Self-reported Dementia Research Articles

The interaction between Indigenous identity and rural residency in dementia prevalence among Ecuadorian older adults.

While recognizing the acknowledged difference in dementia prevalence between Indigenous and non-Indigenous populations, most previous studies were conducted in the Global North. This study aims to examine the relationship between Indigenous ethnic identity and dementia status with a special focus on urban-rural differences among older adults in Ecuador. Data came from Ecuador's Survey of Health, Welfare, and Aging (SABE), derived from a probability sample of households in Ecuador. The final sample comprised 4,984 adults aged 60 or older. Dementia status was assessed through three indicators: low scores on a cognitive assessment, the number of difficulties in instrumental activities of daily living (IADL), and self-reported dementia diagnosis. Logistic regression and multivariate decomposition analyses were utilized. Indigenous participants, even after adjusting for sociodemographic and health-related factors, exhibited a higher likelihood of having dementia compared to their non-Indigenous counterparts. Among Indigenous participants, those residing in rural areas had higher likelihood of having dementia status, while there was no significant urban-rural difference observed among non-Indigenous participants. Rural residency and covariates explain 91.0% of the observed difference in dementia prevalence between Indigenous and non-Indigenous older adults. Indigenous older adults who are living in rural areas are at particularly high risk of having dementia largely because they have presently recognized modifiable risk factors. These findings underscore the urgent need to prioritize provisions of appropriate and equitable service related to dementia for Indigenous People living in rural areas outside high-income countries.

The Associations of Sensory Impairment With 10-Year Risk of Dementia and Alzheimer's Disease: The Health and Retirement Study, 2010-2020.

Studies have examined the association between dual sensory impairment and late-life cognitive outcomes in the U.S with inconsistent findings. To examine the associations between sensory impairment and 10-year risk of dementia or Alzheimer's disease among U.S. adults aged ≥ 50. A prospective cohort study based on the Health and Retirement Study from 2010 to 2020. Individuals aged ≥ 50 years without self-reported dementia and Alzheimer's disease in 2010 were included in the analysis. Self-reported visual and hearing impairments were measures in 2010. Main failure events included self-reported incident dementia and Alzheimer's disease over a 10-year follow-up period. Participants were categorized as having no visual or hearing impairment, visual impairment only, hearing impairment only, and dual sensory impairment. Fine-Gray competing risk regression model was applied to estimate the associations of sensory impairment with incident dementia and Alzheimer's disease, adjusted for demographic characteristics, health behaviors, and health conditions at baseline. Of 20,248 identified individuals, 14.6% had visual impairment only, 11.2% had hearing impairment only, and 9.1% had dual impairment at baseline. After adjusting for all covariates, dual sensory impairment was associated with higher risk of dementia (HR = 1.46, 95% CI: 1.23-1.73) and Alzheimer's disease (HR = 1.35, 95% CI: 1.03-1.76). Visual impairment only was also associated with incident dementia and Alzheimer's disease among individuals <65 years. Older adults in the U.S. with visual and hearing impairments simultaneously had a particularly greater risk of dementia and Alzheimer's disease, indicating the needs of targeted screening for timely treatment and further prevention of dementia and Alzheimer's disease.

Greater baseline intra-individual variation in telephone-based cognitive screening predicts cognitive and diagnostic outcomes at 2-year follow-up.

Intra-Individual Cognitive Variability (IICV) is an emerging clinical tool that has shown promise in predicting cognitive decline and dementia incidence. The present study aims to assess the predictive validity of IICV in remote cognitive screening tests, using nationally representative data. Two waves of cognitive and diagnostic data from the Health and Retirement Study (collected in 2010 and 2012) were utilized to investigate whether baseline IICV can predict cognitive decline and dementia pathology. Middle-aged and older adults who were cognitively intact and completed all cognitive tests at both baseline and follow-up were recruited in the study, resulting in a sample of 6,050 participants. With the coefficient of variation method, the IICV-dispersion was calculated based on cognitive screeners to predict follow-up mean cognitive performance, global cognition, suspected cognitive impairment, and self-reported dementia diagnosis. After accounting for demographics, depressive symptoms, and baseline cognitive performance, the results provide support for the predictive validity of IICV. Specifically, the study demonstrated that IICV-dispersion significantly predicted cognitive and diagnostic outcomes in a concave pattern where the prediction was more sensitive toward the higher end of IICV. IICV explained about 0.2%-2.3% of the variance of outcomes variables. IICV retrieved from cognitive screening tests in telemedicine settings offers insight into future cognitive functioning and neurocognitive diagnostic status, which can be cost-effective and reduce the burden on both patients and health care providers, especially benefitting individuals with low socioeconomic status and rural residents. Potential avenues for future research were also discussed. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

