Comparative Effectiveness of Clinical-Community Childhood Obesity Interventions: A Randomized Clinical Trial

Key Points

Question  What is the effectiveness of 2 clinical-community interventions in improving child body mass index z score and parent-report of their child’s health-related quality of life?

Findings  In this randomized clinical trial that included 721 children, there were significant improvements in body mass index z score in both intervention groups, as well as some aspects of quality of life. However, no statistically significant differences were found between the 2 intervention arms.

Meaning  Two interventions that included a package of high-quality clinical care for obesity and linkages to community resources resulted in improved family-centered outcomes for childhood obesity and improvements in child body mass index.

Importance  Novel approaches to care delivery that leverage clinical and community resources could improve body mass index (BMI) and family-centered outcomes.

Objective  To examine the extent to which 2 clinical-community interventions improved child BMI z score and health-related quality of life, as well as parental resource empowerment in the Connect for Health Trial.

Design, Setting, and Participants  This 2-arm, blinded, randomized clinical trial was conducted from June 2014 through March 2016, with measures at baseline and 1 year after randomization. This intent-to-treat analysis included 721 children ages 2 to 12 years with BMI in the 85th or greater percentile from 6 primary care practices in Massachusetts.

Interventions  Children were randomized to 1 of 2 arms: (1) enhanced primary care (eg, flagging of children with BMI ≥ 85th percentile, clinical decision support tools for pediatric weight management, parent educational materials, a Neighborhood Resource Guide, and monthly text messages) or (2) enhanced primary care plus contextually tailored, individual health coaching (twice-weekly text messages and telephone or video contacts every other month) to support behavior change and linkage of families to neighborhood resources.

Main Outcomes and Measures  One-year changes in age- and sex-specific BMI z score, child health-related quality of life measured by the Pediatric Quality of Life 4.0, and parental resource empowerment.

Results  At 1 year, we obtained BMI z scores from 664 children (92%) and family-centered outcomes from 657 parents (91%). The baseline mean (SD) age was 8.0 (3.0) years; 35% were white (n = 252), 33.3% were black (n = 240), 21.8% were Hispanic (n = 157), and 9.9% were of another race/ethnicity (n = 71). In the enhanced primary care group, adjusted mean (SD) BMI z score was 1.91 (0.56) at baseline and 1.85 (0.58) at 1 year, an improvement of −0.06 BMI z score units (95% CI, −0.10 to −0.02) from baseline to 1 year. In the enhanced primary care plus coaching group, the adjusted mean (SD) BMI z score was 1.87 (0.56) at baseline and 1.79 (0.58) at 1 year, an improvement of −0.09 BMI z score units (95% CI, −0.13 to −0.05). However, there was no significant difference between the 2 intervention arms (difference, −0.02; 95% CI, −0.08 to 0.03; P = .39). Both intervention arms led to improved parental resource empowerment: 0.29 units (95% CI, 0.22 to 0.35) higher in the enhanced primary care group and 0.22 units (95% CI, 0.15 to 0.28) higher in the enhanced primary care plus coaching group. Parents in the enhanced primary care plus coaching group, but not in the enhanced care alone group, reported improvements in their child’s health-related quality of life (1.53 units; 95% CI, 0.51 to 2.56). However, there were no significant differences between the intervention arms in either parental resource empowerment (0.07 units; 95% CI, −0.02 to 0.16) or child health-related quality of life (0.89 units; 95% CI, −0.56 to 2.33).

Conclusions and Relevance  Two interventions that included a package of high-quality clinical care for obesity and linkages to community resources resulted in improved family-centered outcomes for childhood obesity and improvements in child BMI.

