Efficacy of a Primary Care eHealth Obesity Treatment Pilot Intervention Developed for Vulnerable Pediatric Patients

Abstract

Background:

Challenges to treat excess weight in primary care settings include time constraints during encounters and barriers to multiple visits for patient families, especially those from vulnerable backgrounds. Dynamo Kids! (DK), a bilingual (English/Spanish) e-health intervention, was created to address these system-level challenges. This pilot study assessed the effect of DK use on parent-reported healthy habits and child BMI.

Methods:

In this 3-month, quasi-experimental cohort design, DK was offered to parents with children aged 6–12 years with BMI ≥85th percentile in three public primary care sites in Dallas, Texas. DK included three educational modules, one tracking tool, recipes, and links to internet resources. Parents completed an online survey before and after 3 months. Pre-post changes in family nutrition and physical activity (FNPA) scores, clinic-measured child %BMI_p95, and self-reported parent BMI were assessed using mixed-effects linear regression modeling.

Results:

A total of 73 families (mean child age = 9.3 years; 87% Hispanic, 12% non-Hispanic Black, and 77% Spanish-speaking families) completed the baseline survey (participants) and 46 (63%) used the DK site (users). Among users, pre-post changes (mean [standard deviation]) showed an increase in FNPA scores (3.0 [6.3], p = 0.01); decrease in child %BMI_p95 (−1.03% [5.79], p = 0.22); and decrease in parent BMI (−0.69 [1.76], p = 0.04). Adjusted models showed −0.02% [95% confidence interval: −0.03 to −0.01] change in child %BMI_p95 for each minute spent on the DK website.

Conclusions:

DK demonstrated a significant increase in parent FNPA scores and decrease in self-reported parent BMI. e-Health interventions may overcome barriers and require a lower dosage than in-person interventions.

Introduction

Childhood obesity remains a persistent epidemic in the United States (US) with immediate and future concurrent health risks that include cardiovascular disease; diabetes; cancers such as gastric and colorectal cancers; and death.^1,2 Important aspects of the childhood obesity epidemic in the US include its deepening and widening disparities and high economic costs.^3–9 Specifically, 19.3% of US children aged 2–19 years have a BMI ≥95th percentile adjusted for age and sex (%BMI_P95),¹⁰ defined as having obesity, but the prevalence is even higher among non-Hispanic Black (20.4%) and Hispanic youth (23.6%), compared with non-Hispanic White youth (14.7%).^9,11–13

Even more troubling, the rate of change in BMI doubled from 0.052 to 0.100 kg/m²/month in US children aged 2–19 years during the COVID-19 pandemic, compared with the prepandemic period,¹⁴ heightening the need to address obesity⁷ in postpandemic recovery efforts.

To address childhood obesity, the Obesity Chronic Care Model¹⁵ recommends that (1) children's weight statuses be evaluated at health care practices; (2) providers establish a relationship with the family that focuses on weight in a health context (vs. appearance); and (3) parents or guardians (hereafter “parents”) understand that OW/OB are medical diagnoses. Indeed, studies have shown that children with excess weight benefit from comprehensive, family-based, moderate- to high-intensity weight control interventions.^16–18

Yet, the US health care system is not currently able to deliver intensive interventions that adhere to US Preventive Services Task Force (USPTF) best practices, especially for the most at-risk populations due to limited time in clinical visits, insufficient financial resources, and lifestyle barriers such as lack of time, transportation, or child care.^16,19 Moreover, addressing pediatric obesity through primary care is difficult for providers who often lack the time, knowledge, and skills to be effective.¹⁵

Similarly, families are challenged by disrupted schedules and medical information about weight management that may be delivered in an inconvenient and/or not culturally congruent manner.^20,21 Moreover, vulnerable populations may not have a trusted provider or health insurance coverage to afford the necessary care.¹⁶

While e-health interventions may be a strategy to address the provider and patient barriers to obesity care mentioned above, research on their efficacy is limited. Existing studies focus on mobile alerts or telehealth engagement instead of personalized/independent online programs.²² Yet, according to the USPTF, effective obesity treatment programs should be multicomponent, address behavior change, include at least 26 hours of exposure, and last for 2–12 months.¹⁶ Moreover, parents play a critical role as the family environment is predominant in development of children's eating and physical activity behaviors.^23–25 Yet, it is unknown if these recommendations apply to a self-paced, tailored, e-health delivery model.

