Treating Children and Adolescents With Obesity: Characteristics of Success

Abstract

Background:

Factors related to clinically meaningful outcomes for pediatric patients seeking care for severe obesity are not well known. Examining patient-level and program-level characteristics related to success may inform future care.

Objectives

: To determine factors associated with a clinically significant reduction in weight status measured by %BMIp95 after 6 months of treatment.

Study Design:

This is a retrospective study of youth 5–17 years of age seeking multicomponent weight management care to determine if patient characteristics, treatment recommendations, reported adherence, and additional program-affiliated class participation are associated with 6-month change in %BMIp95.

Results:

Among 170 children with obesity, higher reductions in %BMIp95 were seen in those with medium-high dietary adherence compared to low-none (−10.8 vs. −4.0, p = 0.002). Post hoc analysis showed higher dietary adherence among those with private insurance than public insurance (59% vs. 41%, respectively, p = 0.04).

Conclusion:

Children receiving multidisciplinary multicomponent weight management, who achieve clinically meaningful outcomes, are more likely to be adherent to dietary recommendations regardless of the type. Further study is needed of how best to address social determinants of health to improve dietary adherence. Clinical Trial Registration Number: NCT02121132.

Introduction

The increasing prevalence of obesity and severe obesity in children and adolescents in the United States¹ is a critical public health concern, given its well-documented association with medical and psychological complications during both childhood and adulthood.^2–5 Treatment guidelines recommend a staged approach to care, which includes family-based behavioral and dietary modification, and advanced therapies for those with more severe obesity or with comorbidities and other obesity-related health complications. Those not responsive to management in the primary care setting may be referred for more intensive treatment provided by a pediatric weight management program (PWMP).^6–8 Within these PWMPs, there is little “real-world” data that report programmatic characteristics, which lead to clinically successful outcomes.

Most referred children and adolescents to PWMPs have severe obesity and/or multiple complications of obesity.⁹ Although many PWMPs use a multicomponent care model, program design, staffing, and care delivery may vary from one PWMP to another.^10,11 In addition, some PWMPs offer different “program tracks,” which may vary by frequency of clinic visits and program structure to best meet the needs of patient-families.¹⁰

These findings were based on results of a Program Profile Survey administered to participating multicomponent PWMPs participating in the Pediatric Obesity Weight Evaluation Registry (POWER). POWER is a consortium of 31 sites from across the United States that contribute to a centralized data repository of demographic and clinical data from multicomponent PWMPs.¹¹ The goals of POWER are to provide longitudinal tracking of changes in weight status and other health outcomes in a “real-world,” tertiary care pediatric weight management (PWM) setting, and conduct quality improvement to optimize care for obesity-related medical and psychological complications.

Favorable PWMP outcomes, specific to improved weight status and its association with improvement in other obesity-related health measures, are not well defined due to changes in the use of BMI outcome metrics.¹² Historically, BMI z-score has been widely used as the accepted metric to evaluate change in weight status over time, for children and adolescents with overweight and obesity. BMI z-score is defined as BMI transformed into the number of standard deviations (SDs) above or below the mean population BMI for age and sex.¹³ Using BMI-SD-scores (BMI-SDS), clinically meaningful outcomes (e.g., improvement in cardiovascular risk factors, such as hypertension, hypertriglyceridemia, and low high-density lipoprotein [HDL]-cholesterol) have been reported with reductions of greater than 0.25 BMI-SDS based on a large European registry cohort with mostly severe obesity.¹⁴

However, growing evidence and consensus among experts in the field of PWM suggest that evaluating change in BMI z-score may lead to erroneous conclusions, particularly when applied to those in early childhood and older adolescents with severe obesity.^5,13,15–17 Hence, given the POWER registry comprises patient data with very high BMI values, changes are expressed relative to the CDC 95th percentile as we evaluate obesity interventions.

To the best of our knowledge, clinically meaningful changes using %BMIp95 have not been widely reported. However, based on the aggregate POWER dataset (N = 6454 patients), we reported a 5-percentage point decrease in %BMIp95 was associated with improvement in cardiometabolic risk factors. In addition, patients of older age, increased severity of obesity, and Hispanic race/ethnicity were associated with better improvement in %BMIp95.⁹

Given the diversity in care delivery among PWMPs and interest in other patient- and treatment-level characteristics beyond the existing aggregate POWER dataset, this study included an expanded set of patient- and treatment-level characteristics for a subset of patients enrolled across several POWER sites. The aims of this study were to determine the association of clinically significant reduction in %BMIp95 with a broader set of (1) patient characteristics, (2) treatment-level characteristics, and (3) participation in program-associated group classes offered outside of clinical care.