Trial Registration  clinicaltrials.gov Identifier: NCT02124460

Childhood overweight and obesity place a significant burden on morbidity and quality of life. In the United States, the prevalence of childhood overweight and obesity appears to have plateaued and may even be decreasing among 2- to 5-year-old children as of 2012.¹ Yet, overall prevalence remains at historically high levels.^2,3 Clinical interventions to reduce obesity have been somewhat effective but are often limited in their effectiveness owing to the myriad social and environmental barriers that impede improvement in obesity-related behaviors.^4,5

An important, but often overlooked, aspect of interventions to improve obesity is the careful consideration of the socioenvironmental context in which decisions related to health behaviors are being made and in which behavior change is expected to occur. Neighborhood socioeconomic characteristics and environmental resources can significantly influence health behaviors and may contribute to childhood obesity in vulnerable populations.^6-10 Sophisticated geographic information systems methods and community mapping can provide community-level data on environmental resources and this information can assist in developing a tailored clinical-community intervention that could be adapted to an individual’s environment and needs.

Another underused approach to obesity management is to identify innovative strategies from positive outliers. Positive outliers are defined as individuals who have succeeded where many others have not to change their health behaviors, reduce their body mass index (BMI; calculated as weight in kilograms divided by height in meters squared), and develop resilience in the context of adverse built and social environments.^11,12 The premise of the positive outlier approach is that solutions to problems that face a community often exist among certain individuals within that community and that these successful members possess strategies that can be generalized and promoted to improve the outcomes of others in the same community. Such individuals could help guide intervention development for other families in their same neighborhoods who have struggled with behavior change.

We designed the Connect for Health Trial to leverage clinical and community resources to improve obesity and family-centered outcomes. The intervention was built on practices of positive outlier families as well as strategies recommended by a diverse group of stakeholders representing parents, children, pediatricians, and community members.^13,14 In this report, we summarize the main outcomes of Connect for Health.

Connect for Health was conducted in 6 pediatric practices of Harvard Vanguard Medical Associates (HVMA), a multispecialty group practice in Massachusetts. The study design, eligibility, and recruitment have been previously described.¹⁵ Briefly, we randomly assigned patients to 1 of 2 arms: (1) enhanced primary care or (2) enhanced primary care plus contextually tailored, individual health coaching. The enhanced primary care group served as the control arm, although these patients received some intervention previously incorporated into standard practice at HVMA. The primary outcomes included improvements in child BMI z score and family-centered outcomes. Study activities were approved by the Partners Health Care institutional review board. The trial has been recorded in clinicaltrials.gov. The full trial protocol is available in the Supplement.

Eligibility included the following: (1) child age 2 to 12.9 years, (2) BMI in the 85th or greater percentile, and (3) family not planning to leave HVMA within the study time frame. Recruitment occurred from June 2014 to March 2015; data collection ended March 2016. At visits with a child whose BMI was greater than the 85th percentile, clinicians received an alert in the electronic health record (EHR) with a link to refer the patient to the study. After receiving the referral, research assistants called parents to establish eligibility, obtain verbal consent, and complete a telephone survey. We then randomized participants and mailed an enrollment letter.

We randomized participants using 6 separate randomization lists, 1 for each practice. We organized the lists into blocks of 4 to maintain balance between the 2 study arms, and participants were randomized according to the order in which their consent was obtained. The lists were generated by the study biostatistician (E.J.O.) and maintained by the study project manager (C.H.).

All pediatric clinicians received a computerized clinical decision support alert during primary care visits identifying children with a BMI in the 85th or greater percentile, and 2 additional clinical decision support tools to assist in treating children with overweight or obesity.^15-17 Clinicians also gave parents a set of evidence-supported educational materials focusing on specified behavioral targets to support self-guided behavior change.¹⁷ The materials focused primarily on decreasing in screen time and sugar-sweetened beverages; improving diet quality; increasing moderate and vigorous physical activity; and improving sleep duration and quality. Based on our qualitative work with positive outlier families and feedback from our Parent and Youth Advisory Board, we also developed materials to promote social-emotional wellness.