The AlzMatch Pilot Study - Feasibility of Remote Blood Collection of Plasma Biomarkers for Preclinical Alzheimer’s Disease Trials

BackgroundAdvances in plasma biomarkers to detect Alzheimer’s disease (AD) biology allows researchers to improve the efficiency of participant recruitment into preclinical trials. Recently, protein levels of plasma amyloid-beta and tau proteins have been shown to be predictive of elevated amyloid in brain. Online registries, such as the Alzheimer’s Prevention Trials (APT) Webstudy, include and follow participants using remote assessments to facilitate efficient screening and enrollment of large numbers of individuals who may be at higher risk for AD.ObjectivesThe AlzMatch Pilot Study investigated the feasibility of recruiting individuals from an online registry for blood sample collection at community-based phlebotomy centers and plasma biomarker quantification to assess an individual’s eligibility for AD preclinical trials.DesignPilot feasibility study with co-primary outcomes.SettingThis pilot feasibility study included participants from the APT Webstudy, the remote assessment arm of the Trial-ready cohort for Preclinical and Prodromal AD (TRC-PAD) Platform. Novel design included collection of electronic consent, use of community laboratories for plasma collection, mass spectrometry-based biomarker assay, and telephone communication of plasma biomarker screening eligibility.ParticipantsParticipants invited to the AlzMatch pilot feasibility study were active in the APT Webstudy, 50 years of age or older, resided within 50 miles of both a Quest Diagnostics Patient Services Center (a national diagnostic laboratory with convenient locations for sample collection and processing) and one of six TRC-PAD vanguard clinical trial sites, had no self-reported dementia diagnosis, were able to communicate in English and engaged with the APT Webstudy within the prior 6 months.MeasurementsPrimary feasibility outcomes were completion of electronic consent (e-consent) for invited participants and collection of usable blood samples. Additional feasibility outcomes included invitation response rate, plasma biomarker eligibility status (based on amyloid beta-42/40 [Aβ42/40] concentration ratio), ApoE proteotype, and trial inclusion criterion), and completion of telephone contact to learn eligibility to screen for a study.Results300 APT Webstudy participants were invited to consent to the AlzMatch study. The AlzMatch e-consent rate was 39% (n=117) (95% CI of 33.5%–44.5%) overall, which was higher than the expected rate of 25%. Similar consent rates were observed across participants based on self-defined sex (41% Female (n=75), 37% Male (n=42)) and race and ethnicity (37% from underrepresented groups (URG) (n=36), 40% not from URG (n=79)). Among those that consented (n=117), plasma was successfully collected from 74% (n=87) (95% CI of 66%–82%), with similar rates across sex (76% Female (n=57), 71% Male (n=30)) and race and ethnicity (75% URG (n=27) and 75% not from URG (n=59)). 60% (n=51) of participants with plasma biomarker results were eligible to screen for future preclinical AD trials.ConclusionElectronic consent of participants through an online registry, blood sample collection at community-based centers, plasma biomarker quantification and reporting, and biomarker assessments for study eligibility were all feasible with similar engagement rates across demographic groups. Although this pilot was a small and selective sample, participants engaged and consented at higher than expected rates. We conclude that collecting blood at community laboratories for biomarker analyses may improve accessibility beyond research, and may facilitate broader access for clinical use of AD plasma biomarkers. Based on our results, an expanded version of the AlzMatch study is underway, which involves expanding invitations to additional APT Webstudy participants and clinical trial sites.

Race/ethnicity and the measurement of cognition in NSHAP: Recommendations for robustness.

