Statistical methods for analyzing objectively measured physical activity.

项目总结/摘要 该项目旨在开发新的分析工具，以了解身体活动（PA）的积累模式， 久坐行为（SB）和睡眠在24小时活动周期，并灵活地模拟这些模式和健康 结果，如心血管疾病，并应用这些方法在PA流行病学研究，收集 通过加速度计或可穿戴设备进行客观测量。在体育活动流行病学中，每一个inten- 已经单独研究了与健康结果相关的sity/behavior类别。但这 这种方法未能考虑到这些活动之间的相互关联和相互作用。事实上，2018年 身体活动指南咨询委员会科学报告确定的艾德24小时周期和剂量反应 PA/SB积累模式是最重要和最关键的问题，显然需要进一步研究， 告知下一套身体活动指南。 出于这一关键需求，我们提出了新的统计方法来利用收集的丰富信息， 几个方向的加速度计首先，我们开发了强大的功能数据分析方法，用于灵活建模 加速度计测量的PA轨迹和健康结果之间的关联。第二，我们开发了灵活的 用于表征PA/SB蓄积之间剂量-反应关系的功能数据分析方法 模式和健康结果，使用可解释的功能主成分分析方法。第三、 我们计划开发非参数成分数据分析（NCDA）方法，用于提取24- 小时活动周期，扩展纵向研究的方法，并灵活地调查 24-小时活动组成模式和健康。 所提出的分析工具的直接动机，并将立即应用于辅助研究， 我们参与的妇女健康倡议（WHI），客观的身体活动和心血管 健康（OPACH）研究，WHI强壮和健康试验（WHISH），久坐时间和衰老（星星）计划， 妇女健康加速计合作（WHAC），以及公开提供的国家卫生和 营养检查调查（NHANES）和英国生物银行数据。

Project Summary/Abstract This project aims to develop new analytic tools for understanding accumulation patterns of physical activity (PA), sedentary behavior (SB) and sleep in the 24-hour activity cycle and to flexibly model these patterns and health outcomes, such as cardiovascular diseases, and apply these methods in PA epidiological studies that collected objective measurements by accelerometry or wearable devices. In physical activity epidemiology, each inten- sity/behavior category has been investigated separately in association with health outcomes. However, this approach fails to account for the inter-relatedness and interactions among of these activities. In fact, the 2018 Physical Activity Guidelines Advisory Committee Scientific Report identified 24-hour cycle and dose-response PA/SB accumulation patterns among the most important and critical issues that clearly need future research for informing the next set of physical activity guidelines. Motivated by this critical need, we propose novel statistical methods to utilize the rich information collected by accelerometry in a few directions. First, we develop robust functional data analysis methods for flexibly modeling associations between accelerometry-measured PA trajectories and health outcomes. Second, we develop flexible functional data analysis methods for characterizing dose-response relationships between PA/SB accumulation patterns and health outcomes, using an interpretable functional principal component analysis approach. Third, we plan to develop nonparametric compositional data analysis (NCDA) methods for extracting patterns of the 24- hour activity cycle, extend the approach to longitudinal studies and to flexibly investigate associations between 24-hour activity compositional patterns and health. The proposed analytic tools are directly motivated by and will be immediately applied to ancillary studies of the Women's Health Initiative (WHI) that we are involved in, the Objective Physical Activity and Cardiovascular Health (OPACH) Study, WHI Strong and Healthy Trial (WHISH), the Sedentary Time and Aging (STAR) Program, the Women's Health Accelerometry Collaboration (WHAC), as well as the publicly available National Health and Nutrition Examination Survey (NHANES) and UK Biobank data.