协变量
比例危险模型
人口学
联想(心理学)
分位数回归
医学
老年学
回归分析
心理学
统计
环境卫生
内科学
人口
数学
社会学
心理治疗师
作者
Neha Agarwala,Vadim Zipunnikov,Ciprian Crainiceanu,Andrew Leroux
出处
期刊:BMJ evidence-based medicine
[BMJ]
日期:2024-03-12
卷期号:: bmjebm-112303
标识
DOI:10.1136/bmjebm-2023-112303
摘要
Objectively measuring physical activity (PA) has consistently shown an association with reduced all-cause mortality risk in cross-sectional studies. However, the strength of this association may change over time. We quantify the time-varying, covariate-adjusted association between the total volume of PA and all-cause mortality over a 12-year follow-up period using Cox regression with a time varying effect of population-referenced quantile total activity count adjusted for traditional risk factors. Analyses focus on participants 50–84 years old with adequate accelerometer wear time and without missing covariates. The findings suggest that (1) the use of baseline PA in Cox models with long follow-up periods may be inappropriate without time-varying effects and (2) the use of accelerometry derived volume of PA in risk score calculations may be most appropriate for short-term to medium-term risk scores.
科研通智能强力驱动
Strongly Powered by AbleSci AI