In this study, we examine the measurement of cognition in different racial/ethnic groups to move towards a less biased and more inclusive set of measures for capturing cognitive change and decline in older adulthood. We use data from Round 2 (N=3377) and Round 3 (N=4777) of the National Social Life, Health, and Aging Project (NSHAP) and examine the study's Survey Adjusted version of the Montreal Cognitive Assessment (MoCA-SA). We employ exploratory factor analyses to explore configural invariance by racial/ethnic group. Using modification indexes, two-parameter item response theory models, and split-sample testing, we identify items that seem robust to bias by race. We test the predictive validity of the full (18-item) and short (4-item) MoCA-SAs using self-reported dementia diagnosis, instrumental activities of daily living (IADLs), proxy reports of dementia, proxy reports of dementia-related death, and National Death Index reports of dementia-related death. We found that four measures out of the 18 used in NSHAP's MoCA-SA formed a scale that was more robust to racial bias. The shortened form predicted consequential outcomes as well as NSHAP's full MoCA-SA. The short form was also moderately correlated with the full form. Although sophisticated structural equation modeling techniques would be preferrable for assuaging measurement invariance by race in NSHAP, the shortened form of the MoCA-SA provides a quick way for researchers to carry out robustness checks and to see if the disparities and associations by race they document are "real" or the product of artifactual bias.

Path integration deficits are associated with phosphorylated tau accumulation in the entorhinal cortex.

Alzheimer's disease is a devastating disease that is accompanied by dementia, and its incidence increases with age. However, no interventions have exhibited clear therapeutic effects. We aimed to develop and characterize behavioural tasks that allow the earlier identification of signs preceding dementia that would facilitate the development of preventative and therapeutic interventions for Alzheimer's disease. To this end, we developed a 3D virtual reality task sensitive to the activity of grid cells in the entorhinal cortex, which is the region that first exhibits neurofibrillary tangles in Alzheimer's disease. We investigated path integration (assessed by error distance) in a spatial navigation task sensitive to grid cells in the entorhinal cortex in 177 volunteers, aged 20-89 years, who did not have self-reported dementia. While place memory was intact even in old age, path integration deteriorated with increasing age. To investigate the relationship between neurofibrillary tangles in the entorhinal cortex and path integration deficit, we examined a mouse model of tauopathy (P301S mutant tau-overexpressing mice; PS19 mice). At 6 months of age, PS19 mice showed a significant accumulation of phosphorylated tau only in the entorhinal cortex, associated with impaired path integration without impairments in spatial cognition. These data are consistent with the idea that path integration deficit is caused by the accumulation of phosphorylated tau in the entorhinal cortex. This method may allow the early identification of individuals likely to develop Alzheimer's disease.

RACE/ETHNICITY AND THE MEASUREMENT OF COGNITION IN NSHAP: RECOMMENDATIONS FOR ROBUSTNESS

Abstract In this study, we interrogate measurement of cognition by race, in order to move towards a less-biased and more-inclusive set of measures for capturing cognitive change and decline in older adulthood. We use data from Round 2 (N=3377) and Round 3 (N=4777) of the National Social Life Health and Aging Project (NSHAP), and examine the study’s version of the Montreal Cognitive Assessment (MoCA). We employ exploratory factor analyses to explore configural invariance by racial/ethnic group (Non-Hispanic White, Non-Hispanic Black, Hispanic, All else), and then log-likelihood tests of scalar and metric invariance. Using modification indexes we identify items that seem robust to bias by race. We test the predictive validity of the full (18-item) and short (7-item) scales using self-reported dementia diagnosis, Instrumental Activities of Daily Living (IADLs), proxy reports of dementia death, and National Death Index (NDI) reports of dementia death. We found that 7 measures, out of the 18 used in NSHAP’s MoCA, formed a scale that was more robust to racial bias. The shortened form predicted consequential outcomes equally well, compared to NSHAP’s full MoCA. The short form was also highly correlated with the full form, and displayed lower test-retest correlation between Round 2 and 3. Although sophisticated structural equation modeling techniques could be useful for assuaging measurement invariance by race in NSHAP, the shortened form provides a quick way for researchers to carry out robustness checks – to see if the disparities and associations they document are ‘real’ or the product of artifactual bias.

