Abstract
IntroductionThe world's population is aging rapidly, leading to increased public health and economic burdens due to age-related cardiovascular and neurodegenerative diseases. Early risk detection is essential for prevention and to improve the quality of life in elderly individuals. Plus, health risks associated with aging are not directly tied to chronological age, but are also influenced by a combination of environmental exposures. Past research has introduced the concept of "Phenotypic Age," which combines age with biomarkers to estimate an individual's health risk.MethodsThis study explores which factors contribute most to the gap between chronological and phenotypic ages. We combined ten machine learning regression techniques applied to the NHANES dataset, containing demographic, laboratory and socioeconomic data from 41,474 patients, to identify the most important features. We then used clustering analysis and a mixed-effects model to stratify by sex, ethnicity, and education.ResultsWe identified 28 demographic, biological and environmental factors related to a significant gap between phenotypic and chronological ages. Stratifying for sex, education and ethnicity, we found statistically significant differences in the outcome distributions.ConclusionBy showing that health risk prevention should consider both biological and sociodemographic factors, we offer a new approach to predict aging rates and potentially improve targeted prevention strategies for age-related conditions.
Funding Information
  • National Institutes of Health (P30 AG073105)

This publication has 31 references indexed in Scilit: