Time-Varying Associations Between Device-Based and Ecological Momentary Assessment–Reported Sedentary Behaviors and the Concurrent Affective States Among Adolescents: Proof-of-Concept Study

Abstract

Background Previous studies on affective state–sedentary behavior (SB) associations have not accounted for their potentially time-varying nature and have used inconsistent SB measurement modalities. We investigated whether the strength of the associations between affective states and SB varied as a function of the time of day and by SB measurement modality (device-measured SB vs ecological momentary assessment–reported screen-based SB) in youth. Objective This study aimed to establish a proof of concept that SB–affective state associations may not be static during the day. In addition, we aimed to inform the methodology of future work, which may need to model associations as functions of the time of day and carefully consider how SB is operationalized or measured. Methods A total of 15 adolescents (age: mean 13.07, SD 1.03 years; 10/15, 67% female; 6/15, 40% Hispanic; 10/15, 67% healthy weight) wore thigh-mounted activPAL accelerometers and simultaneously reported their screen-based SBs and concurrent positive and negative affective states via ecological momentary assessment for 7 to 14 days (N=636 occasions). Time-varying effect models (varying slopes) examined how each measure of SB was associated with concurrent affective states from 7 AM to 8 PM. Results Time-varying effect model plots revealed that these associations varied in strength throughout the day. Specifically, device-based SB was related to greater concurrent negative affect only after approximately 5 PM and was unrelated to concurrent positive affect. Screen-based SB was related to greater concurrent negative affect only from 7 AM to approximately 9 AM. This was also related to greater concurrent positive affect from 7 AM to approximately 9:30 AM and from approximately 3 PM to approximately 7 PM. Conclusions We provide preliminary evidence to suggest that future confirmatory studies investigating the SB–affective state relationship should consider the time-varying nature of these associations and SB measurement modality. There may be critical time windows when specific types of SBs co-occur with affect, suggesting that interventions may need tailoring to the time of day and type of SB if future studies using similar methodologies can replicate our findings.