Home-Delivered Meals and Nursing Home Placement Among People With Self-Reported Dementia

Home-delivered meals promote food security and independence among homebound older adults. However, it is unclear which of the 2 predominant modes of meal delivery, daily-delivered vs mailed (or drop-shipped) frozen meals, promotes community living for homebound older adults with dementia. To assess the risk of nursing home admission within 6 months between homebound individuals receiving daily-delivered vs drop-shipped frozen meals. This pilot, multisite, 2-arm, pragmatic clinical trial included older adults with self-reported dementia on waiting lists for meals at 3 Meals on Wheels (MOW) programs in Texas and Florida between April 7 and October 8, 2021, to assess time to nursing home placement. Participants were randomized to receive either meals delivered by an MOW driver or frozen meals that were mailed to participants' homes every 2 weeks. Participants received their assigned intervention for up to 6 months. The primary study outcome was days from randomization to a Minimum Data Set nursing home admission assessment within 6 months. Feasibility of conducting this type of study was examined by tracking enrollment, examining baseline characteristics, monitoring participants' intervention fidelity, measuring the proportion of participants linked with Centers for Medicare & Medicaid Services (CMS) data, and analyzing the primary study outcome. Among 325 eligible participants who were randomized, 243 enrolled in the study (mean [SD] age, 81 [8.0] years; 152 (62.6%) were female): 128 to the daily-delivered meals group and 115 to the drop-shipped frozen meals group; 119 participants (49.0%) lived alone. Among the total participants enrolled, 227 (93.4%) were linked deterministically to their CMS data; probabilistic methods were used to link the remaining 16 participants (6.6%). At 6 months from randomization, 160 participants (65.8%) were still receiving meals, and 25 (10.1%; 95% CI, 6.3%-14.0%) were admitted to a nursing home. After adjusting for sex, race and ethnicity, age, program, and living arrangement and the use of death as a censoring event, the adjusted log hazard ratio of nursing home placement between daily-delivered and drop-shipped frozen meals was -0.67 (95% CI, -1.52 to 0.19). This pilot randomized clinical trial demonstrated the feasibility of enrolling participants with self-reported dementia on waiting lists at MOW programs, linking their data, and evaluating outcomes. While this pilot study was not powered to detect meaningful, statistically significant differences in nursing home placement, its feasibility and initial results warrant exploration in a follow-on, adequately powered trial. ClinicalTrials.gov Identifier: NCT04850781.

Identification of high likelihood of dementia in population-based surveys using unsupervised clustering: a longitudinal analysis

BackgroundDementia is defined as a cognitive decline that affects functional status. Longitudinal ageing surveys often lack a clinical diagnosis of dementia though measure cognition and daily function over time. We used unsupervised machine learning and longitudinal data to identify transition to probable dementia.MethodsMultiple Factor Analysis was applied to longitudinal function and cognitive data of 15,278 baseline participants (aged 50 years and more) from the Survey of Health, Ageing, and Retirement in Europe (SHARE) (waves 1, 2 and 4–7, between 2004 and 2017). Hierarchical Clustering on Principal Components discriminated three clusters at each wave. We estimated probable or “Likely Dementia” prevalence by sex and age, and assessed whether dementia risk factors increased the risk of being assigned probable dementia status using multistate models. Next, we compared the “Likely Dementia” cluster with self-reported dementia status and replicated our findings in the English Longitudinal Study of Ageing (ELSA) cohort (waves 1–9, between 2002 and 2019, 7840 participants at baseline).ResultsOur algorithm identified a higher number of probable dementia cases compared with self-reported cases and showed good discriminative power across all waves (AUC ranged from 0.754 [0.722–0.787] to 0.830 [0.800–0.861]). “Likely Dementia” status was more prevalent in older people, displayed a 2:1 female/male ratio, and was associated with nine factors that increased risk of transition to dementia: low education, hearing loss, hypertension, drinking, smoking, depression, social isolation, physical inactivity, diabetes, and obesity. Results were replicated in ELSA cohort with good accuracy.ConclusionsMachine learning clustering can be used to study dementia determinants and outcomes in longitudinal population ageing surveys in which dementia clinical diagnosis is lacking.

