Personalized predictive modeling for patients with Alzheimer's disease using an extension of Sullivan's life table model

Development and validation of a clinical prediction tool to estimate survival in community-dwelling adults living with dementia: a protocol

INTRODUCTION

A clinical prediction tool to estimate life expectancy in community-dwelling individuals living with dementia could inform healthcare decision-making and prompt future planning. An existing Ontario-based tool for community-dwelling elderly individuals does not perform well in people living with dementia specifically. This study seeks to develop and validate a clinical prediction tool to estimate survival in community-dwelling individuals living with dementia receiving home care in Ontario, Canada.

METHODS AND ANALYSIS

This will be a population-level retrospective cohort study that will use data in linked healthcare administrative databases at ICES. Specifically, data that are routinely collected from regularly administered assessments for home care will be used. Community-dwelling individuals living with dementia receiving home care at any point between April 2010 and March 2020 will be included (N≈200 000). The model will be developed in the derivation cohort (N≈140 000), which includes individuals with a randomly selected home care assessment between 2010 and 2017. The outcome variable will be survival time from index assessment. The selection of predictor variables will be fully prespecified and literature/expert-informed. The model will be estimated using a Cox proportional hazards model. The model's performance will be assessed in a temporally distinct validation cohort (N≈60 000), which includes individuals with an assessment between 2018 and 2020. Overall performance will be assessed using Nagelkerke's R2, discrimination using the concordance statistic and calibration using the calibration curve. Overfitting will be assessed visually and statistically. Model performance will be assessed in the validation cohort and in prespecified subgroups.

ETHICS AND DISSEMINATION

The study received research ethics board approval from the Sunnybrook Health Sciences Centre (SUN-6138). Abstracts of the project will be submitted to academic conferences, and a manuscript thereof will be submitted to a peer-reviewed journal for publication. The model will be disseminated on a publicly accessible website (www.projectbiglife.com).

TRIAL REGISTRATION NUMBER

NCT06266325 (clinicaltrials.gov).

Trajectories and Transitions in Service Use Among Older Veterans at High Risk of Long-Term Institutional Care

BACKGROUND

We aimed to identify combinations of long-term services and supports (LTSS) Veterans use, describe transitions between groups, and identify factors influencing transition.

METHODS

We explored LTSS across a continuum from home to institutional care. Analyses included 104,837 Veterans Health Administration (VHA) patients 66 years and older at high-risk of long-term institutional care (LTIC). We conduct latent class and latent transition analyses using VHA and Medicare data from fiscal years 2014 to 2017. We used logistic regression to identify variables associated with transition.

RESULTS

We identified 5 latent classes: (1) No Services (11% of sample in 2015); (2) Medicare Services (31%), characterized by using LTSS only in Medicare; (3) VHA-Medicare Care Continuum (19%), including LTSS use in various settings across VHA and Medicare; (4) Personal Care Services (21%), characterized by high probabilities of using VHA homemaker/home health aide or self-directed care; and (5) Home-Centered Interdisciplinary Care (18%), characterized by a high probability of using home-based primary care. Veterans frequently stayed in the same class over the three years (30% to 46% in each class). Having a hip fracture, self-care impairment, or severe ambulatory limitation increased the odds of leaving No Services, and incontinence and dementia increased the odds of entering VHA-Medicare Care Continuum. Results were similar when restricted to Veterans who survived during all 3 years of the study period.

CONCLUSIONS

Veterans at high risk of LTIC use a combination of services from across the care continuum and a mix of VHA and Medicare services. Service patterns are relatively stable for 3 years.

Differences in the Distribution of Aβ in the Brain between U.S. Veterans and Adults aged 62+ and suffering from Alzheimer's Disease

Background

Elevated concentration of amyloids in the cerebrum results in elevated risks for cerebral hemorrhage and early AD onset following early depression/dementia onset. In this study, we compare patterns of amyloid depositions across eight regions of interest of the human brain between U.S. Veterans and non-Veterans adults aged 62+.