This study aimed to assess the efficacy of Dynamo Kids! (DK), an e-health, family-based, self-paced, and personalized pilot intervention, on improvement in child and family obesity-mitigating behaviors, as well as child and parent BMI. It was hypothesized that participation in the DK intervention would (1) improve child and family obesity-mitigating behaviors, as measured by family nutrition and physical activity (FNPA) scores; (2) decrease %BMI_P95 in child participants; and (3) decrease self-reported BMI in parent participants.

Methods

Study Design

This quasi-experimental (uncontrolled) cohort study enrolled 100 families from 3 public primary care centers in Dallas, TX. Participant families were asked to utilize DK for 3 months and were evaluated before they were exposed to any content and after they completed the intervention. This study was reviewed and approved by both UTHealth and UT Southwestern institutional review boards (STU 2019–0876) and developed in accordance with the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines (Supplementary Data).²⁶

Study Population

The study population comprised vulnerable racial and ethnic minority families with children aged 6–12 years who attended a primary care center that was part of the public hospital system. The pilot study was intentionally launched in three neighborhoods that are predominantly inhabited by minority populations, which are generally hard to reach; however, the clinics see patients from all racial and ethnic backgrounds.

Eligibility and Recruitment

Parents who met the following eligibility criteria were included in the study: (1) had a child 6–12 years of age; (2) the child was a patient of the participating providers at three outpatient practices; (3) the child had a BMI ≥85th percentile at that appointment; and (4) the parent spoke either English or Spanish. Children with a chronic condition that prevents typical physical activity (e.g., used a wheelchair) or requires an atypical diet (e.g., gastrostomy feeds or type 1 diabetes) were excluded. Recruitment occurred from August 2020 to April 2021.

Providers received an electronic health record (EHR) alert and referral link when family met inclusion criteria at a well-child visit. After receiving the referral, a study coordinator called parents to establish eligibility, obtain verbal consent, and enroll the family in the DK intervention. Consent was obtained only from the parent because the intervention was directed toward the parents of the child with OW/OB rather than to the child.

Before beginning the intervention, participants were required to complete a baseline research survey (Supplementary Appendix SA1).

Study Procedures

Based on child inclusion criteria, an EHR alert identified 575 eligible patients; and providers identified 6 patients outside the EHR alert. Providers referred 223 patients; 100 consented to participate. Of the 100 consented parents, 73 completed the baseline survey and 46 used the DK website (Fig. 1). Baseline child weight and height were extracted from the EHR at the well-child visit when the EHR alert fired.

Figure 1.

Dynamo Kids! CONSORT Flow Diagram. %BMI_P95, BMI as a percent of the 95th percentile; EHR, Electronic Health Record; FNPA, family nutrition and physical activity; mBMI_p95, mean child %BMI_p95.

Intervention Description

DK is a family-based, e-health pilot intervention that is guided by the Obesity Chronic Care Model.^27,28 This technology-based, customized, self-paced, and self-directed intervention includes an online portal in both English and Spanish, which parents access to learn strategies to improve specific health behaviors for their children. Additionally, the website intervention includes two practice-based features: (1) an EHR-embedded alert that appears when DK inclusion criteria are met and (2) a customized report based on patient website use, with talking points for providers at a follow-up clinical visit.²⁹

DK is intended to extend brief primary care counseling using an e-health format, with the capacity for exposing families to hours rather than minutes of weight management support. Specifically, the DK foci were to reduce consumption of sugar-sweetened beverages, increase physical activity levels, and adhere to MyPlate recommendations. Tables 1 and 2 provide details of DK; a description of development and protocol is published elsewhere.^27,29

Table 1.

