Background Digital interventions have shown promise in supporting healthy behaviors among university students; however, few interventions support simultaneous change across multiple health behaviors. Moreover, behavioral interventions are typically evaluated as a whole, making it challenging to disentangle the contribution of individual components to the overall effects. Objective This study estimated the effects of the components of a digital behavior intervention on alcohol, diet, physical activity, and smoking outcomes among university students in Sweden. Methods A double-blind randomized factorial trial with 6 two-level factors was conducted. University students in Sweden were proactively recruited through student health care centers and social media. Participants were eligible if they were aged 18 years or older and had at least one health behavior classified as unhealthy. The effects of 6 components were estimated: screening and feedback; goal-setting and planning; motivation; skills and know-how; mindfulness; and self-authored SMS text messages. Primary outcomes were weekly alcohol consumption and frequency of heavy episodic drinking, average daily fruit and vegetable consumption, weekly sugary drink consumption, weekly moderate-to-vigorous physical activity (MVPA), and 4-week point prevalence of smoking. Results A total of 1704 students were randomized. The effectiveness of individual and pairwise components was estimated using available data from 1118 (65.61%) participants at 2 months and 874 (51.29%) at 4 months, with sensitivity analyses conducted using imputed missing data. Most consistently, the evidence indicated that screening and feedback affected fruit and vegetable consumption (2-month mean difference 0.11, compatibility interval CoI –0.02 to 0.24; probability of effect POE 94.7% and 4-month mean difference 0.12, CoI –0.03 to 0.26; POE 94.4%), as did skills and know-how (2-month mean difference 0.19, CoI 0.06-0.33; POE 99.8% and 4-month mean difference 0.14, CoI 0.01-0.28; POE 96.9%). The combination of these 2 components was even more effective (2-month mean difference 0.30, CoI 0.11-0.48; POE 99.9% and 4-month mean difference 0.26, CoI 0.05-0.46; POE 99.4%). The motivation and mindfulness components, both individually and in combination, increased MVPA at 2 months (combined mean difference 78.0, CoI 28.3-128.2; POE 99.9%); however, this effect was not observed at 4 months. Combining screening and feedback with skills and know-how increased MVPA at 4 months (mean difference 60.1, CoI 3.6-116.5; POE 98.2%). Heavy episodic drinking was reduced at 2 months by screening and feedback (incidence rate ratio 0.87, CoI 0.74-1.02; POE 95.2%), and the effect was greater when combined with goal-setting and mindfulness. There was some evidence that the motivation component was harmful with respect to heavy episodic drinking and that self-authored SMS text messages were harmful with respect to sugary drink consumption. Conclusions We dismantled a complex digital multiple behavior intervention and examined it using a factorial design to provide novel insights into the effectiveness of the intervention’s different components. Both marginal and synergistic effects were observed across multiple behaviors, providing evidence regarding which components are most promising in complex interventions. These findings should be considered in light of the risk of bias introduced by attrition to follow-up, which was high in this effectiveness trial with low barriers to participation. Trial Registration International Standard Randomised Controlled Trial Number (ISRCTN) ISRCTN23310640; https://www.isrctn.com/ISRCTN23310640 International Registered Report Identifier (IRRID) RR2-10.1136/bmjopen-2021-051044
Åsberg et al. (Mon,) studied this question.