Participants randomized to the enhanced primary care group were exposed to the clinical best practices described here. In addition, participants received monthly text messages that contained links to publicly available resources to support behavior change (eg, links to the Let’s Move! program). Participants also received a Neighborhood Resource Guide listing places that support healthy living in their community.

In the enhanced primary care plus coaching arm, families received individualized health coaching tailored to their socioenvironmental context. Four trained health coaches contacted families every other month for 1 year using telephone, videoconference (Vidyo), or in-person visits, according to parent preference. These contacts were approximately 15 to 20 minutes. Details of the coaching training and quality assurance have been previously described.¹⁵ Families also received twice-weekly text messages or emails,¹⁸ as well as mailings following each coaching session with educational materials to support families’ behavior change goals.

Health coaches used a motivational interviewing style of counseling and shared decision-making techniques^19,20 to provide family-centered care in addressing childhood obesity risk factors and management. At each contact, health coaches used an online community resource map developed for the study²¹ to identify resources within each family’s community that could support behavior change. In addition, health coaches offered families a 1-month free family membership to area YMCAs to encourage physical activity and community connections. Families were also invited to attend a healthy grocery shopping program led by Cooking Matters (https://cookingmatters.org/). To engage parents and children in setting behavior change goals, health coaches used a behavior change decision aid tool, developed by our study team, that helped families identify outcomes that mattered most to them and potential motivators for engaging in behavior change.

We obtained height and weight from children’s EHR at baseline and at 1 year. In routine practice standardized across all sites, medical assistants measured children’s weight, without shoes, using electronic calibrated scales, and height using a stadiometer. The primary outcome of age- and sex-specific BMI z score was calculated on the basis of the 2000 US Centers for Disease Control and Prevention growth charts. In addition, as a secondary outcome, we used Centers for Disease Control and Prevention–defined cutoffs to categorize BMI as normal (≥5th percentile to <85th percentile); overweight (≥85th percentile to <95th percentile); obesity (95% percentile to <120% of the 95th percentile); and severe obesity (≥120% of the 95th percentile).^1,22

Parent-reported outcomes were assessed using telephone surveys at baseline and 1 year. Parents reported their child’s health-related quality of life using the 4 subscales of the Pediatric Quality of Life 4.0.^23,24 We also assessed parental resource empowerment using the child weight management subscale of the Parent Resource Empowerment Scale.²⁴ The 5 items in the scale assessed parents’ perceived knowledge of, ability to access, comfort accessing, knowledge of finding, and ability to acquire resources related to child weight management. Response options were strongly disagree, disagree, agree, or strongly agree, worth 1 to 4 points, respectively. Items were averaged to create a summary score (range, 1-4). The Cronbach α was .87.²⁵

Using surveys at baseline, we obtained children’s race/ethnicity; parents’ educational attainment, height, and weight; and annual household income. At 1 year, we assessed the feasibility of the study and parents’ acceptance of and satisfaction with the intervention components. To assess unintended consequences, we also asked parents whether their participation in the program affected their satisfaction with their child’s health care services. To investigate prior trends in child BMI, we also obtained children’s height and weight 1 year prior to the baseline study visit (prebaseline) from the EHR.

Distributions of participant characteristics across the 2 study arms were analyzed using t and χ² tests and were found to be balanced at baseline. We performed multiple imputation using chained equations to impute missing outcomes for the 57 of 721 participants (8%) who did not have BMI outcomes at 1-year follow-up. In intent-to-treat analyses, we assessed the effect of the interventions on BMI z score, Pediatric Quality of Life 4.0 summary score, and the parent resource empowerment score using linear mixed effects repeated-measure models to account for clustering within participants over time. Participant baseline and 1-year measures were used as the outcome variables. Random intercepts were included in the models to account for correlation over time. The primary predictors were fixed effects for the intervention arm, time, and the time-by-intervention interaction term, which determined whether there was greater improvement in the enhanced primary care plus coaching group than the enhanced primary care group. Analogous ordinal logistic repeated-measures models were used to model the effect of the intervention on the odds of being in a lower BMI category at follow-up compared with baseline. All models included indicator variables for clinical site. All models were implemented using SAS version 9.4 (SAS Institute).