Dynamo Kids! Intervention Characteristics and Content

DK components	Description
Module topics 1. SSB 2. PA 3. MyPlate	SSB: reduce consumption of SSB, ideally to none. PA: increase PA to the CDC-recommended duration and intensity and reduce sedentary behavior MyPlate: adhering to MyPlate recommendations of preparing, serving, and consuming food All modules were written at a sixth-grade level, used evidenced-based information primarily from US government websites (i.e., the US Department of Agriculture's MyPlate website and the CDC), and were presented in a user-friendly manner guided by a needs assessment.²⁵ Aligned with parent feedback during the needs assessment,²⁴ the purpose of the intervention and repetitive structure was to improve the child's general health, nutritional health, and weight, in addition to motivating the entire family to have a healthier lifestyle.
Intervention structure: Intervention duration: 3 months Monthly organization: 1. Week 1: knowledge 2. Week 2: strategies 3. Week 3: overcoming barriers 4. Week 4: reflection	The website was designed to be used for 3 months, ∼1 month per module, with each module organized around four general topics. Based on the feedback from the needs assessment,²⁴ it was expected that parents would engage about three times weekly for 20 minutes for 3 months. Participants could select the order of modules and could determine when and how long to use. To encourage full exploration of each module, participants could move to the next module only after 2 weeks. Week 1 introduced information on the topic to educate parents on the topic. Week 2 presented strategies to establish healthy habits for their child. Week 3 presented strategies to overcome common barriers families face when enacting such changes. Week 4 offered a process and space for parents to reflect on what they learned, what they did, and how they can sustain such changes.
Customized programmatic features: 1. Personalized module order 2. Goal setting 3. Tracking regular personalized text messages and e-mails	1. Users chose the order in which they engaged with the modules. 2. In each module, parents learned and practiced setting goals for their families. 3. Parents tracked their progress in achieving their goals. 4. Participants received personalized push messages weekly from the research team that encouraged them to continue on their DK journey based on their current completion level, as tracked and measured by the DK website.

DK, Dynamo Kids!; PA, physical activity; SSB, sugar-sweetened beverages.

Table 2.

Sample Dynamo Kids! Education and Engagement Plan, Stratified by Module

Week	Week's title	SSB example	PA example	MyPlate example
1	Learning more	Learning about the quantity of sugar in SSB	Learning the benefits of PA for physical and mental health and they 60-minute time recommendation	Learning what a healthy MyPlate meal includes and how to identify the different types of food (e.g., carbohydrates, protein, fruits, and vegetables)
2	Building skills	Using flavored water as an alternative	Using the resources section to find a nearby park and go for a family walk for 60 minutes	Using low-cost recipes from the USDA to prepare meals for your family
3	Overcoming barriers	Making it a competition: finding the drink with the least sugar in the grocery store with your child	Thinking through how older siblings may encourage your child with OW/OB to regularly engage in PA	Reading stories about similar families with your child with OW/OB so that he does not feel alone.
4	Moving forward	Reflecting on what drinks you no longer consume and what challenges you faced from your family and exploring additional resources on the topic	Developing a plan that allows you, your family, and child with OW/OB to regularly engage in PA	Reflecting with your partner about what foods are financially feasible and preparing a menu for the upcoming weeks.

Each module (SSB, PA, and MyPlate) has the same 4-week structure. Participants can choose the order in which they engage with each topic, and upon completing one module (e.g., PA), they can select the module they want to continue with in week 5 (e.g., MyPlate) and then begin their final module in week 9.

OW/OB, overweight and obesity.

Outcome Measures

Primary outcome

The pre-post change in FNPA screening composite scores^30,31 used to measure parent-reported healthy home environment and parenting practices was the primary outcome. The FNPA survey³² is a validated 20-question survey, available in both English and Spanish; it has been shown to be effective at predicting children's risk for having OW/OB.²⁵

Secondary outcomes

Secondary outcomes included (1) age- and sex-specific %BMI_p95, calculated based on the 2000 CDC growth charts³³ using child height and weight collected using the Health-o-Meter professional tool during clinical visits, which were extracted from the EHR; and (2) parent self-reported height and weight from baseline and follow-up surveys, which were used to calculate BMI.

Follow-up child measurements were extracted from the EHR at the planned weight-focused visit following the intervention or (if the visit did not occur) at the next visit, up to 14 months after the referral visit, allowing time for an annual well-child visit to occur.

Primary exposure

The primary exposure variable was the total number of minutes that families spent and stored on the DK website.

Covariates

Child sex, time between the well-child visit and follow-up visit, and age were primary covariates collected through the EHR. Child race/ethnicity, parent's educational attainment, parent's age, food insecurity status, parent perception of their child's weight, and the family's language preference (English or Spanish) collected using the baseline participant survey were additional covariates.

Statistical Analyses

Descriptive statistics were performed on all variables, both aggregate and stratified by DK user status (i.e., participants who used the DK website [users] and participants who did not use the DK website [nonusers]). Bivariate analyses included correlation coefficients, t-tests, χ², or Fisher exact tests, as appropriate, to assess the relationship between the three dependent outcomes and all independent variables, including user status.

