Online Clinical Consensus Diagnosis of Dementia: Development and Validation

Assessing Cognitive Impairment in the Health and Retirement Study Harmonized Cognitive Assessment Protocol Project: Comparing a Diagnostic Algorithm With a Diagnostic Consensus Panel

BackgroundAccurate classification of cognitive impairment in population studies is challenging.ObjectiveTo compare the performance of a diagnostic algorithm with a clinical consensus panel.SampleIn 2016, the Health and Retirement Study (HRS) implemented the Harmonized Cognitive Assessment Protocol Project (HRS-HCAP) to streamline cognitive assessments for select HRS participants.MethodsThe Manly-Jones HCAP diagnostic classification was used to classify cognitive status as normal, mild cognitive impairment (MCI), or dementia. For this analysis, a consensus panel of five clinicians reviewed 50 cases with high diagnostic uncertainty, each reviewing 30 cases, blinded to the algorithm's classifications.AnalysisDiagnostic concordance was assessed using unweighted and weighted Cohen's kappa (κ).ResultsUnweighted concordance was 70% (35/50), with discordance mostly among MCI cases. Weighted concordance was 84%. Unweighted κ was 0.56 (95% CI 0.30-0.81) and weighted κ was 0.75 (95% CI 0.49-0.91), indicating moderate to substantial agreement between the two methods.

Design and methodology of the harmonized diagnostic assessment of dementia for the longitudinal aging study in India: Wave 2

The rising burden of dementia calls for high-quality data on cognitive decline and dementia onset. The second wave of the Harmonized Diagnostic Assessment for the Longitudinal Aging Study in India (LASI-DAD) was designed to provide longitudinal assessments of cognition and dementia in India. All Wave 1 participants were recruited for a follow-up interview, and a refresher sample was drawn from the Longitudinal Aging Study in India, a nationally representative cohort of Indians aged 45 and older. Respondents underwent a battery of cognitive tests, geriatric assessments, and venous blood collection. Their health and cognitive status were also assessed through an interview with a close family member or friend. Clinical consensus diagnosis was made based on the Clinical Dementia Rating®, and comprehensive data on risk factors of dementia were collected, including neurodegenerative biomarkers, sensory function, and environmental exposures. A total of 4635 participants were recruited between 2022 and 2024 from 22 states and union territories of India, accounting for 97.9% of the population in India. The response rate was 84.0%, and 71.5% of the participants provided venous blood specimen. LASI-DAD provides rich new data to study cognition, dementia, and their risk factors longitudinally in a nationally representative sample of older adults in India. Longitudinal cognitive data, together with longitudinally assessed biomarker data and novel data on sensory function and environmental exposures, provide a unique opportunity to establish associations between risk factors and biologically defined cognitive aging phenotypes.

The added value of metadata on test completion time for the quantification of cognitive functioning in survey research

OBJECTIVE

Information on the time spent completing cognitive testing is often collected, but such data are not typically considered when quantifying cognition in large-scale community-based surveys. We sought to evaluate the added value of timing data over and above traditional cognitive scores for the measurement of cognition in older adults.

METHOD

We used data from the Longitudinal Aging Study in India-Diagnostic Assessment of Dementia (LASI-DAD) study (N = 4,091), to assess the added value of timing data over and above traditional cognitive scores, using item-specific regression models for 36 cognitive test items. Models were adjusted for age, gender, interviewer, and item score.

RESULTS

Compared to Quintile 3 (median time), taking longer to complete specific items was associated (p < 0.05) with lower cognitive performance for 67% (Quintile 5) and 28% (Quintile 4) of items. Responding quickly (Quintile 1) was associated with higher cognitive performance for 25% of simpler items (e.g., orientation for year), but with lower cognitive functioning for 63% of items requiring higher-order processing (e.g., digit span test). Results were consistent in a range of different analyses adjusting for factors including education, hearing impairment, and language of administration and in models using splines rather than quintiles.