Methods

POWER Site Selection

A retrospective study of POWER enrolled patients 5–17 years of age seeking multicomponent PWM care was conducted among POWER sites. Site participation was voluntary among the 29 POWER sites active during Cycle 3 (2018–2020), based on their interest and resources to conduct manual retrospective reviews of medical records for selected POWER patients at their site. In addition, participating sites were selected based on having enough enrolled patients in POWER, who met the inclusion and exclusion criteria for the analyses of this study. As such, only the highest enrolling centers. All participating sites (N = 5) obtained Institutional Review Board (research review committee) approval before participation in the study.

Subject Selection

Eligible subjects for the study were identified by the POWER data coordinating center (DCC) based on the following inclusion criteria: BMI ≥95th percentile for age and sex based on CDC growth curves,¹⁸ ≥5 years of age and completed an initial assessment and at least one follow-up visit with a medical provider over at least 10 months of treatment. We then used a stratified sampling to be sure that we included representation of those with varied changes in %BMIp95 at 6 months within each site. These were determined based on previous data demonstrating that reductions larger than 5% can lead to improved metabolic biomarkers.⁹

The stratification was large (≥15 percentage point reduction in %BMIp95), moderate (5–14 percentage point reduction in %BMIp95), and no response (increase in %BMIp95). Eligible subjects who did not fall within one of the three categories were not included in the sample. To ensure an even distribution of patients with different categories of treatment response among the sites, up to 12 subjects from each category within a site were included in the study and they were randomly selected if more than 12 met a response category. A total of 170 subjects (34 per site) were included in the study. Five sites had sufficient samples and were willing to participate in this study. They each received a list of subject ID numbers for conducting the retrospective medical record chart review of additional data not collected among all POWER sites.

Subject Variables

Baseline characteristics of selected patients were available in the existing POWER data set. These included age, sex, race, ethnicity, health insurance coverage, and weight status classification [Obesity—Class 1 (OB-1; between 100% and 119% of %BMIp95); Severe Obesity—Class 2 (SO-2; BMI between 120% and 139% BMIp95 or BMI between 35 and 39 kg/m², whichever was lower); and Severe Obesity—Class 3 (SO-3; BMI of ≥ of 140% %BMIp95 or BMI of ≥40, whichever was lower)].¹⁸ Additional baseline variables manually extracted from the subject's medical record included educational setting, academic status, living arrangement for patient, and primary household members. In addition, related medical diagnoses and selected medications used for obesity management, documentation of snoring, lack of satiety/excessive hunger, and sneaking food at night were collected. The outcome at 6 months included weight status changes from baseline to 6 months, with %BMIp95 treated as a continuous variable.

Treatment Variables

Lifestyle treatment recommendations specific to diet and physical activity advised by the provider at baseline and follow-up were collected. Categories for dietary recommendations are listed in Table 1 based on common, evidence-based dietary approaches used for the management of pediatric obesity.¹² When other nutrition-related behavioral strategies were documented, such as changes in the home food environment, patient involvement in home meal preparation, keeping a 7-day food log, reading nutrition labels, trying new recipes, or a combination of multiple dietary strategies, these were grouped in the “other” category. Physical activity recommendations were characterized by the level of step count/day, and/or duration (minutes/day) and frequency (days/week) of moderate-vigorous activity, as shown in Table 1. When other types of physical activity were documented, such as dancing, weight training, jumping rope, boxing, swimming, walking the dog, and yoga, these were grouped in the “other” category.

Table 1.

Treatment Recommendations

Dietary recommendation	Physical activity prescription	Additional program-sponsored treatment encounters (duration: minutes or days)
Healthy nutrition guidelines Carbohydrate restricted Establish regular meal/snack pattern Nutrient balanced Low/reduced glycemic load Calorie restricted Time-limited eating plan Other	Moderate-Vigorous Physical Activity (minutes/d) 30 60 90 Other (days/week) 1–2 3–4 5–7 Step Count (steps/d) 3000 5000 7000 9000 or more Other	Group Nutrition or Group Cooking Class Group exercise/training Combination Group (Nutrition, Support and/or Exercise) Camp Individual counseling sessions offered by site provider Other

Adherence to lifestyle recommendations was determined by provider documentation during follow-up visits. Given the varied approaches to measures of adherence and goal setting language used between sites and clinicians, this rating was subjectively determined based on documented language, inferring the degree of adherence to patient goals from the previous encounter. These were rated as none/almost none, some, almost all/all, or unknown for both dietary and physical activity goals.