Data

Data were taken from the ADNI and DoD-ADNI studies. A pseudo-randomization algorithm was applied to achieve comparability, reduce bias due to age mismatching, and account for non-treatment-related differences between subsamples extracted from DoD-ADNI and ADNI databases. The pool of participants included data about age, race, apolipoprotein ε4 allele (APOE) status, modified Hachinski Ischemic Score, education level, and geriatric depression score, which were used to build a propensity score.

Predictors and outcomes

Aβ concentration, resulting from the PET image analysis, in key brain regions of interest, and two categorical variables describing the 0.79 and 1.11 cutoffs were used as outcomes, while the Veteran and AD status were used as predictors.

Methods

To balance subsamples, we applied a pseudo-randomization algorithm, eliminating the observed sources of heterogeneity. We used a generalized linear model for continuous variables and the logistic regression model for binary variables.

Findings

The pattern of the Aβ distribution in Veteran's brains was found to be different from the classic AD pattern. The amyloid depositions following Veteran status were concentrated in cerebellar gray matter and the cerebellum in general. In contrast, the AD pattern shows more Aβ depositions in the frontal lobe, cingulate cortex, parietal, and temporal lobes, along with higher whole-cerebrum concentration of amyloids. Since Florbetapir PET cannot distinguish between senile plaques and depositions in blood vessels, the elevated concentration of amyloids in a cerebellum for participants with the Veteran status may suppose elevated risks for cerebral hemorrhage and early AD onset following early depression/dementia onset.

Development and External Validation of Models to Predict Need for Nursing Home Level of Care in Community-Dwelling Older Adults With Dementia

Importance

Most older adults living with dementia ultimately need nursing home level of care (NHLOC).

Objective

To develop models to predict need for NHLOC among older adults with probable dementia using self-report and proxy reports to aid patients and family with planning and care management.

Design, Setting, and Participants

This prognostic study included data from 1998 to 2016 from the Health and Retirement Study (development cohort) and from 2011 to 2019 from the National Health and Aging Trends Study (validation cohort). Participants were community-dwelling adults 65 years and older with probable dementia. Data analysis was conducted between January 2022 and October 2023.

Exposures

Candidate predictors included demographics, behavioral/health factors, functional measures, and chronic conditions.

Main Outcomes and Measures

The primary outcome was need for NHLOC defined as (1) 3 or more activities of daily living (ADL) dependencies, (2) 2 or more ADL dependencies and presence of wandering/need for supervision, or (3) needing help with eating. A Weibull survival model incorporating interval censoring and competing risk of death was used. Imputation-stable variable selection was used to develop 2 models: one using proxy responses and another using self-responses. Model performance was assessed by discrimination (integrated area under the receiver operating characteristic curve [iAUC]) and calibration (calibration plots).

Results

Of 3327 participants with probable dementia in the Health and Retirement Study, the mean (SD) age was 82.4 (7.4) years and 2301 (survey-weighted 70%) were female. At the end of follow-up, 2107 participants (63.3%) were classified as needing NHLOC. Predictors for both final models included age, baseline ADL and instrumental ADL dependencies, and driving status. The proxy model added body mass index and falls history. The self-respondent model added female sex, incontinence, and date recall. Optimism-corrected iAUC after bootstrap internal validation was 0.72 (95% CI, 0.70-0.75) in the proxy model and 0.64 (95% CI, 0.62-0.66) in the self-respondent model. On external validation in the National Health and Aging Trends Study (n = 1712), iAUC in the proxy and self-respondent models was 0.66 (95% CI, 0.61-0.70) and 0.64 (95% CI, 0.62-0.67), respectively. There was excellent calibration across the range of predicted risk.

Conclusions and Relevance

This prognostic study showed that relatively simple models using self-report or proxy responses can predict need for NHLOC in community-dwelling older adults with probable dementia with moderate discrimination and excellent calibration. These estimates may help guide discussions with patients and families in future care planning.

