 and Additional file 2: Figure S1).

Table 3

Comparison of diagnostic accuracy between participants with clinically diagnosed with cognitive impairment and those who are cognitive impairment due to AD
	Phase 2 Diagnostic accuracy in participants with clinically diagnosed cognitive impairment	Phase 3 Diagnostic accuracy in participants with cognitive impairment due to AD
Number	754	83
Sensitivity	0.90	1.00
Specificity	0.67	0.78
Positive likelihood ratio	2.73	4.54
Negative likelihood ratio	0.15	0.00

3.3.2 Sensitivity and specificity in assessing cognitive impairment due to AD

The CogSAS was subsequently used to classify participants in phase 3 with cognitive impairment due to AD or CU based on the cut-off value. The ROC curve showed a sensitivity of 1.00, specificity of 0.78, and an AUC of 0.93, outperforming those of participants who were clinically diagnosed in phase 2. The likelihood ratio was 4.54, and the negative likelihood ratio was 0.00 (Table 3 and Fig. 3). The optimal cut-off, sensitivity and specificity for each subtest were also calculated (Additional file 1: Table S5 and Additional file 2: Figure S2).

For those with cognitive impairment due to AD, the cut-off point was used to classify the participants according to their ATN classification system-based diagnosis. In phase 3, 85.5% of participants were correctly classified by the CogSAS.

The aim of the present study was to develop and validate the CogSAS, which includes the assessment of cognitive domains impaired in AD patients. Compared with previous mobile health scales, the CogSAS is suitable for use in large populations and has good validity, reliability, and feasibility. The scale also demonstrated high diagnostic accuracy for cognitive impairment due to AD, supporting its feasibility for screening and early detection of cognitive impairment due to AD.

The scale was developed through a rigorous process that ensures its credibility and feasibility. During the scale development process, we used Delphi techniques to ensure scale expertise. In addition, CogSAS items were optimized for use in large populations to improve diagnostic accuracy, whereas other scales were mainly reviewed by an expert panel.[22] Unlike in the opinion of experts, database optimization of the CogSAS was more credible. Furthermore, we recruited participants with AD pathology, which improved the credibility of identifying cognitive impairment due to AD. Additionally, we used speech recognition and other human–computer interaction techniques on this scale to reduce the influence of education level. Compared to text, the use of voice prompting and speech recognition could broaden the application of this tool in community screening, regardless of education level. We have also made efforts in Mandarin accent recognition to promote its application in more regions of China. Overall, due to the rigorous development process, the CogSAS has solid credibility and feasibility.

Through the development process, the CogSAS was validated and showed relatively good validity, reliability, and high diagnostic accuracy. The Cronbach’s alpha was 0.81, and the test-retest reliability was 0.82; both of these values are within the acceptable range (0.70–0.90).[23] The validity of previous digital scales was comparable to 0.77 for CANS-MCI, while some others were not well validated.[24] Furthermore, construct validity was 0.74, and criterion validity was 0.77 in the CogSAS. This test thus reached acceptable validity.[25] Previous digital scales showed variable criterion validity (0.21–0.62) and construct validity (0.76–0.77).[24, 26, 27] These results suggest that the CogSAS is credible for identifying people with cognitive impairment.

Diagnostic accuracy was reflected in the sensitivity, specificity, and likelihood ratio. An ideal screening scale should have high sensitivity to ensure follow-up diagnosis in vulnerable individuals,[28] while moderate specificity is acceptable. The CogSAS showed high sensitivity (up to 1.00) and acceptable specificity, suggesting that it is an ideal tool for community screening. Compared with existing digital scales, the sensitivity of these scales varies from 0.41–1.00, and the specificity varies from 0.6–0.96.[4] The CogSAS also showed acceptable likelihood ratios (positive likelihood 2.73–4.54, negative likelihood ratio 0.00–0.15)[29].

The CogSAS showed good performance for discriminating cognitive impairment from AD pathology. First, the scale was designed primarily for participants with cognitive impairment due to AD, using minimal assessments and the most feasible methods. Memory loss is a hallmark of AD, with an emphasis on the characteristic effect on episodic and working memory.[30, 31] Here, we used the sentence memory task to assess episodic memory and a contradicting direction test to assess both working memory and executive function. Second, in assessing criterion validity, a high correlation coefficient was found with both the MMSE and the MoCA, which are well recognized for diagnosing MCI and dementia due to AD.[32] Finally, a higher sensitivity (1.00) was found for detecting cognitive impairment due to AD pathology. Compared with the sensitivities of 0.7–0.9 for existing scales that discriminate cognitive impairment,[4] the CogSAS showed higher sensitivity, suggesting that it may allow more accurate screening for cognitive impairment due to AD.

Despite its strengths, this study was not without limitations. First, the number of subgroups with biomarkers was relatively small; therefore, replication with larger samples is a priority. Second, sensitivity and specificity were assessed only in the clinic; future studies should thus focus on longitudinal community monitoring and mass screening. Finally, the study recruited mainly participants with MCI and mild AD, while subjective cognitive decline (SCD) was included with earlier intervention.