We used broad categories for adherence to avoid over interpreting an inherently subjective measure. Longitudinal adherence was then defined dichotomously as medium-high adherence (at least 50% of visits rated as “almost all/all”) or low-no adherence (<50% of visits rated as “almost all/all”). Visits marked as “unknown” were not included in the calculation described above. In addition, participation in program-associated group classes on nutrition and physical activity was collected. This was analyzed dichotomously as participating for any length of time or not at all.

Procedures

Data dictionary definitions were created to guide the manual collection of additional data. These were shared among the site personnel responsible for this task. All manually extracted data elements were entered into a password-protected REDCap^19,20 electronic data capture system housed by the POWER DCC. Sites only had access to their own site-specific data.

Statistical Analyses

To identify factors associated with change in %BMIp95, univariate mixed effects models were run with “site” as a random factor. Median and interquartile range (IQR) were presented. Based on univariate analyses results, factors with a p-value <0.10 for the association with change in %BMIp95 were included in a multivariable mixed effects model with site as a random factor that included baseline weight status, age group, and sex as covariates. Least square means and standard errors are presented for multivariable results. All tests were done at the two-sided 5% level of significance. The Tukey-Kramer p-value adjustment was used for factors that required multiple comparison adjustment (e.g., baseline weight status).

A post hoc analysis was performed to determine factors associated with dietary adherence using the Chi-Square test with Benjamini-Hochberg adjustment for multiple group comparisons, as needed.

Results

Five POWER sites participated in the study from the central region of the United States and included Children's Hospital of San Antonio, San Antonio, TX; Cincinnati Children's Hospital Medical Center (CCHMC), Cincinnati, OH; Children's Mercy Kansas City, Kansas City, MO; Children's Hospital of Illinois (CHOI), Peoria, IL; and Helen DeVos Children's Hospital, Grand Rapids, MI.

Baseline demographics of subjects (N = 170) comprised a median (IQR) age of 10.5 years (8.0–13.0), 55% female, 64% White, 15% African American, 71% with public health insurance, and 39% Hispanic. Participant weight status frequencies included 33% with OB-1, 34% with SO-2, and 34% with SO-3 (Table 2).

Table 2.

Demographics and Changes in Percentage of the 95th Percentile for BMI

Demographic factor (N = 170)	n (%)	Change in %BMIp95, median (Q1, Q3)	p ^*
Age at baseline
Age categories
5–11 Years	103 (61)	−7.86 (−17.9, 7.5)	0.16
12–17 Years	67 (39)	−10.0 (−19.2, −5.6)	0.16
Sex
Female	93 (55)	−7.94 (−17.8, 5.8)	0.28
Male	77 (45)	−9.81 (−20.0, 5.7)
Ethnicity
Hispanic	66 (39)	−9.41 (−20.2, 2.6)	0.24
Not Hispanic	99 (58)	−8.61 (−16.1, 9.1)
Unknown/Not Reported	5 (3)	−9.13 (−14.9, −6.2)
Race
White	109 (64)	−9.04 (−20.0, 3.4)	0.84
Black	26 (15)	−9.24 (−13.9, 1.9)
Other	34 (20)	−8.31 (−16.6, 6.3)
Unknown	1 (0.6)	—
Insurance
Private	46 (27)	−9.06 (−15.9, −5.6)	0.60
Public	120 (71)	−8.96 (−19.4, 5.9)
Self-pay/None	2 (1)	−17.84 (−27.8, −7.8)
Unknown	2 (1)	1.95 (−9.1, 13.0)
Weight status at baseline
Obese—Class 1	56 (33)	−7.09 (−14.5, 5.9)	0.12
Severe Obese—Class 2	57 (34)	−9.08 (−16.6, 3.4)
Severe Obese—Class 3	57 (34)	−11.69 (−21.1, 2.7)
Site
1	34 (20)	−8.66 (−16.6, 10.7)	0.13
2	34 (20)	−8.82 (−20.0, 7.8)
3	34 (20)	−7.74 (−11.2, 8.3)
4	34 (20)	−10.13 (−16.4, 7.2)
5	34 (20)	−12.37 (−23.0, 1.9)

p-Value for univariate analysis using the ANOVA mixed effects with site as a random factor for all comparisons except site. p < 0.05 is significant.

%BMIp95, percent of the 95th percentile for BMI; ANOVA, analysis of variance.