Life expectancy and healthy life expectancy of Korean registered disabled by disability type in 2014-2018: Korea National Rehabilitation Center database

BACKGROUND

Conducting a distinct comparison between the life expectancy (LE) and healthy life expectancy (HALE) of people with disabilities (PWDs) and the general population is necessary due to the various environmental and health conditions they encounter. Therefore, this study aimed to develop the life table for PWDs and calculate those of LE and HALE based on sex, severity, and disability types among the registered Korean PWDs.

METHODS

We used aggregated data of registered PWDs from the Korea National Rehabilitation Center database between 2014 and 2018. Overall, 345,595 deaths were included among 12,627,428 registered PWDs. First, we calculated the LE for total PWDs and non-disabled people using a standard life table, extending the old age mortality among nine models. Subsequently, we calculated the LE for each type of disability using the relationship between the mortality of total PWDs and those of each type of disability. Finally, HALE was calculated using the Sullivan method for three types as follows: disability-free and perceived health (PH) using the National Survey, and hospitalized for  ≥ 7 days using the Korea National Health Insurance System (NHIS) database.

RESULTS

The calculated LE/HALE-NHIS (years) at registration in males and females were 81.32/73.32 and 87.38/75.58, 68.54/58.98 and 71.43/59.24, 73.87/65.43 and 78.25/67.51, and 61.53/50.48 and 62.41/49.72 years among non-disabled, total PWDs, mild disabled, and severe disabled, respectively. LE/HALE-NHIS was lowest and highest in respiratory dysfunction and hearing disabilities, respectively.

CONCLUSIONS

Males with disabilities had shorter LE and HALE at registration than females, except for those with severe disabilities, and there were variabilities in the LE based on the disability types.

Associations between cognitive function and a range of significant life events in an elderly essential tremor cohort study

BACKGROUND

Although essential tremor (ET) is associated with cognitive decline, we know little about how specific cognitive changes predict significant events in patients' lives. We examined the relations of attention, executive function, language, memory, and visuospatial performance to the occurrence of near falls, falls, walking aid use, home health aide use, non-independent living and hospitalizations within a prospective, longitudinal study of ET cases. We expected executive function and memory to be most strongly associated with these events.

METHODS

131 ET cases (mean age at baseline = 76.4 ± 9.4 years; 109 normal cognition; 17 mild cognitive impairment, 5 demented) completed questionnaires (clinical history and occurrence of life events) and a battery of neuropsychological tests at baseline and at 18, 36, and 54 months. We assessed associations between cognitive functioning and outcomes via regression equations.

RESULTS

Cases with lower baseline levels of executive function reported more near falls, p < 0.006, and were more likely to use a walking aid, p < 0.03, odds ratio (OR) = 2.89 during the follow-up period, than were other cases. Decline in executive function was associated with home health aide use during follow-up, p < 0.04, OR = 3.34. Baseline visuospatial performance also bore a marginally significant association with non-independent living arrangements during follow-up, p < 0.06, OR = 2.13. These effects were independent of age and tremor severity.

CONCLUSION

These data establish the important role that cognitive decline, and executive function specifically, play in the experiences of ET patients. Moreover, these associations are of sufficient magnitude to have significant clinical implications.

Validation of a Multivariate Prediction Model of the Clinical Progression of Alzheimer's Disease in a Community-Dwelling Multiethnic Cohort

BACKGROUND

The major aims of the three Predictors Studies have been to further our understanding of Alzheimer's disease (AD) progression sufficiently to predict the length of time from disease onset to major disease outcomes in individual patients with AD.

OBJECTIVES

To validate a longitudinal Grade of Membership (L-GoM) prediction algorithm developed using clinic-based, mainly white patients from the Predictors 2 Study in a statistically representative community-based sample of Hispanic (N = 211) and non-Hispanic (N = 62) older adults (with 60 males and 213 females) from the Predictors 3 Study and extend the algorithm to mild cognitive impairment (MCI).