A novel, interactive cognitive assessment scale, the CogSAS, was developed. The results indicate that the scale has good reliability, validity, feasibility, and good diagnostic accuracy. The scale is able to discriminate cognitive impairment, especially in the presence of AD pathology. The CogSAS may improve mass screening for cognitive impairment, leading to earlier diagnosis.

MCI mild cognitive impairment

AD Alzheimer’s disease

CU cognitively unimpaired

CogSAS cognitive self-assessment scale

MMSE Mini-Mental State Examination

MoCA Montreal Cognitive Assessment

CDR Clinical Dementia Rating

MTA medial temporal lobe atrophy

ROC receiver operating characteristic

AUC area under the curve

SD standard deviation

MRI magnetic resonance imaging

KMO Kaiser–Meyer–Olkin

SEN Sentence Memory Task

VIS visual memory task

EXE executive function task

ATN amyloid, tau, neurodegeneration

SCD subjective cognitive decline

Ethics approval and consent to participate

The study protocols were approved by the Ethics Committee and Institutional Review Board of Xuanwu Hospital, Capital Medical University, Beijing (Approval No. LXS(2018)093). All methods and experiments were performed following the relevant guidelines and regulations. All participants enrolled in the study, or their guardians, signed an informed written consent form, giving specific approval for this study before the study commenced.

Consent for publication

Not applicable.

Availability of data and materials

The data sets analysed in the current study can be made available from the corresponding author upon reasonable request.

Competing interest

The authors declare no competing interests.

Funding

This work was supported by the Beijing Municipal Science and Technology Committee [D171100008217005] and by the National Natural Science Foundation of China [81971011, 82271464].

Authors’ Contributions

DL, YC, XC, FL, LW, TX, MT, DP, JW, JL.L and WC were responsible for the study concept and design of the scale; KX, MT, DP, JW, JL, WC, YW, MC and LL established the two cohorts and coordinated data collection; JB.L and TC were responsible for the creation of new digital cognitive scale; KX and JH performed the statistical analysis. KX drafted the paper and prepared the figures. All authors reviewed and critically revised the draft manuscript. Authors approved the final manuscript.

Acknowledgements

We thank all the study participants and community social workers who contributed to the study.