Univariate analyses demonstrated that dietary adherence, irrespective of dietary strategy used, was associated with improved %BMIp95 outcomes. Analysis of which strategy was most effective was not possible to ascertain because of varied and evolving serial or simultaneous application of these recommendations. No other characteristic studied, including sex, race, ethnicity, or insurance type, was significantly associated with %BMIp95 changes. Baseline factors such as household composition, type of education, food-seeking behaviors, lack of satiety, and presence of snoring were also not significantly associated with %BMIp95 reduction. Medical diagnoses, pharmacotherapy for weight loss treatments, and other medications (Supplementary Table S1) demonstrated no significant association with %BMIp95 reductions, with the exception of those without an asthma diagnosis, which trended toward greater %BMIp95 reductions (−9.6 vs. −7.8; p = 0.06) (Table 3). Given the small number of patients reporting participation in program-associated group classes, these data could not be analyzed.

Table 3.

Changes in Percentage of the 95th Percentile for BMI by Patient Baseline Characteristics, Adherence Status, Diagnoses, and Medications

Category	n	Change in %BMIp95, median (Q1, Q3)	p ^a
Baseline status
Education setting			0.93
Private school	11	−7.63 (−9.7, 5.9)
Public school	131	−9.18 (−19.2, 5.7)
Home school	9	−10.5 (−14.9, 8.6)
Unknown	19	−9.13 (−20.4, 9.6)
Living arrangement			0.37
One home	146	−8.49 (−18.0, 5.9)
Not one home	20	−10.92 (−17.6, −2.2)
Missing	4	−1.86 (−18.1, 13.3)
Snoring			0.25
Occasionally/Always	56	−9.90 (−19.9, 1.3)
Rarely/Never	86	−8.03 (−18.3, 7.2)
Missing	28	—
Lack of satiety/excessive hunger			0.44
Yes	26	−12.49 (−22.7, 8.0)
No	80	−8.13 (−17.3, 2.5)
Missing	64	—
Sneaking food at night			0.80
Yes	24	−8.93 (−17.3, 8.0)
No	77	−8.75 (−18.0, 2.3)
Missing	69	—
Adherence status
Physical activity adherence			0.15
Medium-high	36	−11.39 (−21.2, 4.2)
Low-none	93	−7.63 (−15.6, 3.4)
No recommendation	41	−9.18 (−20.4, 7.2)
Dietary Adherence			0.011^*
Medium-high	78	−11.39 (−20.4, −6.6)
Low-none	90	−7.49 (−14.3, 5.9)
Missing	2	7.89 (6.3, 9.4)
Diagnoses
ADHD			0.92
Yes	24	−9.98 (−17.9, 8.6)
No	146	−8.68 (−18.0, 5.3)
Anxiety, Depression, Bipolar, ODD			0.25
Yes	34	−13.44 (−20.0, 5.7)
No	136	−8.28 (−16.7, 5.8)
OSA			0.41
Yes	35	−11.24 (−21.3, 2.1)
No	135	−8.61 (−16.6, 5.9)
Asthma			0.06
Yes	39	−7.84 (−11.7, 9.4)
No	131	−9.63 (−19.6, 3.3)
Genetic, Developmental Delay			0.11
Yes	18	−13.78 (−19.3, −7.8)
No	152	−8.49 (−17.3, 6.1)
Medications
Metformin			0.93
Yes	29	−9.81 (−16.8, 9.6)
No	141	−8.38 (−18.0, 5.7)
Sedation^b			0.47
Yes	16	−12.57 (−21.3, 8.2)
No	154	−8.68 (−16.8, 5.7)
Atypical Antipsychotic^b			0.30
Yes	6	−17.08 (−21.2, −9.7)
No	164	−8.68 (−17.9, 5.8)
Selective serotonin reuptake inhibitors^b			0.99
Yes	22	−9.39 (−16.4, 2.3)
No	148	−8.82 (−18.1, 5.8)
Stimulant^b			0.27
Yes	23	−14.60 (−21.3, 5.7)
No	147	−8.38 (−16.8, 5.9)
Weight loss, antiepileptic^b			0.12
Yes	14	−14.89 (−19.3, −8.6)
No	156	−8.28 (−18.0, 6.1)

p-Value for univariate analysis using the ANOVA mixed effects with site as a random factor.

