Braun Maya, Crombez Geert, De Backere Femke, Tack Emma, De Paepe Annick L
Department of Experimental Clinical and Health Psychology, Ghent University, Ghent, Belgium.
Institute for Implementation Science in Health Care, University of Zurich, Zurich, Switzerland.
Health Psychol Behav Med. 2024 Dec 1;12(1):2434140. doi: 10.1080/21642850.2024.2434140. eCollection 2024.
Personalising recommendations for physical activity coping plans can help bridging the physical activity intention-behaviour gap. Data-driven 'black-box' approaches result in recommendations that prove difficult to explain, and may have undesired consequences. This study aimed to explicitly link barriers and coping strategies using end-user input.
152 participants (85 female) took part in an online task. Participants were asked to judge the relevance of coping strategies for barriers to physical activity, and under which circumstances coping strategies were relevant for a given barrier. Data was aggregated and heat maps were produced. Necessary conditions for the relevance of each combination were coded and their frequencies were reported.
Relevance of 1570 combinations of barriers and coping strategies were assessed, with 2 combinations rated 'always relevant' by all participants, and 37 combinations rated as 'always relevant' by no participants. Barriers differ strongly in how many coping strategies are relevant for them, and coping strategies differ strongly in how many barriers they are relevant for. Resulting aggregates concerning the average rating as 'never relevant', 'always relevant' and 'relevant under certain conditions' are shared for each barrier coping strategy combination, as are the conditions associated with different barriers and coping strategies.
This study introduces a novel method to create rules for recommendations using input from stakeholders. The datasets created throughout this research are available for re-use in future research, as well as for clinical practice and (digital) intervention development. This data can be used as a base for explainable personalised recommendations for physical activity coping plans.
针对体育活动应对计划进行个性化推荐有助于弥合体育活动意图与行为之间的差距。数据驱动的“黑箱”方法得出的推荐难以解释，且可能产生不良后果。本研究旨在利用最终用户的输入明确关联障碍和应对策略。
152名参与者（85名女性）参与了一项在线任务。参与者被要求判断应对策略与体育活动障碍的相关性，以及应对策略在何种情况下与特定障碍相关。数据进行了汇总并生成了热图。对每种组合相关性的必要条件进行了编码，并报告了其频率。
评估了1570种障碍与应对策略组合的相关性，其中2种组合被所有参与者评为“始终相关”，37种组合没有参与者评为“始终相关”。不同障碍与之相关的应对策略数量差异很大，应对策略与之相关的障碍数量差异也很大。每种障碍应对策略组合的“从不相关”“始终相关”和“在某些条件下相关”的平均评分汇总结果，以及与不同障碍和应对策略相关的条件均已列出。
本研究引入了一种新颖的方法，利用利益相关者的输入创建推荐规则。本研究过程中创建的数据集可在未来研究中重复使用，也可用于临床实践和（数字）干预开发。这些数据可作为体育活动应对计划可解释的个性化推荐的基础。
