Genetics of changes in population pyramids: Implications for health forecasting with a focus on Alzheimer's Disease and related dementias

Duke University, Durham NC

Investigators

Linked publications, trials & patents

Abstract

The objective of the proposed research is to provide a series of gender-specific forecasts of the future burden of Alzheimer?s disease and related dementia (AD/ADRD) in connection with the changing trends in other major diseases common in U.S. older adults. This objective will be reached by incorporating relevant information on the demographic, social, economic, health, and genetic structure of the U.S. population into the forecasting models. This information will be obtained from four diverse sources of Medicare data: The Surveillance Epidemiology and End Results (SEER-M), The Medicare 5% Research Identifiable Files (5%-Medicare), The Health and Retirement Study (HRS-M) and the National Long Term Care Survey (NLTCS-M) as well as from other supplementary sources. This work will build on our recent methodological developments in modeling and prediction of health-related traits to account for changes in health-related components and estimate the associated projected changes in utilization of the Medicare system. In these analyses, model parameters and resultant forecasts will be generated using Medicare-5% then validated on SEER-M, HRS-M and NLTCS-M. In analyses requiring individual-level social economic or genetic information, HRS-M will be the primary and NLTCS-M the validation dataset. For each study scenario, short-term (to 2028), medium-term (to 2038) and long-term (to 2050) forecasts will be made. Three specific aims will be addressed: 1. Develop new forecasting models of AD/ADRD burden that incorporate the effects of disease severity, fraction of AD/ADRD under- diagnosis, and mixed (multiple) dementia as well as the time patterns of AD/ADRD using a McKendrick-von Foerster equation-based approach. Apply other approaches including the multistate approach (a popular forecasting method), statistical projections, time series methods, Lee-Carter-based approaches, and microsimulation and compare their performance to that of our primary method. Construct realistic projections under multiple hypothetical scenarios. 2. Develop a McKendrick-von Foerster equation-based approach to incorporate the effects of co- and multi-morbidity including cerebrovascular diseases, type 2 diabetes mellitus, depression, cancer and other health disorders that show strong connections with AD/ADRD. Generate new scenarios based on information on health disorders included in forecasting models. Update and compare previous forecasts of AD/ADRD burden. 3. Create distinct groups of individuals based on their demographic, socioeconomic, behavioral and genetic characteristics. Estimate the group-specific AD/ADRD burden and evaluate the differences in outcomes, disease severity and fraction of under-diagnosis between these groups.. Update and compare previous forecasts of AD/ADRD burden. Integrate findings and compare the results obtained study-wide under all chosen scenarios.