Sedation medications: trazodone, melatonin; atypical antipsychotic: aripiprazole, lithium, lurasidone, olanzapine, quetiapine, risperidone, valproate, ziprasidone; SSRI: citalopram, escitalopram, fluoxetine, fluvoxamine, paroxetine, sertraline, methylphenidate, dexmethylphenidate, amphetamine, dextroamphetamine, methamphetamine, atomoxetine, guanfacine, nuvigil; weight loss/antiepileptic: topiramate, phentermine, orlistat, phentermine-topiramate, naltrexone-bupropion.

p < 0.05 is significant.

%BMIp95, percent of the 95th percentile for BMI; ADHD, Attention Deficit Hyperactivity Disorder; ODD, Oppositional Defiant Disorder; OSA, obstructive sleep apnea; SSRI, selective serotonin reuptake inhibitor.

After applying a multivariable analysis, dietary adherence remained significantly associated with %BMIp95 changes. Specifically, we observed a greater reduction in %BMIp95 in those with medium-high dietary adherence compared to low-none (−10.8 vs. −4.0; p = 0.002). On the contrary, we did not demonstrate statistically significant associations with adherence to physical activity recommendations and reduction in %BMIp95. Related to baseline patient characteristics, we observed that those with SO-3 had greater reductions in %BMIp95 than those with OB-1 (−11.7 vs. −3.6; p = 0.008). The trend for greater %BMIp95 reductions in patients without an asthma diagnosis vs. those with an asthma diagnosis remained in the multivariable results (−9.8 vs. −5.0; p = 0.06) (Table 4).

Table 4.

Multivariable Analysis for Changes in Percentage of the 95th Percentile for BMI

Patient characteristics (N = 168 ^a )	n	LSMean ± SE	p ^b
Baseline weight status			p_{1 vs. 2} = 0.40p_{1 vs. 3} = 0.008p_{2 vs. 3} = 0.17
Obese—Class 1	56	−3.60 ± 2.25
Severe Obese—Class 2	56	−6.94 ± 2.23
Severe Obese—Class 3	56	−11.66 ± 2.18
Dietary compliance			0.0021
None/Low: 0%–49% of visits	90	−4.00 ± 1.90
Medium/High: 50%–100% of visits	78	−10.80 ± 2.02
Diagnosis: Asthma			0.06
Yes	37	−4.98 ± 2.48
No	131	−9.82 ± 1.58
Sex			0.14
Female	93	−5.84 ± 1.89
Male	75	−8.96 ± 2.00
Age group at baseline			0.19
5–11 Years	101	−5.98 ± 1.83
12–17 Years	67	−8.82 ± 2.07

Two patients had unknown dietary compliance for all visits and were not included in the multivariable analysis.

p-Values ANOVA mixed effects model with site as random variable. Tukey-Kramer p-value adjustment used for pairwise comparisons for baseline weight status.

LSMean, least squares mean; SE, standard error.

Given that dietary adherence was most associated with clinical success, a post hoc analysis explored other patient characteristics associated with better dietary adherence, defined by those who reported dietary adherence in at least 50% of the visits. Patients with private insurance demonstrated better dietary adherence compared to those with public insurance, 59% vs. 41%, respectively (p = 0.04). Patients with OB-1 weight status at baseline had better dietary adherence than those with SO-3, 59% vs. 30%, respectively (p = 0.007). In addition, lower dietary adherence was correlated with the presence of a prescribed selective serotonin reuptake inhibitor (SSRI), 23% vs. none 50% (p = 0.02).

Discussion

We were not able to identify an association between available patient-level characteristics and clinically meaningful %BMIp95 reductions. Furthermore, we were unable to evaluate associations between group nutrition and physical activity classes and %BMIp95 reduction because few sites reported patient attendance in these offerings. However, we did find a correlation between greater adherence to dietary recommendations and %BMIp95 reductions during treatment. Our study fills a gap in “real-world” data about PWMP care. We found only one published study from a single site reporting on increased physical activity and a lack of depression, but not dietary adherence associated with a >5%BMIp95 reduction. Thus, reported factors associated with clinical success in a “real-world” setting are not consistent.²¹

In contrast to the single-site study, we assessed the use of different dietary approaches, rather than a single dietary intervention, which not only varied across different sites but also differed among patients seen at the same site. We also observed that dietary approaches were sometimes changed during the course of treatment with the same patient or multiple strategies were used simultaneously, suggesting adaptability in provider practices such that different treatment options may have been used in response to barriers and needs of the patient-family.