Bivariate analyses were conducted to determine if missing data were missing completely at random. Given that the multiple imputation estimates and standard errors were very different from those in the complete case analysis, the missing data mechanism was assumed to be informative.³⁴

Hence, we limited our interpretation to multiple imputation to estimate the missing data: we produced 11 values for parent baseline BMI, 6 values for food insecurity, 14 values for time between baseline and follow-up measurements, 31 values for the change in FNPA survey, 14 values for the change in child %BMI_p95, and 37 values for the change in parent BMI. Then, we used mixed-effects modeling to identify significant predictors of the change in the three outcomes of interest.

Linear mixed-effects repeated-measures models assessed the effect of DK on the three outcomes of interest and accounted for clustering within participants over time. Multiple imputation using chained equations imputed missing outcomes for the six variables with missing data.²⁵ Covariates were selected for inclusion based on literature^35,36 reporting an association with weight status or when significantly correlated (at a p ≤ 0.05 level).

All data analyses were conducted using Stata 16 (StataCorp 2019, College Station, TX).

Results

For the children of the 73 participants (46 [63%] users and 27 [37%] nonusers), the baseline mean (standard deviation [SD]) age was 9.3 (1.79) years; 60% were male; 88% were Hispanic and 12% were non-Hispanic Black; 78% preferred Spanish content; 88% had annual household incomes <$40,000; and 56% reported food insecurity in the previous year (Tables 3 and 4). Slightly more than half (55%) of the participants completed the follow-up survey (33 [72%] users and 7 [26%] nonusers).

Table 3.

Descriptive Characteristics of the Children and Family Household Resources in the Study Sample

	All survey completers, n = 73	Nonuser, survey completer, n = 27	User, survey completer, n = 46	p ^*
Child age,
years, mean (SD)	9.31 (1.80)	9.63 (1.88)	9.13 (1.74)	0.18^a
Child sex, n (%)
Male	44 (60.3)	22 (81.5)	22 (47.8)	0.005 ^b
Female	29 (39.7)	5 (18.5)	24 (52.2)
Race/ethnicity, n (%)
Hispanic	64 (87.7)	24 (88.9)	40 (87)	0.56^c
Non-Hispanic Black	9 (12.3)	3 (11.1)	6 (13.0)
Non-Hispanic White	0 (0)	0 (0)	0 (0)
Household income, n (%)^d
Less than $20,000	28 (38.4)	14 (51.9)	14 (30.4)	0.38^b
$20,000–$29,999	20 (27.4)	7 (25.9)	13 (28.3)
$30,000 or more	24 (32.9)	6 (22.2)	18 (39.1)
Parent access to a car, n (%)^e
No car available	5 (6.8)	1 (3.7)	4 (8.7)	0.73^c
Car available sometimes	19 (26.0)	8 (29.6)	11 (23.9)
Car available most of the time	48 (65.8)	18 (66.7)	30 (65.2)
Home access to the internet, n (%)
No access to the internet	1 (1.4)	1 (3.7)	0 (0)	0.35^c
Cell phone access only	14 (19.2)	6 (22.2)	8 (17.4)
Cell and home internet (Wi-Fi) access	58 (79.5)	20 (74.1)	38 (82.6)
Food insecurity, positive screen, n (%)^f
Yes	41 (61.2)	16 (59.3)	25 (54.3)	0.75^b
No	26 (38.8)	8 (29.6)	18 (39.1)

Bold text denotes statistically significant value.

Difference between users and nonusers.

t-Test; ^bχ² test; ^cFisher's exact test; ^dmissing data for one user (2.2%); ^emissing data for one user (2.2%); and ^fmissing data for three nonusers (11.1%) and three users (6.5%).

Table 4.

