JMIR Publications

ABSTRACT

Background:

Wearable fitness technologies, like the Oura Ring, provide physiological metrics, like sleep and heart rate data, to a growing user base of young adults. However, these technologies and connected mobile applications do not directly measure young adults’ risky behaviors, such as alcohol use, that contribute to these metrics.

Objective:

For this RCT, we evaluated the acceptability, feasibility, and perceived effectiveness of a wearable personalized feedback intervention for alcohol reduction in young adults that integrated physiological and behavioral data (clinicaltrials.gov, NCT05090995).

Methods:

Participants (N = 60) wore the Oura Ring for 6 weeks and completed daily behavioral smartphone diaries. Only the Feedback group (n = 30) had full access to the Oura Ring mobile application and biweekly feedback reports from the study team. The Assessment group (n = 30) received one delayed feedback report. We used AI-driven convergent mixed methods to evaluate exit surveys and interviews, including descriptive and predictive analysis with surveys and natural language processing (NLP) and rapid qualitative analyses with interviews.

Results:

Half of participants (50.0%) were male, 81.6% were White, and they had a mean age of 22.02 (SD = 1.98). On a 100-pt scale, the overall program was described as highly acceptable (M(SD) = 84.17 (17.81)), feasible (77.92 < M < 91.67), and effective (70.42 < M < 75.83). Wearing the Oura Ring was highly acceptable (76.25 < M < 90.83) and feasible (68.33 < M < 93.52). Along with the mobile app, the Oura Ring was described as moderately effective for the Feedback group (52.88 < M < 67.00). Completing the smartphone diaries was moderately acceptable (M(SD) = 59.17 (25.20)), moderately highly feasible (64.17 < M < 82.50), and highly effective (interview descriptions). The feedback reports were highly acceptable (75.00 < M < 92.67), feasible (70.26 < M < 84.38), and effective (69.64 < M < 88.36). Feedback participants had high adherence: using the mobile app daily, and 80% read all 3 biweekly feedback reports. Per NLP, the most common topics among Feedback participant interviews related to their behavior change due to multiple intervention components: Multimodal general change (n = 9) and Multimodal sleep/alcohol change (n = 6). These contrasted to the most common topics from Assessment participants that pertained to pre-change learning: Self-guided report use (n = 6) and Multimodal insights, good sleep (n = 6). Therefore, Feedback participants may have reported higher behavior change post-intervention.

Conclusions:

Commercial fitness wearables that integrate behavioral data may promote readiness to change drinking in young adults who are generally unconcerned about risky behaviors. Wearable feedback interventions integrating physiological and behavioral data may be acceptable, feasible, and effective for young adult alcohol reduction. Clinical Trial: clinicaltrials.gov, NCT05090995