These findings are consistent with other studies reporting on adherence to different dietary regimens and improvement in weight status in children and adolescents with obesity. A systematic review with meta-analysis that included 14 studies conducted over a 38-year period (1975–2013) found improved weight status can be achieved in children with overweight or obesity regardless of the macronutrient distribution of a reduced-energy diet, when intervention adherence is high.²²

Although not directly associated with %BMIp95 changes, having public insurance was linked to lower dietary adherence. Due to its ubiquitous availability in the medical record, insurance type is often utilized as a proxy for socioeconomic status (SES). While imperfect, public insurance is clearly associated with education level, household income, and neighborhood poverty.²³ Previous research suggests adolescents on public insurance were more likely to have an elevated weight status, and that youth on public insurance engaged in weight management were more likely to drop out of treatment.^24,25

Similarly, low SES has been associated with poorer nutrition habits, excess screen time, and less physical activity.²⁶ Youth in low SES settings often encounter additional barriers to healthy eating, including food insecurity, less access to healthy foods, and fewer healthy family meal habits,^27–29 and it is likely that these factors also impede adherence to nutritional treatment approaches to pediatric obesity. Thus, it is not surprising that the current results showed associations between public insurance and suboptimal nutritional adherence during treatment.

To our knowledge, the association of SSRI prescription and lower dietary adherence is a novel one. Prospective research has confirmed increased gains in adiposity among youth with depression and among those with SSRI prescriptions,³⁰ suggesting SSRIs may provide additional risk of elevated adiposity, above that of depression itself. In addition, SSRI-associated weight gain among adolescents with overweight has been shown to be stronger among those with unhealthy behaviors, such as consuming a western diet and excess time spent sedentary.³¹

Beyond dietary adherence, no other statistically significant patient-level or treatment-level characteristic was associated with significant improvement in weight status. Our previous outcomes article⁹ reported an insignificant trend of BMI reductions and increased number of visits, but given the heterogeneity of sites with wide ranges of encounter duration, example one visit at one site might be 1 hour duration and another site might be 4 hours, dose response was not included in our analysis.

Regarding the use of antiobesity medications, our sample had very few receiving this form of treatment (n = 14), which limited our statistical power to detect a potentially meaningful difference. Moreover, the dataset did not include information to indicate successful adherence to the medications or information on dose, so these null findings should be interpreted with caution.³² In addition, those with class 3 obesity had higher %BMIp95 reductions than those with OB-1 and class 2 obesity, but paradoxically seemed to have worse dietary adherence. A possible explanation for this is that an equivalent proportional loss for those with class 3 obesity was greater than those with OB-1.

Interestingly, we also failed to detect an association between lack of satiety and change in %BMIp95 or dietary adherence. However, our results were limited to the clinician's chart notes regarding patient-reported satiety and hunger, and therefore may not provide a robust metric (e.g., a validated hyperphagia questionnaire). Further research is needed to determine if tailored nutrition plans or advanced therapies can improve outcomes among patients with abnormal hunger/satiety cues.

Strengths of this study include a target population from which the sample was drawn from a large and nationally representative group of programs; however, the actual sample size was from five different PWMPs located in a large central region of the United States. Due to the relatively large POWER data registry, we were able to investigate factors of patients with clinically meaningful %BMIp95 reductions at 6 months vs. those without this level of change, and also added manual data abstraction from records with more detailed information on patient- and treatment-level characteristics.

Study limitations include reliance on data collection through a retrospective chart review meaning data accuracy depends on the quality of documentation at the time of the visit by multiple providers. In addition, although trained according to an established protocol with an agreed upon data dictionary, using different personnel at each site may have introduced variation in data extraction and interpretation of goal adherence. Furthermore, our data are based on patient self-reporting on both dietary and physical activity adherence without more objective measures, which may contribute to the null findings. Finally, the sample size from only five POWER sites may have limited our ability to discern potentially meaningful differences among less common patient characteristics or treatment practices and may not be fully representative of all POWER sites, thus limiting its generalizability.

Based on data extracted from medical record documentation, we found treatment-seeking youth with obesity who had meaningful reductions in %BMIp95 differed from their peers only in their reported improved adherence to recommended dietary interventions. Prospective research would likely benefit this area of study, which would allow investigators to target factors likely to be relevant to treatment success. Such research is warranted to further establish best practices for PWMPs to pursue, and to provide useful metrics for evaluating their progress in this pursuit.

Impact Statement

Little is known about which intensive lifestyle treatments and patient characteristics increase the likelihood of clinically meaningful outcomes within PWMPs. Adherence to dietary recommendations is associated with meaningful change, and confirms that the type of dietary intervention selected was not essential for this outcome.