Descriptive Characteristics of Parents in the Study Sample

	All survey completers, n = 73	Nonuser, survey completer, n = 27	User, survey completer, n = 46	p ^*
Age, n (%)
<25 years	1 (1.4)	0 (0)	1 (2.2)	0.34^b
25–34 years	26 (35.6)	8 (29.6)	18 (39.1)
35–44 years	40 (54.8)	15 (55.6)	25 (54.3)
45–54 years	6 (8.2)	4 (14.8)	2 (4.3)
Sex, n (%)
Male	1 (1.4)	0 (0)	1 (2.2)	0.63^b
Female	72 (98.6)	27 (100)	45 (97.8)
Self-reported existing health condition, n (%)
Self-reported, having at least one health condition	18 (24.7)	7 (25.9)	11 (23.9)	0.41^a
Marital status, n (%)
Married/living with a partner	56 (76.7)	20 (74.1)	36 (78.3)	0.89^b
Divorced/separated	10 (13.7)	4 (14.8)	6 (13.0)
Single, never married	7 (9.6)	3 (11.1)	4 (8.7)
Language preference, n (%)
English	17 (23.3)	5 (18.5)	12 (26.1)	0.33^b
Spanish	56 (76.7)	22 (81.5)	34 (73.9)
Educational level, n (%)
Some high school or less	28 (38.4)	13 (48.1)	15 (32.6)	0.40^a
High school degree or GED	24 (32.9)	8 (29.6)	16 (34.8)
More than high school (some college or more)	21 (28.8)	6 (22.2)	15 (32.6)
Work for pay, n (%)
Yes, full-time	17 (23.3)	8 (29.6)	9 (19.6)	0.28^b
Yes, part-time	5 (6.8)	3 (11.1)	2 (4.3)
No	51 (69.9)	16 (59.3)	35 (76.1)
Perception of child's weight, n (%)
Approximately the correct weight	5 (6.8)	3 (11.1)	2 (4.3)	0.53^b
Slightly overweight	55 (75.3)	19 (70.4)	36 (78.3)
Very overweight	13 (17.8)	5 (18.5)	8 (17.4)

Difference between users and nonusers.

χ² test.

Fisher's exact test.

A total of 28 (38%) participants attended a planned follow-up visit (21 [46%] users and 7 [26%] nonusers). Among users, 21 (46%) had measures from weight-focused follow-up visits vs. 16 (35%) from other visit types. Among nonusers, 7 (26%) had measures from weight-focused follow-up visits vs. 15 (56%) from other visit types. The mean (SD) days between baseline and follow-up anthropometric assessments was 269 (107.84) days. The mean (SD) number of minutes spent on the DK website was 144.39 (132.08) minutes, with a range from 4 to 476 minutes.

Families with girls were more likely to use the DK website (χ² = 8.05, p = 0.005). Characteristics associated with increase in FNPA scores included younger child age (ρ = −0.39, p = 0.03) and parents older than 35 years (t = 2.48, p = 0.02). Girls (t = 2.01, p = 0.05) were associated with a higher baseline child %BMI_p95, and older parents (ρ = 0.27, p = 0.02) were associated with a decrease in child %BMI_p95. A higher baseline child %BMI_p95 (ρ = 0.35, p = 0.01) and lower educational level (ρ = 0.34, p = 0.01) were associated with a higher baseline parent BMI (Table 5).

Table 5.

Baseline Family Nutrition and Physical Activity Scores, Child BMI_p95 Measures, and Parent BMI

	All survey completers, n = 73	Nonuser, survey completer, n = 27	User, survey completer, n = 46	p
FNPA score
Mean (SD) [range: 31–70]	53 (8)	51 (9)	54 (7)	0.16^a
Child BMI categories, n (%)
Overweight (85th–95th percentile)	15 (20.5)	3 (11.1)	12 (26.1)	0.29^b
Obesity Class 1 (100% BMI_p95–119.9% BMI_p95)	31 (42.5)	11 (40.7)	20 (43.5)
Obesity Class 2 (120% BMI_p95–139.9% BMI_p95)	18 (24.7)	10 (37)	8 (17.4)
Obesity Class 3 (≥140% BMI_p95)	9 (12.3)	3 (11.1)	6 (13.0)
Child BMI_p95
Mean (SD)	116.8% (18.5)	119.9% (17.5)	115.1% (18.9)	0.29^a
Parent BMI (kg/m²) categories, n (%)
Healthy weight (18.5–24.9)	9 (12.3)	3 (11.1)	6 (13.0)	0.80^b
Overweight (25.0–29.9)	14 (19.2)	4 (14.8)	10 (21.7)
Obesity Class 1 (BMI 30.0–34.9)	19 (26.0)	6 (22.2)	13 (28.3)
Obesity Class 2 (BMI 35.0–39.9)	9 (12.3)	4 (14.8)	5 (10.9)
Obesity Class 3 (BMI ≥40.0)	11 (15.1)	3 (11.1)	8 (17.4)
Missing	11 (15.1)	7 (25.9)	4 (8.7)
Parent BMI (kg/m²)
Mean (SD)	33.0 (8.1)	32.6 (6.7)	33.2 (8.7)	0.43^a

Difference between users and nonusers.

t-Test.

