Effectiveness of smartphone-based dyadic interventions to increase physical activity in romantic couples: microrandomized trial

Background Social exchange processes, such as social support and social control, can promote health behavior change. However, these processes are often neglected when studying health behavior change and designing interventions. Intervening on these social exchange processes using dyadic interventions may provide a promising approach to promote health behaviors. Objective This study aimed to investigate the effects of dyadic interventions to increase moderate-to-vigorous physical activity (MVPA) in romantic couples. Furthermore, we explored how the target, type, and timing of the interventions affect their effectiveness. Methods In total, 38 romantic couples (mean age 34.01, SD 11.03 y) were recruited through online advertisements and participated in a smartphone-based microrandomized trial over 55 days consisting of control and intervention phases. The fully automated dyadic interventions included a one-time psychoeducation intervention, weekly dyadic and collaborative planning, and dyadic just-in-time adaptive interventions (JITAIs). MVPA was measured through daily diaries and wrist-worn accelerometers. We used multilevel modeling to estimate the effect of the intervention phase and weighted and centered estimation for microrandomized trials to estimate the treatment effects of dyadic and collaborative planning, as well as the dyadic JITAIs. Results Participants indicated higher device-based (b=5.88, SE=3.04, t3665=1.93; P=.03) and self-reported (b=8.26, SE=3.88, t3904=2.13; P=.01) MVPA during the intervention phase compared with the control phase. Dyadic and collaborative planning did not increase device-based (b=6.31, SE=5.18; P=.12) but only self-reported (b=14.25, SE=5.16; P=.005) MVPA. However, the effects of the 2 kinds of planning on self-reported MVPA disappeared when additional covariates were included (b=0.14, SE=3.32; P=.48). Furthermore, the dyadic JITAIs targeting both the actor and the partner increased device-based (actor: b=11.17, SE=3.18; P<.001; partner: b=7.23, SE=3.60; P=.03) and self-reported (actor: b=17.34, SE=3.65; P<.001; partner: b=11.82, SE=4.10; P<.001) MVPA. However, the effects of the dyadic JITAIs targeting the actor disappeared for self-reported MVPA (b=2.20, SE=3.22; P=0.25) when additional covariates were included. Exploratory analyses revealed that different types and timings of dyadic JITAIs were differentially effective. Conclusions This study demonstrated the promising effects of dyadic interventions to promote MVPA and highlighted the importance of the target, type, and timing of the interventions. Further research should investigate the mechanisms underlying the effects of dyadic interventions on health behaviors. Trial Registration ISRCTN registry ISRCTN15673058; https://www.isrctn.com/ISRCTN15673058