CONCLUSIONS

Response times from cognitive testing may contain important information on cognition not captured in traditional scoring. Incorporation of this information has the potential to improve existing estimates of cognitive functioning.

Dementia ascertainment in India and development of nation-specific cutoffs: A machine learning and diagnostic analysis

Introduction

Cognitive assessments are useful in ascertaining dementia but may be influenced by patient characteristics. India's distinct culture and demographics warrant investigation into population-specific cutoffs.

Methods

Data were utilized from the Longitudinal Aging Study in India-Diagnostic Assessment of Dementia (n = 2528). Dementia ascertainment was conducted by an online panel. A machine learning (ML) model was trained on these classifications, with explainable artificial intelligence to assess feature importance and inform cutoffs that were assessed across demographic groups.

Results

The Informant Questionnaire of Cognitive Decline in the Elderly (IQCODE) and Hindi Mini-Mental State Examination (HMSE) were identified as the most impactful assessments with optimal cutoffs of 3.8 and 25, respectively.

Discussion

An ML assessment of clinician dementia ratings identified IQCODE and HMSE to be the most impactful assessments. Optimal cutoffs of 3.8 and 25 were identified and performed excellently in the overall sample, though did decrease in specific, more difficult-to-diagnose subgroups.

Highlights

Pioneers use of explainable artificial intelligence in the diagnosis of dementia.Creates assessment cutoffs specific to the nation of India.Highlights differences in cutoffs across nations.

Insights from a 7-Year Dementia Cohort (VALCODIS): ApoE Genotype Evaluation

Background: The VALCODIS (Valencian Cognitive Diseases Study) cohort was designed and studied at the Hospital Universitari i Politècnic La Fe (Valencia, Spain) for the research of cognitive diseases, especially in the search for new biomarkers of Alzheimer's disease (AD). Methods: Participants in the VALCODIS cohort had cerebrospinal fluid (CSF) and blood samples, neuroimaging, and neuropsychological tests. The ApoE genotype was evaluated to identify its relationship with CSF biomarkers and neuropsychological tests in AD and non-AD participants. Results: A total of 1249 participants were included. They were mainly AD patients (n = 547) but also patients with other dementias (frontotemporal lobar dementia (n = 61), Lewy body dementia without AD CSF signature (n = 10), vascular dementia (n = 24) and other specific causes of cognitive impairment (n = 442), and patients with subjective memory complaints (n = 165)). In the ApoE genotype evaluation, significant differences were found for Aβ42 levels between genotypes in both AD and non-AD patients, as well as a negative correlation between tau values and a cognitive test in non-carriers and ε4 heterozygous. Conclusions: The VALCODIS cohort provides biologically diagnosed patients with demographical, clinical and biochemical data, and biological samples for further studies on early AD diagnosis. Also, the ApoE genotype evaluation showed correlations between CSF biomarkers and neuropsychological tests.

Considerations for the use of the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) in cross-country comparisons of cognitive aging and dementia

INTRODUCTION

Informant reports are a critical component of dementia diagnoses, but the comparability of informant reports across countries is not well understood.

METHODS

We compared the performance of the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) using population-representative surveys in the United States (N = 3183), England (N = 1050), and India (N = 4047).

RESULTS

Analyses of regression splines and comparisons of model fit showed strong associations between IQCODE and objective cognition at low cognitive functioning in the United States and England; in India, the association was weaker but consistent over the range of cognition. Associations between IQCODE score and informant generation (analysis of variance [ANOVA] p = 0.001), caregiver status (p < 0.001), and years known by the informant (p = 0.015) were different across countries after adjusting for objective cognition.

DISCUSSION

In India, the IQCODE was less sensitive to impairments at the lowest levels of cognitive functioning. Country-specific adjustments to IQCODE scoring based on informant characteristics may improve cross-national comparisons.

HIGHLIGHTS

Associations between IQCODE and cognitive testing were similar in the United States and England but differed in India. In India, the IQCODE may be less sensitive to impairments among those with low cognition and no education. Informant characteristics may differentially impact informant reports of decline across countries. Adjustments or culturally sensitive adaptations may improve cross-national comparability.