Footnotes

Authors' Contributions

A.C., S.K., S.C., B.S., J.T., E.K., and M.F. conceptualized/designed the study and methodology, A.C., S.C., S.K., B.S., J.T., and M.F. conducted investigation and data curation, E.K. conducted data curation and formal analysis, A.C. and S.K. conducted supervision/oversight, A.C., S.C., S.K., J.T., and M.F. drafted the initial article. A.C., S.C., S.K., B.S., J.T., and M.F. reviewed and edited the article and gave final approval of submitted drafts.

Funding Information

No funding was received for this article.

Author Disclosure Statement

B.S. serves as a speaker for Rhythm Pharmaceuticals and member of the POWER Governance Board. S.C. serves as a speaker for Rhythm Pharmaceuticals and serves as an advisor for Novo Nordisk. E.K. has stock ownership in Procter & Gamble Company. S.K. serves as a consultant to Nestle and Eli Lilly and Company and the principal investigator (PI) and member of the POWER Governance Board. J.T. receives institutional payments for program evaluation at the Salvation Army Kroc Center. A.C. is a member of the POWER Governance Board. All other authors have no conflict to report.

Supplementary Material

References

Skinner

, Ravanbakht

, Skelton

, et al. Prevalence of obesity and severe obesity in US children, 1999–2016. Pediatrics, 2018; 141(3):e20173459; doi: 10.1542/peds.2017-3459

Smith

, Fu

, Kobayashi

. Prevention and management of childhood obesity and its psychological and health comorbidities. Annu Rev Clin Psychol, 2020; 16:351–378; doi: 10.1146/annurev-clinpsy-100219-060201

US Preventive Services Task Force; Grossman

, Bibbins-Domingo

, et al. Screening for obesity in children and adolescents: US Preventive Services Task Force recommendation statement. JAMA, 2017; 317(23):2417–2426; doi: 10.1001/jama.2017.6803

Skinner

, Perrin

, Moss

, et al. Cardiometabolic risks and severity of obesity in children and young adults. N Engl J Med, 2015; 373(14):1307–1317; doi: 10.1056/NEJMoa1502821

Kelly

, Barlow

, Rao

, et al. Severe obesity in children and adolescents: Identification, associated health risks, and treatment approaches: A scientific statement from the American Heart Association. Circulation, 2013; 128(15):1689–1712; doi: 10.1161/CIR.0b013e3182a5cfb3

Fox

, Gross

, Bomberg

, et al. Severe obesity in the pediatric population: Current concepts in clinical care. Curr Obes Rep, 2019; 8(3):201–209; doi: 10.1007/s13679-019-00347-z

Styne

, Arslanian

, Connor

, et al. Pediatric obesity-assessment, treatment, and prevention: An Endocrine Society clinical practice guideline. J Clin Endocrinol Metab, 2017; 102(3):709–757; doi: 10.1210/jc.2016-2573

Barlow SE; Expert Committee. Expert Committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: Summary report. Pediatrics, 2007; 120(Suppl 4):164; doi: 10.1542/peds.2007-2329C

Kumar

, King

, Christison

, et al. Health outcomes of youth in clinical pediatric weight management programs in POWER. J Pediatr, 2019; 208:57.e4–65.e4; doi: 10.1016/j.jpeds.2018.12.049

10.

Tucker

, Stratbucker

, King

, et al. Characteristics of paediatric weight management in the United States: Associations with program retention and BMI outcomes in the paediatric obesity weight evaluation registry (POWER). Pediatr Obes, 2022; 17(2):e12848; doi: 10.1111/ijpo.12848

11.

Kirk

, Armstrong

, King

, et al. Establishment of the Pediatric Obesity Weight Evaluation Registry: A national research collaborative for identifying the optimal assessment and treatment of pediatric obesity. Child Obes, 2017; 13(1):9–17; doi: 10.1089/chi.2016.0060

12.

Kirk

, Ogata

, Wichert

, et al. Treatment of pediatric overweight and obesity: Position of the Academy of Nutrition and Dietetics based on an umbrella review of systematic reviews. J Acad Nutr Diet, 2022; 122(4):848–861; doi: 10.1016/j.jand.2022.01.008

13.

Freedman

, Butte

, Taveras

, et al. BMI z-scores are a poor indicator of adiposity among 2- to 19-year-olds with very high BMIs, NHANES 1999–2000 to 2013–2014. Obesity (Silver Spring), 2017; 25(4):739–746; doi: 10.1002/oby.21782

14.

Reinehr

, Lass

, Toschke

, et al. Which amount of BMI-SDS reduction is necessary to improve cardiovascular risk factors in overweight children?. J Clin Endocrinol Metab, 2016; 101(8):3171–3179; doi: 10.1210/jc.2016-1885

15.

