Associations of device-measured and self-reported physical activity with alcohol consumption: Secondary analyses of a randomized controlled trial (FitForChange).

BACKGROUND: Physical activity (PA) is increasingly used as an adjunct treatment for alcohol use disorder (AUD). Previous studies have relied on self-report measures of PA, which are prone to measurement error. In the context of a randomized controlled trial of PA for AUD, we examined: (1) associations between device-measured and self-reported PA, (2) associations between PA measurements and alcohol use, and (3) the feasibility of obtaining device-measured PA data in this population.

METHOD: One-hundred and forty individuals with clinician-diagnosed AUD participated in a 12-week intervention comparing usual care (phone counselling) to yoga-based exercise and aerobic exercise. Device-measured PA (Actigraph GT3x), self-reported PA (International Physical Activity Questionnaire) and alcohol consumption (Timeline Follow Back Method) were assessed before and after the trial. Effects of the interventions on PA levels were assessed using linear mixed models.

RESULTS: In total, 42% (n=59) of participants returned usable device-measured PA data (mean age= 56±10 years, 73% male). Device-measured and self-reported vigorous-intensity PA were correlated (β= -0.02, 95%CI= -0.03, -0.00). No associations were found for moderate-intensity PA. Compared to usual care, time spent in device-measured light-intensity PA increased in the aerobic exercise group (∆= 357, 95%CI= 709, 5.24). Increases in device-measured light-intensity PA were associated with fewer standard drinks (∆= -0.24, 95%CI= -0.03, -0.44), and fewer heavy drinking days (∆= -0.06, 95%CI=-0.01, -0.10).

CONCLUSION: Increases in light-intensity/habitual PA were associated with less alcohol consumption in adults with AUD. Self-reported PA data should be interpreted with caution. Incentives are needed to obtain device-measured PA data in AUD populations.