Clinicians referred 1752 children; we attempted to contact the parents of 1545 children to assess eligibility. We enrolled 721 children; 361 were randomized to the enhanced primary care group and 360 were assigned to the enhanced primary care plus coaching group (Figure 1). During the intervention period, 1 participant disenrolled from the enhanced primary care plus coaching arm, citing insufficient time for the study activities. At 1 year, we obtained BMI from 664 children (92%) and surveys from 657 parents (91%). Table 1 shows the characteristics of the study sample. The baseline mean (SD) age was 8 (3) years; 35.0% were white (n = 252), 33.3% were black (n = 240), 21.8% were Hispanic (n = 157), and 9.9% were other race/ethnicity (n = 71); 45.4% (n = 327) lived in households with annual incomes less than $50 000.

Table 2 and Table 3 show participants’ adjusted changes in BMI z score and in being in a lower BMI category from baseline to 1-year follow-up. In the enhanced primary care group, the adjusted mean (SD) BMI z score was 1.91 (0.56) at baseline and 1.85 (0.58) at 1 year, an improvement of −0.06 BMI z score units (95% CI, −0.10 to −0.02). In the enhanced primary care plus coaching group, the adjusted mean (SD) BMI z score was 1.87 (0.56) at baseline and 1.79 (0.58) at 1 year, an improvement of −0.09 BMI z score units (95% CI, −0.13 to −0.05). Although we observed slightly more improvement in BMI z score among the enhanced primary care plus coaching group, there was no statistically significant difference between the 2 intervention arms (−0.02 units; 95% CI, −0.08 to 0.03; P = .39). These results reflect multiple-imputed data for children with missing 1-year follow-up visits.

At 1 year, we found that 9.3% of children in the enhanced primary group and 11.6% of children in the enhanced primary care plus coaching group no longer had a BMI in the overweight or obese range. Overall, we observed higher odds of being in a lower BMI category than they were at baseline in both the enhanced primary care group (odds ratio, 1.18; 95% CI, 1.03-1.35) and the enhanced primary care plus coaching group (odds ratio, 1.23; 95% CI, 1.08-1.40).

We conducted post-hoc analyses to examine whether our observations of improved BMI z score in both intervention arms could be explained by an underlying temporal trend toward improvement. This was not the case. Among 560 children with BMI z scores available 1 year prior to baseline (prebaseline), at baseline, and at 1-year follow up, we found that BMI z score was increasing in the year prior to enrollment in the enhanced primary care group (0.23 units; 95% CI, 0.18-0.29) and the enhanced primary care plus coaching group (0.16 units; 95% CI, 0.11-0.22) and then decreased in both groups in the year following enrollment (Figure 2).

Table 4 shows changes in participants’ health-related quality of life and parental resource empowerment during the intervention. Parents in the enhanced primary care plus coaching group (1.53 units; 95% CI, 0.51 to 2.56), but not in the enhanced care alone group (0.65 units; 95% CI, −0.38 to 1.67), reported significant improvements in their child’s health-related quality of life. Parental resource empowerment increased in both intervention arms (Table 4). However, there were no statistically significant differences in either outcome between the 2 intervention arms.

Among participants in the enhanced primary care group, 91% of parents reported they received the study text messages and 53% were satisfied with their content. Approximately 60% reported receiving the Neighborhood Resource Guide and of those, 66% reported being very satisfied with its content.

For the enhanced primary care plus coaching group, 100% of participants reported receiving the study text messages and 72% were very satisfied with their content. Among the 360 participants in the enhanced primary care plus coaching group, 65% completed all 6 visits with a health coach; 96% reported receiving neighborhood resource information and 76% were very satisfied with the information. Eighty-one parents (23%) reported joining their local YMCA and 64 parents (18%) reported attending one of the Cooking Matters workshops.

