ABSTRACT
OBJECTIVE: To conduct a randomized controlled trial examining the effects of a social network intervention on health. PARTICIPANTS AND METHODS: The Microclinic Social Network Program randomized controlled trial (implemented from June 1, 2011, through December 31, 2014) delivered weekly social-health classroom interventions for 9 to 10 months vs standard of care. Longitudinal multilevel analyses examined end-of-trial and 6-month post-intervention outcomes. Social network effects were estimated via a novel social induction ratio. RESULTS: We randomized 494 participants, comprising 27 classroom clusters from five neighborhood cohorts. Compared with controls, the intervention showed decreased body weight -6.32 pounds (95% CI, -8.65 to -3.98; overall P<.001), waist circumference -1.21 inches (95% CI, -1.84 to -0.58; overall P<.001), hemoglobin A1c % change -1.60 (95% CI, -1.88 to -1.33; overall P<.001), mean arterial blood pressure -1.83 mm Hg (95% CI, -3.79 to 0.32; overall P<.01), borderline-increased high-density lipoprotein cholesterol 1.09 (95% CI, 0.01-2.17; P=.05; overall P=.01). At 6 months post-intervention, net improvements were: weight change 97% sustained (P<.001), waist circumference change 92% sustained (P<.001), hemoglobin A1c change 82.5% sustained (P<.001), high-density lipoprotein change 79% sustained (overall P=.01), and mean arterial blood pressure change greater than 100% sustained improvement of -4.21 mm Hg (P<.001). Mediation analysis found that diet and exercise did not substantially explain improvements. In the intent-to-treat analysis of social causal induction, the weight-change social induction ratio (SIR) was 1.80 for social-network weight change-meaning that social networks explained the greater weight loss in the intervention than controls. Furthermore, we observed an even stronger weight-loss SIR of 2.83 at 6 months post-intervention. CONCLUSION: Results show intervention effectiveness for improving health in resource-limited communities, with SIR demonstrating that social-network effects helped induce such improvements. TRIAL REGISTRATION: Clinicaltrials.gov Identifier NCT01651065.
