Purpose This randomized investigation tested three different forms of personalized assignment to SMS-based (text messaging) physical activity (PA) advisors to promote PA among midlife and older Latino adults. Design An experimental study-within-a-trial (SWAT) from the On the Move randomized clinical trial. Setting San Francisco Bay Area, California. Sample Insufficiently active midlife and older Latino adults with overweight or obesity (n = 141). Intervention Participants receiving SMS-based PA coaching and were randomized to one of three PA advising conditions: Non-Personalized, Sex-Matched, or Choice-of-Sex. Measures PA was measured using the validated Community Health Activities Model Program for Seniors (CHAMPS) instrument, which assessed 12-month change in total weekly walking minutes, weekly moderate-to-vigorous physical activity (MVPA) minutes, and total weekly PA minutes. Analysis Repeated-measures linear regression analyses that employed fully conditional specification intention-to-treat principles were used to examine between-condition PA changes. Results Significant 12-month increases in all PA outcomes were observed across the three SMS conditions (all P < 0.0001). The Non-Personalized (B = −37.34, P < 0.0001) and Choice-of-Sex conditions (B = 20.25, P < 0.01) had greater increases in MVPA minutes than the Sex-Matched condition. In sex-specific subgroup analyses, the increased MVPA minutes for both the Non-Personalized and Choice-of-Sex conditions remained significantly greater than the Sex-Matched condition for men (Non-Personalized: B = −54.01, P = 0.004; Choice-of-Sex: B = 70.11, P < 0.0001), while only the increases in MVPA minutes for the Non-Personalized condition remained significantly greater than the Sex-Matched condition for women (Non-Personalized: B = −33.33, P = 0.0005; Choice-of-Sex: B = 6.62, P = 0.51). Conclusion Random assignment to either the Non-Personalized or Choice-of-Sex advisors yielded the greatest gains in 12-month MVPA minutes. Simplified digital messaging strategies that are adapted to participant preferences may enhance intervention effectiveness in under-resourced and similar populations with implications for scaling digital PA interventions.
Royer et al. (Wed,) studied this question.