Prevalence of dementia, heart disease and stroke in community-dwelling adults in Canada, 2016–2021: opportunities for joint prevention

IntroductionThis aim of this study is to provide updated estimates on the prevalence of dementia, heart disease, and stroke in Canadian communities. Targeting all three conditions together, at the community level, may be key to disease prevention and health aging in the Canadian population.MethodsUsing nationwide health survey data, we calculated the age-standardized prevalence of self-reported dementia, heart disease and stroke in adults aged 18 years and over residing in Canadian communities from 2016 to 2021. Poisson regression models were used to detect statistically significant changes in the prevalence of all three conditions from 2016 to 2021.ResultsLess than 1% (~ 175,000 individuals) of adults residing in Canadian communities reported dementia, 5% (~ 1.5 million individuals) reported heart disease, and more than 1% (~ 370,000 individuals) reported stroke annually from 2016 to 2021. Overall, the age-standardized prevalence for stroke decreased minimally from 2016 to 2021 (p = 0.0004). Although the age-standardized prevalence of heart disease and dementia decreased from 2016 to 2018, subsequent increases in prevalence from 2018 to 2021 led to a lack of overall statistically significant changes from 2016 to 2021 (p = 0.10 for heart disease and p = 0.37 for dementia).ConclusionRecent increases in the prevalence of dementia, heart disease and stroke in Canadian communities threaten to reverse any gains in vascular disease prevention over the past six years. Findings reveal the urgent need for intensified prevention efforts that are community-based with a focus on joint reduction in the shared risk factors contributing to all three diseases.

Cognitive Impairment and Self-Reported Dementia in UK Retired Professional Soccer Players: A Cross Sectional Comparative Study

BackgroundPrevious studies based on death certificates have found professional soccer players were more likely to die with neurodegenerative diseases, including dementia. Therefore, this study aimed to investigate whether retired professional male soccer players would perform worse on cognitive tests and be more likely to self-report dementia diagnosis than general population control men.MethodsA cross-sectional comparative study was conducted between August 2020 and October 2021 in the United Kingdom (UK). Professional soccer players were recruited through different soccer clubs in England, and general population control men were recruited from the East Midlands in the UK. We obtained self-reported postal questionnaire data on dementia and other neurodegenerative diseases, comorbidities and risk factors from 468 soccer players and 619 general population controls. Of these, 326 soccer players and 395 general population controls underwent telephone assessment for cognitive function.ResultsRetired soccer players were approximately twice as likely to score below established dementia screening cut-off scores on the Hopkins Verbal Learning Test (OR 2.06, 95%CI 1.11–3.83) and Verbal Fluency (OR 1.78, 95% CI 1.18–2.68), but not the Test Your Memory, modified Telephone Interview for Cognitive Status, and Instrumental Activities of Daily Living. Analyses were adjusted for age, education, hearing loss, body mass index, stroke, circulatory problems in the legs and concussion. While retired soccer players were younger, had fewer cardiovascular diseases and other morbidities and reported healthier lifestyles, 2.8% of retired soccer players reported medically diagnosed dementia and other neurodegenerative disease compared to 0.9% of controls (OR = 3.46, 95% CI 1.25–9.63) after adjustment for age and possible confounders.ConclusionsUK male retired soccer players had a higher risk of performing below established cut-off scores of dementia screening tests and were more likely to self-report medically diagnosed dementia and neurodegenerative diseases, despite having better overall physical health and fewer dementia risk factors. Further study is needed to determine specific soccer-related risk factors.

Identifying dementia using medical data linkage in a longitudinal cohort study: Lothian Birth Cohort 1936