Jia L, Quan M, Fu Y, et al. Dementia in China: epidemiology, clinical management, and research advances. Lancet Neurol 2020;19.
Rasmussen J, Langerman H. Alzheimer's Disease - Why We Need Early Diagnosis. Degener Neurol Neuromuscul Dis 2019;9.
Chan JYC, Kwong JSW, Wong A, Kwok TCY, Tsoi KKF. Comparison of Computerized and Paper-and-Pencil Memory Tests in Detection of Mild Cognitive Impairment and Dementia: A Systematic Review and Meta-analysis of Diagnostic Studies. J Am Med Dir Assoc 2018;19.
Chan JYC, Yau STY, Kwok TCY, Tsoi KKF. Diagnostic performance of digital cognitive tests for the identification of MCI and dementia: A systematic review. Ageing Res Rev 2021;72.
Alden EC, Pudumjee SB, Lundt ES, et al. Diagnostic accuracy of the Cogstate Brief Battery for prevalent MCI and prodromal AD (MCI A(+) T(+) ) in a population-based sample. Alzheimers Dement 2021;17.
Katz MJ, Wang C, Nester CO, et al. T-MoCA: A valid phone screen for cognitive impairment in diverse community samples. Alzheimers Dement (Amst) 2021;13.
Hammers D, Spurgeon E, Ryan K, et al. Validity of a brief computerized cognitive screening test in dementia. J Geriatr Psychiatry Neurol 2012;25.
Thabtah F, Peebles D, Retzler J, Hathurusingha C. Dementia medical screening using mobile applications: A systematic review with a new mapping model. Journal of Biomedical Informatics 2020;111.
Fukui Y, Yamashita T, Hishikawa N, et al. Computerized touch-panel screening tests for detecting mild cognitive impairment and Alzheimer's disease. Intern Med 2015;54.
Cabinio M, Rossetto F, Isernia S, et al. The Use of a Virtual Reality Platform for the Assessment of the Memory Decline and the Hippocampal Neural Injury in Subjects with Mild Cognitive Impairment: The Validity of Smart Aging Serious Game (SASG). J Clin Med 2020;9.
Jing L. Development report on the quality of life for the elderly in China.2019.
APA. Diagnostic and Statistical Manual of Mental Disorders: DSM-V. 1994.
Petersen RC. Mild cognitive impairment as a diagnostic entity. Journal of Internal Medicine 2004;256.
Jack CR. A/T/N: An unbiased descriptive classification scheme for Alzheimer disease biomarkers. Neurology 2016;87.
Murray, Alison D. A new biomarker classification system for AD, independent of cognition. Neurology 2016;87.
Scheltens P, Launer LJ, Barkhof F, Weinstein HC, Gool WAV. Visual assessment of medial temporal lobe atrophy on magnetic resonance imaging: Interobserver reliability. Journal of Neurology 1995;242.
Wahlund LO, Julin P, Johansson SE, Scheltens P. Visual rating and volumetry of the medial temporal lobe on magnetic resonance imaging in dementia: a comparative study. Journal of neurology, neurosurgery, and psychiatry;69.
Boateng GO, Neilands TB, Frongillo EA, Melgar-Quiñonez HR, Young SL. Best Practices for Developing and Validating Scales for Health, Social, and Behavioral Research: A Primer. Frontiers in Public Health 2018;6.
Raykov T, Marcoulides GA. Introduction to Psychometric Theory: Introduction to psychometric theory.
Rousson V, Gasser T, Seifert B. Assessing intrarater, interrater and test-retest reliability of continuous measurements. Statistics in Medicine 2010;21.
Kaiser HF, Rice J. Little Jiffy, Mark Iv. Educational and Psychological Measurement 2016;34.
Ismail Z, Aguera-Ortiz L, Brodaty H, et al. The Mild Behavioral Impairment Checklist (MBI-C): A Rating Scale for Neuropsychiatric Symptoms in Pre-Dementia Populations. J Alzheimers Dis 2017;56.
Tavakol M, Dennick R. Making sense of Cronbach's alpha. Int J Med Educ 2011;2.
Maruff P, Lim YY, Darby D, et al. Clinical utility of the cogstate brief battery in identifying cognitive impairment in mild cognitive impairment and Alzheimer’s disease. BMC Psychology 2013;1.
Portney LG, Watkins MP. Foundation of Clinical Research. Application to Practice. Appleton & Lange 2009.
Eraslan Boz H, Limoncu H, Zygouris S, et al. A new tool to assess amnestic mild cognitive impairment in Turkish older adults: virtual supermarket (VSM). Neuropsychol Dev Cogn B Aging Neuropsychol Cogn 2020;27.
Park JH, Jung M, Kim J, Park HY, Kim JR, Park JH. Validity of a novel computerized screening test system for mild cognitive impairment. Int Psychogeriatr 2018;30.
Sabbagh MN, Boada M, Borson S, et al. Early Detection of Mild Cognitive Impairment (MCI) in Primary Care. J Prev Alzheimers Dis 2020;7.
Grimes DA, Schulz KF. Refining clinical diagnosis with likelihood ratios. Lancet 2005;365.
Weintraub S, Besser L, Dodge HH, Teylan M, Morris JC. Version 3 of the Alzheimer Disease Centers' Neuropsychological Test Battery in the Uniform Data Set (UDS). Alzheimer Disease & Associated Disorders 2017;32.
Weiss LG, Saklofske DH, Coalson DL, Raiford SE. Theoretical, Empirical and Clinical Foundations of the WAIS-IV Index Scores. WAIS-IV Clinical Use and Interpretation 2010.
Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Chertkow H. The Montreal Cognitive Assessment, MoCA: A Brief Screening Tool For Mild Cognitive Impairment. Journal of the American Geriatrics Society 2005;53.

No competing interests reported.

Additionalfile1.docx
Additional file 1: Table S1 Composition of the expert panel; Table S2 Demographic data of phase 1 participants who are cognitively unimpaired and those with cognitive impairment; Table S3 Items optimization of initial CogSAS version; Table S4 The diagnostic accuracy of items in clinically diagnosed participants from phase 2; Table S5 The diagnostic accuracyof items in cognitive impairment due to AD from phase 3
Additionalfile2.pdf
Additional file 2: Material S1 results of the round 1 Delphi survey; Material S2 results of the round 2 Delphi survey; Material S3 results from the round 3 Delphi survey
Additionalfile3.pdf
Additional file 3: Pictures involve in Task 2
Additionalfile4.tif
Additional file 4:FIGURE S1Receiver operating characteristic curve of items comparing participants with CU and cognitively impaired patients due to AD in phase 2
Additionalfile5.tif
Additional file 5: FIGURE S2Receiver operating characteristic curve of items comparing participants with CU and cognitively impaired patients due to AD in phase 3

Download PDF

Version 1

posted

You are reading this latest preprint version

A mobile interactive cognitive self-assessment scale for screening Alzheimer’s disease

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

1 Background

2 Methods

2.1 Procedures

2.1.1 Scale development

2.1.1.1 Design of the initial CogSAS version

2.1.1.2 Item optimization

2.1.2 Scale validation in participants clinically diagnosed with cognitive impairment

2.1.3 Scale validation in participants with cognitive impairment due to AD

2.2 Statistical analysis

3 Results

3.1 Scale development

3.2 Scale validity and reliability analyses

3.2.1 Reliability analysis

3.2.2 Validity analysis

3.3 Sensitivity and specificity analyses

3.3.1 Sensitivity and specificity in clinically diagnosed participants

3.3.2 Sensitivity and specificity in assessing cognitive impairment due to AD

4 Discussion

5 Conclusion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1