Learning network for Advanced Behavioural Data Analysis (LABDA)

最近，出现了一种范式转变，从孤立地关注单一行为(即PA、久坐行为或睡眠)对健康的影响，转向将这些全天候运动行为结合在一起，以实现最大的健康益处。然而，目前的公共卫生准则在很大程度上是基于不准确的自我报告数据，因此相当笼统(例如，“多运动，少坐”)。技术进步导致了可穿戴传感器技术的出现，在更长的时间段内提供了丰富的时间序列数据。因此，需要新的分析方法来详细了解多维全天候运动行为特征与健康之间的联系；哪些子组需要特别关注；哪些行为特征是干预中最重要的目标。开发这种新的分析方法对于创建最佳、量身定制的指南和反馈所需的证据库至关重要，需要将流行病学、数据科学、方法开发和公共卫生方面的知识和技能具体结合起来，并彻底了解将知识转化为指南和改进可穿戴技术反馈所需的内容。因此，在LABDA，我们将培训10名博士生，以推动这一跨学科领域的发展，并提供基于传感器的行为数据的高级分析方法工具箱，以及供其他研究人员和政策制定者决定使用哪些方法来解决哪个(研究)问题的指南。

Recently, there has been a paradigm shift from the isolated focus on the health impact of a single behaviour (i.e. PA, sedentary behaviour or sleep) to the combination of these 24/7 movement behaviours for maximum health benefits. However, current public health guidelines are largely based on inaccurate self-report data and are, therefore, rather general (e.g. "move more and sit less"). Technological advancements have led to wearable sensor techniques providing rich time-series data over longer periods. Consequently, novel analysis methods are required to provide detailed insight into the links between multi-dimensional 24/7 movement behaviour profiles and health; which subgroups need particular attention; and what behavioural profiles are most important to target in interventions.Developing such novel analysis methods, essential for creating the evidence base needed for optimal, tailored guidelines and feedback, requires a specific combination of knowledge and skills in epidemiology, data science, method development, and public health with a thorough understanding of what is needed to translate knowledge to guidelines and improve wearable technology feedback. In LABDA, we will therefore train 10 doctoral fellows to advance this interdisciplinary field and deliver a toolbox of advanced analysis methods for sensor-based behavioural data, together with a guide for other researchers and policy makers to decide which methods to use for which (research) question.