Family Intervention to Prevent Childhood Obesity

Research Design The proposed research design is a prospective randomized controlled trial to evaluate the effectiveness of a family-based childhood obesity intervention that uses a "hybrid" approach of in-person and mobile phone components compared to usual WIC care. Approximately, 260 families (130 in each of the groups and oversampled for a 30% attrition rate) will be recruited from WIC centers in Los Angeles County in the neighborhoods of East Los Angeles, where >95% of clients are of Mexican descent, and from MacArthur Park ("Little Central America") near downtown Los Angeles, where > 80% of clients are of Central American descent. These WIC centers which serve over 5,000 individuals per month (4,000 Latino children) will ensure a sufficient number of Latino participants and ethnic diversity within the Latino population.

Recruitment Eligibility to participate will be open to adult caregivers of a young child who meet study inclusion and exclusion criteria. Eligible caregivers will be randomly assigned to the intervention or usual care arm. All recruitment, enrollment, and incentive strategies will be vetted by IRB.

Study Randomization Study randomization will occur after baseline data collection. A computer-generated permuted block randomization scheme with a block size of 4 will be used to ensure balanced treatment allocation within strata based on child overweight and obesity status, and caregiver structure. These caregiver structure groups will be: (1) two caregivers in same household; (2) two caregivers in different households; and (3) single parent without a second caregiver to participate (although anticipate a very small percentage in this group as neighborhood demographic data indicate >70% two-caregiver households). Single parents will be reminded that former spouses, extended family, friends, and other caregivers are eligible and encouraged to participate.

Intervention Group The intervention will include in-person and mobile phone components focusing on Latino caregivers of young children, to support evidence-based and age appropriate dietary, media-viewing, and physical activity practices among 2- to 5-year old children, in order to decrease ethnic disparities in childhood obesity. The intervention is 4-week in-person sessions plus mobile phone component, followed by 2-motnsh of mobile phone "booster" messages to reinforce content.

Usual Care Control Group After enrollment into the study and completion of baseline data, caregivers will be randomly assigned into the intervention or control arm of the study. Both groups of caregivers will continue to receive ongoing usual care through WIC. WIC "usual care" for caregivers of 2- to-5-year-olds includes: (1) monthly vouchers for nutritious food; (2) individual nutrition counseling at least twice per year; (3) nutrition education two times per year (1:1 counseling or online modules); and (4) referrals to family services.

Data Collection and Management Research assistants will be blinded to participation condition to minimize bias in outcome assessments. All surveys will be available in English and Spanish, have been used with Latinos, and have a high degree of internal consistency and validity. All data will be collected using REDCap software on tablets. Families will come to WIC centers or choose a home visit for data collection. Evaluation of the intervention's effectiveness will include receiving the intervention and the repeated factor of time (baseline, 1-, 6-, 12-months). The Data Coordinating Team in collaboration with the PI will oversee and instruct the study personnel (e.g. project manager and research assistants) on data collection training, be responsible for conducting the study randomization, and monitor data entry, and quality assurance. Study data will be collected and managed using REDCap.

Analysis of Primary Outcome:

Child's BMI. BMI scores for each child will be calculated based on the average height and weight from two measurements made at each time point (baseline, 1-, 6-, and 12 months post baseline). Changes in the child's BMI over time will be examined using a mixed effects linear regression model with family random intercept and nested random slope for time period of measurement. Additionally, fixed effects for study arm (intervention vs usual care) and the interaction of group-by-time will be specified. A significant (p<0.05) interaction term will indicate that the changes in BMI over time are reliably different between the two groups. Marginal estimates will be used to characterize this interaction with respect to the time specific difference between the groups; with notable comparisons made for the changes by 6 months (primary outcome) and 12 months (secondary outcome). Models will be fit as both unadjusted and adjusted for covariates. The random effect for time will be specified to allow for individual differences in the BMI change over time. We will utilize an intention-to-treat principle in these models, including participants with any missing values at the study time points.