BackgroundThe Lothian Birth Cohort 1936 (LBC1936) is a longitudinal study of ageing with well-characterised assessments, but until now, it has relied on self-report or proxies for dementia such as cognitive tests. Our aims were twofold:a) to describe a framework for identifying dementia in a cohort study.b) to report the age-specific incidence and prevalence of all-cause dementia and dementia subtypes in 865 individuals in the LBC1936.MethodsElectronic Health Records (EHR) of all participants were reviewed, and relevant information was extracted to form case vignettes for everyone with any record of cognitive dysfunction. The EHR data sources include hospital and clinic letters, general practitioner and hospital referrals, prescribed medications, imaging and laboratory results. Death certificate data were obtained separately. Clinician assessments were performed when there was concern about a participant's cognition. A diagnosis of probable dementia, possible dementia, or no dementia was agreed upon by a consensus diagnostic review board, comprised of a multidisciplinary team of clinical dementia experts who reviewed case vignettes and clinician assessment letters. For those with probable dementia, a subtype was also determined, where possible. We compared the agreement between our newly ascertained dementia diagnoses with the existing self-reported dementia diagnoses.ResultsSelf-reported dementia diagnoses were positive in only 17.8% of ascertained dementia diagnoses. The EHR review identified 163/865 (18.8%) individuals as having cognitive dysfunction. At the consensus diagnostic review board, 118/163 were diagnosed with probable all-cause dementia, a prevalence of 13.6%. Age-specific dementia prevalence increased with age from 0.8% (65–74.9 years) to 9.93% (85–89.9 years). Prevalence rates for women were higher in nearly all age groups. The most common subtype was dementia due to Alzheimer disease (49.2%), followed by mixed Alzheimer and cerebrovascular disease (17.0%), dementia of unknown or unspecified cause (16.1%), and dementia due to vascular disease (8.5%).ConclusionsWe present a robust systematic framework and guide for other cohort teams wanting to ascertain dementia diagnoses. The newly ascertained dementia diagnosis provides vital data for further analyses of LBC1936 to allow exploration of lifecourse predictors of dementia.

Comparing Dementia Classification by Self-Report and Administrative Records in the National Core Indicators-Aging and Disability Survey: A Predictive Modeling Approach.

Policymakers are interested in the long-term services and supports (LTSS) needs of people living with dementia. The National Core Indicators-Aging and Disability (NCI-AD) survey is conducted to evaluate LTSS care needs. However, dementia reporting in NCI-AD varies across states, and is either obtained from state administrative records or self-reported during the survey. We explored the implications of identifying dementia from administrative records versus self-report. We analyzed 24,569 NCI-AD respondents age 65+, of which 22.4% had dementia. To assess dementia accuracy by data source, we fit separate logistic regression models using the administrative and self-reported subsamples. We applied model coefficients to the population whose dementia status came from the opposite source. Using the administrative model to predict self-reported dementia resulted in higher sensitivity than using the self-report model to predict administrative dementia (43.8% vs. 37.9%). The self-report model's diminished sensitivity suggests administrative records may capture cases of dementia missed by self-report.

ASSESSING DEMENTIA CLASSIFICATION IN THE NATIONAL CORE INDICATORS AGING AND DISABILITY SURVEY

Abstract Improving the quality of long-term care services for older adults is a national priority. The National Core Indicators Aging and Disability (NCI-AD) collects data from multiple states to evaluate long-term care recipients’ service satisfaction and needs. Many analysts are interested in exploring the service environments of people living with dementia. However, dementia measurement and reporting varies in NCI-AD between states. Dementia status may be obtained from administrative records or self-reported during the survey. We explored the measurement of dementia in NCI-AD and the implications of relying on administrative or self-reported dementia status. We analyzed NCI-AD data from 2015-2018 representing 24,569 respondents age 65+, of which 5,502 (22.4%) were identified as having dementia. 42.9% of respondents had dementia status determined administratively and 57.1% had it determined during the survey. To assess dementia accuracy by data source, we fit separate LASSO models for both the administrative and survey subpopulations predicting dementia status using demographic and functional predictors. We then used each model to predict dementia status in the subpopulation with discordant dementia data source. Using the administrative model to predict survey reported status resulted in a higher sensitivity than using the survey model to predict administrative status (44.6% vs 32.2%). The diminished predictive accuracy of the survey model suggests administrative records may capture cases of dementia diagnosis that would be missed by self-report. These findings highlight analytical caveats for researchers interested study long-term care quality for people with dementia using NCI-AD, and emphasize the importance of moving towards more standardized dementia reporting.