Mental Health Impacts of COVID-19: Does Prepandemic Cognition and Dementia Status Matter?

BACKGROUND

The coronavirus disease 2019 (COVID-19) pandemic disrupted daily life and led to sharp shocks in trends for various health outcomes. Although substantial evidence exists linking the pandemic and mental health outcomes and linking dementia and mental health outcomes, little evidence exists on how cognitive status may alter the impact of COVID-19 on mental health.

METHODS

We used prepandemic data from the Longitudinal Aging Study in India-Diagnostic Assessment of Dementia study and 9 waves of data from the Real-Time Insights of COVID-19 in India study (N = 1 182). We estimated associations between measures of prepandemic cognition (continuous cognition based on 22 cognitive tests, dementia status) and mental health measures during the pandemic (Patient Health Questionnaire [PHQ]-4 [9 time points], PHQ-9 [2 time points], Beck Anxiety Inventory [3 time points]), adjusting for age, gender, rural/urban residence, state, education, and prepandemic mental health.

RESULTS

Summarizing across time points, PHQ-9 score was marginally or significantly associated with prepandemic cognition (PHQ-9 difference: -0.38 [-0.78 to 0.14] points per SD higher cognition; p = .06), and prepandemic dementia (PHQ-9 difference: 0.61 [0.11-1.13] points for those with dementia compared to no dementia; p = .02). Associations with BAI were null, whereas associations with PHQ-4 varied over time (p value for interaction = .02) and were strongest during the delta wave, when pandemic burden was highest.

CONCLUSIONS

We present initial evidence that mental health impacts of COVID-19 or other acute stressors may be unequally distributed across strata of cognitive outcomes. In dynamically changing environments, those with cognitive impairment or dementia may be more vulnerable to adverse mental health outcomes.

Prevalence of dementia in India: National and state estimates from a nationwide study

INTRODUCTION

Prior estimates of dementia prevalence in India were based on samples from selected communities, inadequately representing the national and state populations.

METHODS

From the Longitudinal Aging Study in India (LASI) we recruited a sample of adults ages 60+ and administered a rich battery of neuropsychological tests and an informant interview in 2018 through 2020. We obtained a clinical consensus rating of dementia status for a subsample (N = 2528), fitted a logistic model for dementia status on this subsample, and then imputed dementia status for all other LASI respondents aged 60+ (N = 28,949).

RESULTS

The estimated dementia prevalence for adults ages 60+ in India is 7.4%, with significant age and education gradients, sex and urban/rural differences, and cross-state variation.

DISCUSSION

An estimated 8.8 million Indians older than 60 years have dementia. The burden of dementia cases is unevenly distributed across states and subpopulations and may therefore require different levels of local planning and support.

HIGHLIGHTS

The estimated dementia prevalence for adults ages 60+ in India is 7.4%. About 8.8 million Indians older than 60 years live with dementia. Dementia is more prevalent among females than males and in rural than urban areas. Significant cross-state variation exists in dementia prevalence.

Deep phenotyping and genomic data from a nationally representative study on dementia in India

The Harmonized Diagnostic Assessment of Dementia for the Longitudinal Aging Study in India (LASI-DAD) is a nationally representative in-depth study of cognitive aging and dementia. We present a publicly available dataset of harmonized cognitive measures of 4,096 adults 60 years of age and older in India, collected across 18 states and union territories. Blood samples were obtained to carry out whole blood and serum-based assays. Results are included in a venous blood specimen datafile that can be linked to the Harmonized LASI-DAD dataset. A global screening array of 960 LASI-DAD respondents is also publicly available for download, in addition to neuroimaging data on 137 LASI-DAD participants. Altogether, these datasets provide comprehensive information on older adults in India that allow researchers to further understand risk factors associated with cognitive impairment and dementia.