Overall, 48% of participants in the enhanced primary care arm and 63% of participants in the enhanced primary care plus health coaching arm reported that participation in Connect for Health increased their satisfaction with their child’s health care services. Only 7 participants (1.1%) reported their participation in the program decreased their satisfaction with their child’s health care services and there were no differences across study arms.

In this randomized clinical trial, we found that 2 interventions that delivered enhanced primary care and leveraged clinical and community resources for childhood obesity support resulted in modest improvements in child BMI z score and greater resolution of elevated BMIs. However, while the magnitude of reduction in BMI z score was higher in the intervention group that additionally received interactive contextually tailored health coaching, the difference was not statistically significant compared with the group who were exposed to enhanced care alone. Both interventions led to improved family-centered outcomes. However, there were no statistically significant differences in either family-centered outcome between the 2 intervention groups. Overall, the intervention components were feasible to deliver, acceptable to parents, and did not have adverse effects on parents’ perceptions of their children’s health care services.

The Connect for Health Trial was designed with the hypothesis that the intervention group receiving both enhanced primary care and health coaching tailored to children’s community resources and social context would be more effective than the group receiving enhanced primary care alone. Yet, our findings did not support this hypothesis and there are several potential reasons. First, the enhanced primary care group was not a typical usual care control group. The practices where we delivered the study had already made several updates to their EHR to include clinical decision support tools and to provide families with educational materials for self-guided behavior change support. It would have been unethical to undo these practice changes once they were already established and after evidence supported their effectiveness in improving child BMI.¹⁷ Second, based on feedback from our Parent and Youth Advisory Board, we made the decision to add content on social and emotional wellness to existing parent educational materials and to provide passive information in the form of a booklet on neighborhood resources. Both of these enhancements could have further strengthened the effects of the control group on improving BMI. Third, it is possible that the number of contacts, frequency, or content of the health coaching provided in the enhanced primary care plus coaching group was insufficient to produce greater effects than the enhanced primary care group alone.

While our findings did not support the original hypothesis of a greater intervention effect among the group that was individually coached, our findings do suggest that both intervention groups experienced improved BMI. Without a traditional control group, our results could be attributed to temporal trends or regression to the mean. However, post-hoc analyses of BMI changes prior to and after enrollment in the trial suggest that the temporal trend was for BMI z score to continue increasing after enrollment. Thus, our results are unlikely due to secular trends but regression toward the mean may still be a possibility.

The magnitudes of effect on BMI z score in our study (eg, −0.06 to −0.09 units) are similar and only modestly higher than those previously summarized (−0.04 units) in a meta-analysis of brief interventions in primary care.²⁶ While these magnitudes of effect interrupted the increasing BMI trends in our population, questions remain about their clinical significance. There is currently a lack of direct evidence for any specific threshold for clinical significance.²⁷ An expert panel has suggested that a BMI z score reduction of 0.20 units is associated with clinically significant improvement.²⁸ Other studies suggest that changes of 0.15 BMI z score units led to more healthful cardiometabolic profiles.²⁷ As suggested by a recent evidence review of childhood obesity management, regardless of what the threshold of clinical significance will be determined to be, simply arresting gain in excess BMI likely constitutes a clinically important benefit for many children.²⁷

In addition to BMI, we examined family-centered outcomes of importance to parents and children. We found that parent-reported child health-related quality of life improved by 1.53 units among the enhanced primary care plus coaching group and appeared to be driven by large improvements in the psychosocial score of the Pediatric Quality of Life, comparably higher than previous pediatric obesity trials.²⁶ These effects were not greater than the enhanced primary care group. These findings suggest that the educational content delivered in both intervention arms related to social and emotional wellness, including content on stress reduction, positive thinking, and bullying, may have driven the observed improvements in child quality of life. Both interventions also improved parents’ perception of empowerment related to their child’s weight management, a novel family-centered measure that has been shown to drive changes in food-, physical activity–, and screen-related parenting among parents of children with obesity.^29,30