Woo

. Using body mass index Z-score among severely obese adolescents: A cautionary note. Int J Pediatr Obes, 2009; 4(4):405–410; doi: 10.3109/17477160902957133

16.

Kelly

, Daniels

. Rethinking the use of body mass index z-score in children and adolescents with severe obesity: Time to kick it to the curb?. J Pediatr, 2017; 188:7–8; doi: 10.1016/j.jpeds.2017.05.003

17.

Freedman

, Butte

, Taveras

, et al. The limitations of transforming very high body mass indexes into z-scores among 8.7 million 2- to 4-year-old children. J Pediatr, 2017; 188:50.e1–56.e1; doi: 10.1016/j.jpeds.2017.03.039

18.

Flegal

, Wei

, Ogden

, et al. Characterizing extreme values of body mass index-for-age by using the 2000 Centers for Disease Control and Prevention growth charts. Am J Clin Nutr, 2009; 90(5):1314–1320; doi: 10.3945/ajcn.2009.28335

19.

Harris

, Taylor

, Thielke

, et al. Research electronic data capture (REDCap)—A metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform, 2009; 42(2):377–381; doi: 10.1016/j.jbi.2008.08.010

20.

Harris

, Taylor

, Minor

, et al. The REDCap consortium: Building an international community of software platform partners. J Biomed Inform, 2019; 95:103208; doi: 10.1016/j.jbi.2019.103208

21.

Gorecki

, Feinglass

, Binns

. Characteristics associated with successful weight management in youth with obesity. J Pediatr, 2019; 212:35–43; doi: 10.1016/j.jpeds.2019.05.039

22.

Gow

, Ho

, Burrows

, et al. Impact of dietary macronutrient distribution on BMI and cardiometabolic outcomes in overweight and obese children and adolescents: A systematic review. Nutr Rev, 2014; 72(7):453–470; doi: 10.1111/nure.12111

23.

Casey

, Pollak

, Glymour

, et al. Measures of SES for electronic health record-based research. Am J Prev Med, 2018; 54(3):430–439; doi: 10.1016/j.amepre.2017.10.004

24.

Haas

, Lee

, Kaplan

, et al. The association of race, socioeconomic status, and health insurance status with the prevalence of overweight among children and adolescents. Am J Public Health, 2003; 93(12):2105–2110; doi: 10.2105/ajph.93.12.2105

25.

Walker

, Smolkin

, O'Leary

MLL

, et al. Predictors of retention and BMI loss or stabilization in obese youth enrolled in a weight loss intervention. Obes Res Clin Pract, 2012; 6(4):e330–e339; doi: 10.1016/j.orcp.2011.08.157

26.

Poulain

, Vogel

, Sobek

, et al. Associations between socio-economic status and child health: Findings of a large German cohort study. Int J Environ Res Public Health, 2019; 16(5):677; doi: 10.3390/ijerph16050677

27.

Tarasuk

, Fafard St-Germain

, Mitchell

. Geographic and socio-demographic predictors of household food insecurity in Canada, 2011–2012. BMC Public Health, 2019; 19(1):12; doi: 10.1186/s12889-018-6344-2

28.

Cushon

, Creighton

, Kershaw

, et al. Deprivation and food access and balance in Saskatoon, Saskatchewan. Chronic Dis Inj Can, 2013; 33(3):146–159.

29.

Attorp

, Scott

, Yew

, et al. Associations between socioeconomic, parental and home environment factors and fruit and vegetable consumption of children in grades five and six in British Columbia, Canada. BMC Public Health, 2014; 14:150; doi: 10.1186/1471-2458-14-150

30.

Schwartz

, Glass

, Pollak

, et al. Depression, its comorbidities and treatment, and childhood body mass index trajectories. Obesity (Silver Spring), 2016; 24(12):2585–2592; doi: 10.1002/oby.21627

31.

Shi

, Atlantis

, Taylor

, et al. SSRI antidepressant use potentiates weight gain in the context of unhealthy lifestyles: Results from a 4-year Australian follow-up study. BMJ Open, 2017; 7(8):e016224; doi: 10.1136/bmjopen-2017-016224

32.

Borzutzky

, King

, Fox

, et al. Trends in prescribing anti-obesity pharmacotherapy for paediatric weight management: Data from the POWER Work Group. Pediatr Obes, 2021; 16(1):e12701; doi: 10.1111/ijpo.12701

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.38 MB