Estimating the Prevalence of Dementia in India Using a Semi-Supervised Machine Learning Approach

INTRODUCTION

Accurate estimation of dementia prevalence is essential for making effective public and social care policy to support individuals and families suffering from the disease. The purpose of this paper is to estimate the prevalence of dementia in India using a semi-supervised machine learning approach based on a large nationally representative sample.

METHODS

The sample of this study is adults 60 years or older in the wave 1 (2017-2019) of the Longitudinal Aging Study in India (LASI). A subsample in LASI received extensive cognitive assessment and clinical consensus ratings and therefore has diagnoses of dementia. A semi-supervised machine learning model was developed to predict the status of dementia for LASI participants without diagnoses. After obtaining the predictions, sampling weights and age standardization to the World Health Organization (WHO) standard population were applied to generate the estimate for prevalence of dementia in India.

RESULTS

The prevalence of dementia for those aged 60 years and older in India was 8.44% (95% CI: 7.89%-9.01%). The age-standardized prevalence was estimated to be 8.94% (95% CI: 8.36%-9.55%). The prevalence of dementia was greater for those who were older, were females, received no education, and lived in rural areas.

DISCUSSION

The prevalence of dementia in India may be higher than prior estimates derived from local studies. These prevalence estimates provide the information necessary for making long-term planning of public and social care policy. The semi-supervised machine learning approach adopted in this paper may also be useful for other large population aging studies that have a similar data structure.

Future Directions for the HRS Harmonized Cognitive Assessment Protocol

In the absence of effective pharmacological treatment to halt or reverse the course of Alzheimer's disease and related dementias (ADRDs), population-level research on the modifiable determinants of dementia risk and outcomes for those living with ADRD is critical. The Harmonized Cognitive Assessment Protocol (HCAP), fielded in 2016 as part of the U.S. Health and Retirement Study (HRS) and multiple international counterparts, has the potential to play an important role in such efforts. The stated goals of the HCAP are to improve our ability to understand the determinants, prevalence, costs, and consequences of cognitive impairment and dementia in the U.S. and to support cross-national comparisons. The first wave of the HCAP demonstrated the feasibility and value of the more detailed cognitive assessments in the HCAP compared to the brief cognitive assessments in the core HRS interviews. To achieve its full potential, we provide eight recommendations for improving future iterations of the HCAP. Our highest priority recommendation is to increase the representation of historically marginalized racial/ethnic groups disproportionately affected by ADRDs. Additional recommendations relate to the timing of the HCAP assessments; clinical and biomarker validation data, including to improve cross-national comparisons; dropping lower performing items; enhanced documentation; and the addition of measures related to caregiver impact. We believe that the capacity of the HCAP to achieve its stated goals will be greatly enhanced by considering these changes and additions.

A guide for researchers seeking training in retrospective data harmonization for population neuroscience studies of Alzheimer's disease and related dementias

Due to needs surrounding rigor and reproducibility, subgroup specific disease knowledge, and questions of external validity, data harmonization is an essential tool in population neuroscience of Alzheimer's disease and related dementias (ADRD). Systematic harmonization of data elements is necessary to pool information from heterogeneous samples, and such pooling allows more expansive evaluations of health disparities, more precise effect estimates, and more opportunities to discover effective prevention or treatment strategies. The key goal of this Tutorial in Population Neuroimaging Curriculum, Instruction, and Pedagogy article is to guide researchers in creating a customized population neuroscience of ADRD harmonization training plan to fit their needs or those of their mentees. We provide brief guidance for retrospective data harmonization of multiple data types in this area, including: (1) clinical and demographic, (2) neuropsychological, and (3) neuroimaging data. Core competencies and skills are reviewed, and resources are provided to fill gaps in training as well as data needs. We close with an example study in which harmonization is a critical tool. While several aspects of this tutorial focus specifically on ADRD, the concepts and resources are likely to benefit population neuroscientists working in a range of research areas.

Learning From Clinical Consensus Diagnosis in India to Facilitate Automatic Classification of Dementia: Machine Learning Study

BACKGROUND

The Harmonized Diagnostic Assessment of Dementia for the Longitudinal Aging Study in India (LASI-DAD) is the first and only nationally representative study on late-life cognition and dementia in India (n=4096). LASI-DAD obtained clinical consensus diagnosis of dementia for a subsample of 2528 respondents.