As in any study, this study is subject to potential limitations. First, as previously described, our post-hoc analyses showing an increase in BMI z score prior to intervention enrollment suggests either that we were successful in reversing an upward trend or that our results reflect regression to the mean. We are unable to rule out the possibility of the latter. Second, the study setting—a multisite delivery system with a robust EHR—may not be representative of smaller pediatric practices in the United States. However, as a relatively large medical group, HVMA is a typical primary care setting for many children, and meaningful use incentives are promoting increases in EHR adoption in both large and small pediatric practices.³¹ Thus, the Connect for Health interventions are likely to generalize to more pediatric settings in the future. Third, our intervention did not decrease the percentage of children with severe obesity. Previous studies have suggested that the magnitude of decreases in net daily energy intake necessary for children with severe obesity to achieve a healthy weight is considerably greater than the pediatric weight management that can be delivered in primary care–based interventions such as Connect for Health.^32,33 Our findings support the urgent recommendation for evidence-based, more-aggressive weight management approaches for children with severe obesity.³³

Two interventions that included a package of high-quality clinical care for obesity and linkages to community resources resulted in improved parent-reported outcomes for childhood obesity and improvements in child BMI. While individualized health coaching led to improvements in health-related quality of life, it did not have significantly greater effects on child BMI than enhanced primary care alone.

Accepted for Publication: April 12, 2017.

Correction: This article was corrected online August 7, 2017, to fix a data error in the Abstract, text, and Table 2, and to omit a reference.

Corresponding Author: Elsie M. Taveras, MD, MPH, Division of General Academic Pediatrics, Department of Pediatrics, Massachusetts General Hospital for Children, 125 Nashua St, Ste 860, Boston, MA 02114 (elsie.taveras@mgh.harvard.edu).

Published Online: June 5, 2017. doi:10.1001/jamapediatrics.2017.1325

Author Contributions: Dr Taveras had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Taveras, Marshall, Sharifi, Avalon, Fiechtner, Price, Sequist, Slater.

Acquisition, analysis, or interpretation of data: Taveras, Marshall, Sharifi, Fiechtner, Horan, Gerber, Orav, Sequist, Slater.

Drafting of the manuscript: Taveras, Sharifi, Avalon, Horan, Gerber, Price.

Critical revision of the manuscript for important intellectual content: Taveras, Marshall, Sharifi, Fiechtner, Horan, Orav, Sequist, Slater.

Statistical analysis: Gerber, Orav.

Obtained funding: Taveras, Sharifi, Sequist.

Administrative, technical, or material support: Marshall, Avalon, Fiechtner, Horan, Price, Sequist, Slater.

Study supervision: Taveras, Sharifi, Fiechtner.

Conflict of Interest Disclosures: None reported.

Funding/Support: This work was supported through award IH-1304-6739 from the Patient-Centered Outcomes Research Institute. Dr Taveras was also supported by K24 grant DK10589 from the National Institutes of Health. Dr Sharifi was supported by grant K12 HS 022986 from the Agency for Healthcare Research and Quality. Dr Fiechtner was supported by training grant T32 DK 007747 from the National Institute of Diabetes and Digestive and Kidney Diseases to the Division of Gastroenterology and Nutrition and grant K12 HS022986 from the Agency for Healthcare Research and Quality.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: All statements in this article, including its findings and conclusions, are solely those of the authors and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute or its board of governors or methodology committee.

Additional Contributions: We thank the health care professionals and staff at Harvard Vanguard Medical Associates for their ongoing collaboration in pediatric obesity research efforts. We thank the Connect for Health clinical research coordinators and health coaches for their assistance with the study. We thank the parents and children who serve on our advisory boards and offered their input to help shape the intervention. Last, we thank our community partners, Cooking Matters, and multiple Massachusetts YMCAs for making their resources available to study participants.