Comparing Two Programs to Help Families Living in Medically Underserved Areas Address Childhood Obesity [Internet]

Excerpt

Background: Although there is a large body of literature documenting the efficacy of family-based childhood obesity treatment interventions, there is little evidence of their systematic translation into regular practice, particularly in health-disparate regions.

Objectives: To address this research-practice gap, a community-based participatory research approach, guided by the RE-AIM (reach, effectiveness, adoption, implementation, and maintenance) framework, was used in the medically underserved Dan River region (i.e., south central Virginia). The primary aim was to determine the relative effectiveness of 2 family-based childhood obesity interventions (Family Connections and iChoose) based on changes to the primary outcome: reduced child body mass index (BMI) z scores at 6 months. Relative to Family Connections children, it was hypothesized that iChoose children would achieve significantly higher reductions in BMI z scores. Secondary effectiveness aims included changes in parent weight and child and parent health behaviors. Additional secondary aims, focused on RE-AIM, included reach, implementation fidelity and costs, individual-level maintenance of anthropometric and behavioral changes, and potential organizational capacity and program maintenance.

Methods: A Community Advisory Board and a Parent Advisory Team collaboratively planned and implemented the trial. From 2017 to 2020, 3 consecutive cohorts of families were enrolled in the trial, and eligibility requirements were at least 1 eligible child with a BMI percentile ranking of 85 or higher, aged between 8 and 12 years. Recruitment occurred through medical chart review/referral and an open referral system across all cohorts; however, child age eligibility was extended to 5 to 12 years for cohort 3 because of lower-than-anticipated recruitment in cohorts 1 and 2. Children were randomly assigned to iChoose (n = 70) or Family Connections (n = 69). iChoose included family sessions, interactive voice response (IVR) support calls, physical activity sessions, and child newsletters over 6 months for a total potential of 76 contact hours. iChoose also had a 6-month supported maintenance period, with 12 IVR calls and child newsletters. Family Connections included 2 parent sessions and 10 IVR support calls over 6 months for a potential 5 contact hours, with no additional support during the 6-month maintenance period. Child BMI z scores, parent BMI values, child and parent health behavior outcomes, and intervention engagement were assessed. Implementation aims were measured through observed and self-reported fidelity checklists and community-level costs. Maintenance aims were determined through child and parent changes in outcomes at 12 months. COVID-19 restricted in-person anthropometric data collection for cohort 3. Maintenance at the organizational level was analyzed through qualitative data focused on community capacity for sustained implementation. The 6-month effectiveness data were analyzed using an intention-to-treat Heckman selection model. Also, the 6-month and 12-month effectiveness data were analyzed using completers at either time point (ie, mixed-models repeated-measures analysis). Descriptive, nonparametric, and parametric analytic methods were used to examine reach, engagement, and implementation. An inductive-deductive approach was used for qualitative data analysis.

Results: Overall study reach was 8% of total referrals (n = 1802) and 18% of eligible referrals contacted during the screening process (n = 790). Enrolled children (n = 139; mean age 10.1 ± 1.7 years, 27% who were overweight, 73% with obesity, 45% Black, 62% Medicaid recipients) were randomly assigned to iChoose (n = 70) or Family Connections (n = 69). Enrolled children had significantly greater weight status, were older, and were less likely to be White than children whose parents declined enrollment. At 6 months and 12 months, respectively, 63% and 59% of iChoose and 84% and 83% of Family Connections participants completed study outcome assessments. Changes in BMI z scores from baseline to 6 months were not statistically significant within iChoose (BMI z score = 0.03 [95% CI, −0.13 to 0.19]) or Family Connections (BMI z score = 0.00 [95% CI, −0.16 to 0.16]) or between study arms (BMI z score = 0.03 [95% CI, −0.17 to 0.23]). Null findings were found for completer analysis models and for parents' BMI 6-month changes. Among parents, an average of 6.9 unhealthy days (95% CI, 0.6-13.2) differences were observed between study arms over time in favor of improvement for Family Connections (−3.9 vs 3.1 days). No other 6-month secondary self-reported outcomes were statistically significant. Relative to iChoose, Family Connections participants completed a significantly higher percentage of family sessions and IVR calls, but iChoose participants received a significantly higher exposure to intervention content (average 533 minutes [95% CI, 401-666]) than Family Connections (average 122 minutes [95% CI, 103-141]). At 12 months, between–study arm effects were not significant for child BMI z score or parent BMI. Individual-level self-reported maintenance outcomes revealed no consistent patterns between study arms. Both iChoose (group classes = 98%, physical activity classes = 87%) and Family Connections (group classes = 99%) had high implementation fidelity. Also, IVR calls were delivered with 100% fidelity in both programs. The cost of iChoose ($68 289) across cohorts was approximately twice that of Family Connections ($32 533). Organizational-level maintenance analysis suggested a need to develop systems to support recruitment and engagement.

Conclusions: Contrary to the hypothesis, there was no significant difference in child BMI z scores between the 2 family-based childhood obesity treatment interventions. Likewise, neither yielded significant improvements from baseline in the primary outcome or secondary BMI-related effectiveness outcomes and only minimal differences in secondary health behavior outcomes. Yet, both interventions were delivered with high fidelity. Relative to iChoose, descriptive data indicated higher retention, better engagement, and lower costs for Family Connections. This may tentatively suggest that Family Connections is a better fit for a rural, lower-resourced region, like the Dan River region; however, regardless of resource intensity of either intervention, insights from key stakeholders stress the importance of strengthening recruitment and engagement systems before adoption and future sustainability of any family-based childhood obesity intervention. Additionally, future efforts in this region should explore asynchronous and technology-based programmatic options, consider integrating more intensive self-monitoring strategies, and consider engaging clinical providers at the point of care for families.

Limitations: Given the challenges with recruitment and retention in a small, medically underserved area, the interpretation of the study results is uncertain. Caution should be applied when interpreting them.

Note: Portions of this report have been published in Contemporary Clinical Trials, Pediatric Obesity, and Frontiers in Public Health. Per our journal author rights, we credit the manuscripts below and repurpose our own work for this PCORI report.

1. Zoellner JM, You W, Hill JL, et al. . A comparative effectiveness trial of two family-based childhood obesity treatment programs in a medically underserved region: rationale, design & methods. Contemp Clin Trials. 2019;84:105801. doi:10.1016/j.cct.2019.06.015

2. Zoellner JM, You W, Hill JL, et al. . Comparing two different family-based childhood obesity treatment programmes in a medically underserved region: effectiveness, engagement and implementation outcomes from a randomized controlled trial. Pediatr Obes. 2022;17(1):e12840. doi:10.1111/ijpo.12840

3. Brock DJP, Estabrooks PA, Yuhas M, et al. . Assets and challenges to recruiting and engaging families in a childhood obesity treatment research trial: insights from academic partners, community partners, and study participants. Front Public Health. 2021;9:631749. doi:10.3389/fpubh.2021.631749