ADJUSTING FOR COUNTRY-LEVEL VARIATION IN DEMENTIA PREVALENCE WITH CLASSIFICATION ALGORITHMS IN SHARE

Abstract Background. Population-level dementia prevalence depends on societal factors and individual-level risk and protective factors. To improve our understanding of how these factors interact, we can use cross-national surveys such as the Survey of Health, Ageing and Retirement in Europe (SHARE). However, in absence of validated cognitive assessments, adjusting for underdiagnosis of dementia is needed. The present study sought to explore the usefulness of the Langa-Weir and alternative algorithms to detect probable dementia while accounting for country-level variation in estimated prevalence and underdiagnosis of dementia.Method. Data from 57,880 respondents aged 60 years and older to wave 7 of SHARE (2017) with non-missing data on variables related to sociodemographics and cognition were used. Adaptations of the Langa-Weir classification algorithm were compared to a weighted logistic regression model and an XGBoost classifier applying the synthetic minority oversampling technique. Different specifications of algorithms were tested globally and for individual countries and compared with the World Alzheimer’s Report (2015)’s country-level projections of dementia prevalence for 2018. Results. All algorithms accurately classified self-reported diagnosis of dementia (accuracy = 0.90-0.96), with the Langa-Weir classification based on recall and a cutoff reflecting country-specific prevalence outperforming other algorithms regarding compensation of underdiagnosis. Algorithmically detected probable dementia is associated with newly self-reported dementia diagnosis, drop-out and death two years later. Discussion. Identifying probable dementia through classification algorithms can increase statistical power and improve validity in cross-national investigations. Further research is needed to replicate the findings in validated cognitive assessments and to identify causes of cross-national variation in dementia underdiagnosis.

EXPLORING THE USE OF MACHINE LEARNING TO IMPROVE DEMENTIA DETECTION: AUTOMATING CODING OF THE CLOCK-DRAWING TEST

Abstract Alzheimer’s disease and related dementias, a leading cause of disability among older adults, has become a critical public health concern. The clock-drawing test (CDT), which asks subjects to draw a clock, typically with hands showing 11:10, has been widely used as a screening tool to detect dementia in clinical research and surveys. The clock-drawings are often coded into binary (e.g., normal vs. abnormal) or ordinal scores. A limitation in large-scale studies is that the manual-coding of CDT could result in biases if coders interpret and implement coding rules differently. Several small-scale studies have explored the use of machine learning methods to automate CDT coding. Such studies either have had limited success with ordinal coding or have used methods that are not designed specifically for complex images. This study aims to create and evaluate an intelligent CDT-scoring system that automatically codes ordinal CDT scores using deep learning neural networks (DLNN) methods including resnet101, EfficientNet-B3, and Visual Transformer. We used a large repository of CDT images from the 2011-2019 National Health and Aging Trends Study, a panel study of Medicare beneficiaries ages 65 and older. Results show that 1) DLNN has achieved a high scoring accuracy; 2) DLNN generates more consistent and accurate scores than trained lay coders, compared to expert coders with a neuropsychology background; 3) DLNN-coded CDT scores are highly correlated with self-reported dementia diagnosis and performance-based assessments of memory. This study offers a model for automating coding of other drawing tests used to evaluate a variety of cognitive functions.

Falls Among High- and Low- Frequency Sleep Medication Users With and Without Dementia

Difficulty with sleep and falls are prevalent among older adults. Sleep medication use is associated with falls in older adults, but little is known about its impact in older adults with dementia. We used data from the 2011 National Health and Aging Trends Study to assess the association of low- versus high- frequency sleep medication use with falls in older adults with self-reported dementia. In our fully adjusted model, among those with dementia, high-frequency sleep medication users were more likely to fall than low-frequency sleep medication users (OR=3.86, 95% CI: 1.31, 11.37). Among those without dementia, high-frequency sleep medication users were more likely to fall than low-frequency sleep medication users (OR=1.40, 95% CI: 1.11, 1.77). Reducing sleep medication use in older adults with and without dementia may help reduce the risk of falls and fall-related outcomes in older adults.