METHODS

The L-GoM model was applied to data collected at the initial Predictors 3 visit for 150 subjects with AD and 123 with MCI. Participants were followed annually for up to seven years. Observed rates of survival and need for full-time care (FTC) were compared to those predicted by the algorithm.

RESULTS

Initial MCI/AD severity in Predictors 3 was substantially higher than among clinic-based AD patients enrolled at the specialized Alzheimer's centers in Predictors 2. The observed survival and need for FTC followed the L-GoM model trajectories in individuals with MCI or AD, except for N = 32 subjects who were initially diagnosed with AD but reverted to a non-AD diagnosis on follow-up.

CONCLUSION

These findings indicate that the L-GoM model is applicable to community-dwelling, multiethnic older adults with AD. They extend the use of the model to the prediction of outcomes for MCI. They also justify release of our L-GoM calculator at this time.

Development and External Validation of a Mortality Prediction Model for Community-Dwelling Older Adults With Dementia

Importance

Estimating mortality risk in older adults with dementia is important for guiding decisions such as cancer screening, treatment of new and chronic medical conditions, and advance care planning.

Objective

To develop and externally validate a mortality prediction model in community-dwelling older adults with dementia.

Design, Setting, and Participants

This cohort study included community-dwelling participants (aged ≥65 years) in the Health and Retirement Study (HRS) from 1998 to 2016 (derivation cohort) and National Health and Aging Trends Study (NHATS) from 2011 to 2019 (validation cohort).

Exposures

Candidate predictors included demographics, behavioral/health factors, functional measures (eg, activities of daily living [ADL] and instrumental activities of daily living [IADL]), and chronic conditions.

Main Outcomes and Measures

The primary outcome was time to all-cause death. We used Cox proportional hazards regression with backward selection and multiple imputation for model development. Model performance was assessed by discrimination (integrated area under the receiver operating characteristic curve [iAUC]) and calibration (plots of predicted and observed mortality).

Results

Of 4267 participants with probable dementia in HRS, the mean (SD) age was 82.2 (7.6) years, 2930 (survey-weighted 69.4%) were female, and 785 (survey-weighted 12.1%) identified as Black. Median (IQR) follow-up time was 3.9 (2.0-6.8) years, and 3466 (81.2%) participants died by end of follow-up. The final model included age, sex, body mass index, smoking status, ADL dependency count, IADL difficulty count, difficulty walking several blocks, participation in vigorous physical activity, and chronic conditions (cancer, heart disease, diabetes, lung disease). The optimism-corrected iAUC after bootstrap internal validation was 0.76 (95% CI, 0.75-0.76) with time-specific AUC of 0.73 (95% CI, 0.70-0.75) at 1 year, 0.75 (95% CI, 0.73-0.77) at 5 years, and 0.84 (95% CI, 0.82-0.85) at 10 years. On external validation in NHATS (n = 2404), AUC was 0.73 (95% CI, 0.70-0.76) at 1 year and 0.74 (95% CI, 0.71-0.76) at 5 years. Calibration plots suggested good calibration across the range of predicted risk from 1 to 10 years.

Conclusions and Relevance

We developed and externally validated a mortality prediction model in community-dwelling older adults with dementia that showed good discrimination and calibration. The mortality risk estimates may help guide discussions regarding treatment decisions and advance care planning.

Targeted Screening for Alzheimer's Disease Clinical Trials Using Data-Driven Disease Progression Models

Heterogeneity in Alzheimer's disease progression contributes to the ongoing failure to demonstrate efficacy of putative disease-modifying therapeutics that have been trialed over the past two decades. Any treatment effect present in a subgroup of trial participants (responders) can be diluted by non-responders who ideally should have been screened out of the trial. How to identify (screen-in) the most likely potential responders is an important question that is still without an answer. Here, we pilot a computational screening tool that leverages recent advances in data-driven disease progression modeling to improve stratification. This aims to increase the sensitivity to treatment effect by screening out non-responders, which will ultimately reduce the size, duration, and cost of a clinical trial. We demonstrate the concept of such a computational screening tool by retrospectively analyzing a completed double-blind clinical trial of donepezil in people with amnestic mild cognitive impairment (clinicaltrials.gov: NCT00000173), identifying a data-driven subgroup having more severe cognitive impairment who showed clearer treatment response than observed for the full cohort.