OBJECTIVE

This study develops a machine learning model that uses data from the clinical consensus diagnosis in LASI-DAD to support the classification of dementia status.

METHODS

Clinicians were presented with the extensive data collected from LASI-DAD, including sociodemographic information and health history of respondents, results from the screening tests of cognitive status, and information obtained from informant interviews. Based on the Clinical Dementia Rating (CDR) and using an online platform, clinicians individually evaluated each case and then reached a consensus diagnosis. A 2-step procedure was implemented to train several candidate machine learning models, which were evaluated using a separate test set for predictive accuracy measurement, including the area under receiver operating curve (AUROC), accuracy, sensitivity, specificity, precision, F1 score, and kappa statistic. The ultimate model was selected based on overall agreement as measured by kappa. We further examined the overall accuracy and agreement with the final consensus diagnoses between the selected machine learning model and individual clinicians who participated in the clinical consensus diagnostic process. Finally, we applied the selected model to a subgroup of LASI-DAD participants for whom the clinical consensus diagnosis was not obtained to predict their dementia status.

RESULTS

Among the 2528 individuals who received clinical consensus diagnosis, 192 (6.7% after adjusting for sampling weight) were diagnosed with dementia. All candidate machine learning models achieved outstanding discriminative ability, as indicated by AUROC >.90, and had similar accuracy and specificity (both around 0.95). The support vector machine model outperformed other models with the highest sensitivity (0.81), F1 score (0.72), and kappa (.70, indicating substantial agreement) and the second highest precision (0.65). As a result, the support vector machine was selected as the ultimate model. Further examination revealed that overall accuracy and agreement were similar between the selected model and individual clinicians. Application of the prediction model on 1568 individuals without clinical consensus diagnosis classified 127 individuals as living with dementia. After applying sampling weight, we can estimate the prevalence of dementia in the population as 7.4%.

CONCLUSIONS

The selected machine learning model has outstanding discriminative ability and substantial agreement with a clinical consensus diagnosis of dementia. The model can serve as a computer model of the clinical knowledge and experience encoded in the clinical consensus diagnostic process and has many potential applications, including predicting missed dementia diagnoses and serving as a clinical decision support tool or virtual rater to assist diagnosis of dementia.

Design and Methodology of the Longitudinal Aging Study in India-Diagnostic Assessment of Dementia (LASI-DAD)

BACKGROUND/OBJECTIVES

To provide high-quality data on older adults in India that will enable an in-depth study of late-life cognition and dementia in India and cross-country analysis of risk factors for cognitive aging and dementia.

DESIGN

The Longitudinal Aging Study in India (LASI) is a nationally representative survey of health, economic, and social well-being of the Indian population aged 45 and older. Its large sample of more than 70,000 older adults represents not only the country as a whole but also each state. LASI-Diagnostic Assessment of Dementia (DAD) is an in-depth study of late-life cognition and dementia, drawing a subsample of over 3,000 LASI respondents aged 60 and older.

SETTING

Participants were interviewed at home or in a participating hospital according to their preferences.

PARTICIPANTS

Adults aged 60 and older (N = 3,224), along with 3,191 informants.

MEASUREMENTS

Respondents underwent a battery of cognitive tests, and informants were interviewed about their cognitive and health conditions. A common set of cognitive tests was selected to enable international comparisons, and additional cognitive tests suitable for illiterate and innumerate populations were also selected. Rich data on risk factors of dementia were collected through health examination, venous blood assays, and genotyping.

RESULTS

The response rate was 82.9%, varying across sex, education, and urbanicity. Data are available to other researchers.

CONCLUSION

LASI-DAD provides an opportunity to study late-life cognition and dementia and their risk factors in the older population in India and to gain further insights through cross-country analysis by pooling data from its international sister studies. J Am Geriatr Soc 68:S5-S10, 2020.