Generational Multimorbidity Disease Clusters for British Cohorts Born 1921 – 1960

The aim of this study is the first step in our understanding of the uniqueness and stability of multmorbdity disease patterns for different generations. The unique historical context that each generation has been exposed to is thought to have systemic health impacts and differences in epidemiological make-up (Clouston et al. 2021). Literature suggests that multimorbidity disease patterns, are similar across countries (Hernandez et al, 2021 – in press) and observational points, and that migration into complex disease clusters is more common as people age (Cassell et al, 2018, Kingston et al. 2018). Most commonly reported are Cardiovascular and Metabolic disease clusters which lead to lower quality of life, mortality and morbidity (Kudesia, 2021). We asked: Do multimorbidity disease patterns differ for unique generations? Using the ELSA, the disease clusters of three cohorts were examined; an older cohort, born 1921-1930, a middle cohort born 1931-1940 a younger cohort born 1941-1950 and the ”newest” cohort, born 1951-1960. Self-reported dementia and memory problems lead a specific cluster for the middle cohort, those born in 1931-1940, but not for the other cohorts. While disease patterns were different between sex for other clusters, the disease cluster of dementia and memory problems held similar disease patterns for males and females, with a prevalence of 3%. The dementia/memory problem cluster loaded with cardio/metabolic diseases. This suggests that complex multimorbidity for the British 1931-1940 cohort has had an impact related to dementia and memory problem diagnoses for this specific generation, for males and females alike.

Dementia knowledge and associated factors among older Chinese adults: a cross-national comparison between Melbourne and Beijing.

This study compared dementia knowledge between older Chinese adults in Melbourne, Australia, and Beijing, China, and explored factors associated with dementia knowledge between these two groups. Ultimately, this study aimed to inform the development of tailored dementia education programs for older Chinese adults. A cross-sectional design was employed in this study. Participants were recruited from 5 Chinese community senior groups in Melbourne and 10 community health centers in Beijing from March to May 2019. A total of 379 older Chinese adults aged 50 and over completed the questionnaire, including 153 from Melbourne and 226 from Beijing. Dementia knowledge was assessed using the Alzheimer's Disease Knowledge Scale (ADKS). Demographic characteristics, dementia-related experience, and the mental health status of participants were collected. Stepwise linear regression was used to analyze the factors associated with dementia knowledge. In general, older Chinese adults in Melbourne and Beijing reported similar levels of dementia knowledge for both the overall ADKS scale (mean ± SD: 17.2 ± 2.9 in Melbourne vs. 17.5 ± 2.9 in Beijing, p > 0.05) and the seven subdomains. Of the subdomains, the highest correct response rates were observed in the life impact of the dementia subdomain, and the lowest rates were observed in the caregiving subdomain. Stepwise linear regression analysis revealed that younger age and self-reported dementia worry were significantly associated with higher levels of dementia knowledge in the Melbourne group, whereas a positive family history of dementia was significantly associated with higher levels of dementia knowledge in the Beijing group. Older Chinese adults living in Melbourne and Beijing share similar levels of dementia knowledge, but factors associated with their knowledge are different. These findings will inform the development of culturally and socially appropriate dementia education programs for older Chinese populations in different countries.

The relationship of cognitive change over time to the self-reported Ascertain Dementia 8-item Questionnaire in a general population.

The aim of the study was to examine the relationship between longitudinally assessed cognitive functioning and self-reported dementia status using the Ascertain Dementia 8-item questionnaire (AD8) in a national population-based sample. The analysis included 14,453 participants from the REasons for Geographic and Racial Differences in Stroke study. A validated cutoff of ≥2 symptoms endorsed on the AD8 (administered 10years after enrollment) represented positive AD8 status. Incident cognitive impairment was defined as change from intact to impaired status in the Six-Item Screener score, and cognitive decline was defined by trajectories of Letter "F" Fluency from the Montreal Cognitive Assessment, and Animal Fluency, Word List Learning, and Word List Delayed recall, all from the Consortium to Establish a Registry for Alzheimer's Disease battery. Logistic regression models controlled for demographics, health variables, and depressive symptoms. Sensitivity and specificity of the AD8 to detect incident cognitive impairment were 45.2% and 78.4%, respectively. Incident cognitive impairment and a one-word decline in WLL increased the odds of self-reported positive AD8 by 96% (95% CI: 1.68-2.28) and 27% (95% CI: 1.17-1.37), respectively. There was a strong association between high depression risk and self-reported positive AD8 in sensitivity analyses. Incident cognitive impairment and high depression risk were the strongest predictors of self-reported positive AD8 in this population-based sample. Our results inform the utility of the AD8 as a self-report measure in a large, national sample that avoids selection biases inherent in clinic-based studies. The AD8 is screening measure and should not be used to diagnose dementia clinically.