The STAR-MAMA RCT: Bilingual Mobile Health Coaching for Postpartum Weight Loss

Abstract

Introduction: Gestational diabetes and overweight during pregnancy are associated with future type 2 diabetes. Postpartum weight loss can reduce diabetes risk. However, effective interventions for postpartum weight loss are lacking, in particular for Latina populations, despite their disproportionate burdens of gestational diabetes, overweight, and diabetes.

Study design: This was a community-based RCT.

Setting/participants: Researchers recruited pregnant individuals with gestational diabetes or BMI>25 kg/m² from safety-net health care settings and Women, Infants, and Children offices in Northern California in 2014-2018. Of 180 individuals randomized to intervention (n=89) or control (n=91), 78% identified as Latina, 61% were primarily Spanish speaking, and 76% perceived their diabetes risk to be low.

Intervention: The intervention consisted of a 5-month postpartum telephone-based health coaching intervention delivered in English or Spanish.

Main outcome measures: Data were collected through surveys at enrollment and 9-12 months after delivery and chart review up to 12 months after delivery. The primary outcome, weight change from prepregnancy to 9-12 months after delivery, was compared between the groups, overall and within strata defined a priori according to language (Spanish or English) and diabetes risk perception (none/slight or moderate/high).

Results: The intent-to-treat estimated intervention effect was +0.7 kg (95% CI= -2.4 kg, +3.8 kg; p=0.67). In stratified analyses, intervention effects remained nonsignificant but varied in direction: effects were favorable among English speakers and those with higher perceived diabetes risk, and unfavorable among Spanish speakers and those with lower perceived risk. Analyses were conducted in 2021-2022.

Conclusions: A postpartum health coaching intervention, designed for low-income Latina women at increased risk for diabetes, did not reduce postpartum weight gain. Intervention effects were nonsignificantly more favorable among English speakers versus Spanish speakers, and among those who perceived their diabetes risk to be high versus low.

Trial registration: This study is registered at www.

Clinicaltrials: gov NCT02240420.