Development of a Longitudinal Dataset of Persons With Dementia and Their Caregivers Through End-of-Life: A Statistical Analysis System Algorithm for Joining National Health and Aging Trends Study/National Study of Caregiving

Background: Alzheimer’s disease and related dementias (AD/ADRD) are terminal conditions impacting families and caregivers, particularly at end-of-life. Longitudinal, secondary data analyses present opportunities for insight into dementia caregiving and decision-making over time; however, joining complex datasets and preparing them for analysis poses many challenges. Objectives: To describe an approach to linking national survey data of older adults with their primary caregivers to build a prospective, longitudinal dataset, and to share the Statistical Analysis System (SAS) coding statement algorithms with other researchers. Methods: The National Health and Aging Trends Study (NHATS) and National Study of Caregiving (NSOC) are joined using a series of algorithms based on conceptual and operational definitions of dementia, primary caregivers, and the occurrence of death. A series of SAS algorithms resulting in the final longitudinal dataset was created. Results: NHATS/NSOC participants were linked using three preliminary data files (n = 12 427) and one final data join (n = 3305) over nine rounds of data collection. Presence of dementia was defined based on the indicator in the year preceding the last month-of-life (LML) interview. Primary caregivers were defined as the person providing the most frequent care over time. Additional flag variables (LML interview, dementia classification, and cohort (2011 vs 2015)) were created. The SAS algorithms are presented herein. Discussion: The SAS coding statement algorithms provide an opportunity to conduct longitudinal analysis of care for both members of the dyad in the context of dementia and end-of-life. Future research using the proposed dataset can further explore care and caregiving in these populations.

Validation and demonstration of a new comprehensive model of Alzheimer's disease progression

INTRODUCTION

Identifying the course of Alzheimer's disease (AD) for individual patients is important for numerous clinical applications. Ideally, prognostic models should provide information about a range of clinical features across the entire disease process. Previously, we published a new comprehensive longitudinal model of AD progression with inputs/outputs covering 11 interconnected clinical measurement domains.

METHODS

Here, we (1) validate the model on an independent cohort; and (2) demonstrate the model's utility in clinical applications by projecting changes in 6 of the 11 domains.

RESULTS

Survival and prevalence curves for two representative outcomes-mortality and dependency-generated by the model accurately reproduced the observed curves both overall and for patients subdivided according to risk levels using an independent Cox model.

DISCUSSION

The new model, validated here, effectively reproduces the observed course of AD from an initial visit assessment, allowing users to project coordinated developments for individual patients of multiple disease features.

Understanding and predicting the longitudinal course of dementia

PURPOSE OF REVIEW

To date, most research in dementia has focused either on the identification of dementia risk prediction or on understanding changes and predictors experienced by individuals before diagnosis. Despite little is known about how individuals change after dementia diagnosis, there is agreement that changes occur over different time scales and are multidomain. In this study, we present an overview of the literature regarding the longitudinal course of dementia.

RECENT FINDINGS

Our review suggests the evidence is scarce and findings reported are often inconsistent. We identified large heterogeneity in dementia trajectories, risk factors considered and modelling approaches employed. The heterogeneity of dementia trajectories also varies across outcomes and domains investigated.

SUMMARY

It became clear that dementia progresses very differently, both between and within individuals. This implies an average trajectory is not informative to individual persons and this needs to be taken into account when communicating prognosis in clinical care. As persons with dementia change in many more ways during their patient journey, heterogeneous disease progressions are the result of disease and patient characteristics. Prognostic models would benefit from including variables across a number of domains. International coordination of replication and standardization of the research approach is recommended.