A Strategy to Inform Athlete Sleep Support From Questionnaire Data and Its Application in an Elite Athlete Cohort

Abstract

Purpose: Information from the Pittsburgh Sleep Quality Index (PSQI) and Athlete Sleep Behavior Questionnaire (ASBQ) provide the ability to identify the sleep disturbances experienced by athletes and their associated athlete-specific challenges that cause these disturbances. However, determining the appropriate support strategy to optimize the sleep habits and characteristics of large groups of athletes can be time-consuming and resource-intensive. The purpose of this study was to characterize the sleep profiles of elite athletes to optimize sleep-support strategies and present a novel R package, AthSlpBehaviouR, to aid practitioners with athlete sleep monitoring and support efforts.

Methods: PSQI and ASBQ data were collected from a cohort of 412 elite athletes across 27 sports through an electronic survey. A k-means cluster analysis was employed to characterize the unique sleep-characteristic typologies based on PSQI and ASBQ component scores.

Results: Three unique clusters were identified and qualitatively labeled based on the z scores of the PSQI components and ASBQ components: cluster 1, "high-priority; poor overall sleep characteristics + behavioral-focused support"; cluster 2, "medium-priority, sleep disturbances + routine/environment-focused support"; and cluster 3, "low-priority; acceptable sleep characteristics + general support."

Conclusions: The findings of this study highlight the practical utility of an unsupervised learning approach to perform clustering on questionnaire data to inform athlete sleep-support recommendations. Practitioners can consider using the AthSlpBehaviouR package to adopt a similar approach in athlete sleep screening and support provision.

Keywords: ASBQ; PSQI; R package; clustering; machine learning.