Fisher's exact test.

FNPA, family nutrition and physical activity; SD, standard deviation.

Complete case analysis of the outcomes was performed for all enrolled participants and for the subset of DK users (Table 6). FNPA scores (primary outcome) improved among all enrolled participants (3.5 [SD = 6.95], p = 0.002) and among DK users (3.0 [SD = 6.25], p = 0.13).

Table 6.

Change in Three Dependent Outcomes (Family Nutrition and Physical Activity Scores, Child %BMI_p95, and Parent BMI) Among Participants: Complete Case Analysis

	Baseline survey completers (n = 73)	p ^a	Baseline survey completers who used the site^b (n = 46)	p ^a
FNPA score
Complete (pre and post) sample, n	n = 42 (57.5%)		n = 33 (71.7%)
Change mean (SD)	3.5 (7.0)	0.002	3.00 (6.2)	0.01
Child %BMI_p95
Complete (pre and post) sample, n	n = 59 (80.8%)		n = 37 (80.4%)
Change mean (SD)	–0.63% (5.5)	0.39	–1.16% (5.7)	0.22
Parent BMI
Complete (pre and post) sample, n	n = 36 (49.3%)		n = 30 (65.2%)
Change, kg/m², mean (SD)	–0.51 (1.7)	0.08	–0.69 (1.8)	0.04

Bold text indicates statistical significance.

Paired t-test analysis.

Users were defined as participants who logged into the DK website and opened one of the modules at least twice.

Decrease in mean child %BMI_p95 was seen in all enrolled participants and DK users, but was not significant: among all enrolled participants, Δchild %BMI_p95 = −0.53 (SD = 5.52), p = 0.39, and among DK users, Δchild %BMI_p95 = −1.03 (SD = 5.79), p = 0.22. Self-reported parent BMI change was not significant among all enrolled participants (−0.51 [SD = 1.69], p = 0.08), but reached significance among DK users (−0.69 [SD = 1.76], p = 0.04).

In the multiple imputation models for FNPA and self-reported parent BMI, no covariates were significantly associated with change in score. However, in the Δchild %BMI_p95 model, the time between the initial and follow-up visits that were the source of the clinical measurements was significantly associated with an increase in child %BMI_p95, and the amount of time spent on the DK website was significantly associated with a decrease in child %BMI_p95.

Specifically, while adjusting for all other covariates, a 1-day increase in time was associated with an increase in child %BMI_p95 (0.015% confidence interval [95% CI: 0.002–0.03]) and a 1-minute increase in total minutes spent on the DK website was associated with a decrease in child %BMI_p95 (−0.02% [CI 95%: −0.03 to −0.01]) (Table 7).

Table 7.

Mixed-Effects Repeated-Measures Linear Regression Modeling of Demographic and Dynamo Kids! Usage Predictors of Primary Outcome [Family Nutrition and Physical Activity Score Change] and Secondary Outcomes [Child %BMI_p95 and Parent BMI Changes]

	Model 1 outcome: FNPA score		Model 2 outcome: child %BMI_p95		Model 3 outcome: parent BMI
Predictors	β Coefficient (95% CI)	SE	β Coefficient (95% CI)	SE	β Coefficient (95% CI)	SE
Child sex (female)	−3.31 (−7.75 to 1.1)	2.23	2.56 (−0.25 to 5.36)	1.43	1.9 (−0.028 to 3.82)	0.95
Child age, months	0.01 (−0.11 to 0.13)	0.06	−0.07 (−0.14 to 0.005)	0.04	0 (−0.04 to 0.03)	0.02
Baseline child %BMI_p95	8.11 (−3.64 to 19.87)	5.95	−1.48 (−11.20 to 8.24)	4.93	−0.34(−4.27 to 3.59)	1.98
Food insecurity (positive screen)	−3.82 (−8.46 to 0.82)	2.32	−0.26 (−3.27 to 2.74)	1.53	−0.37 (−1.94 to 1.20)	0.78
Baseline parent BMI, kg/m²	−0.01 (−0.41 to 0.39)	0.20	0.07 (−0.22 to 0.36)	0.14	0.09 (−0.05 to 0.22)	0.07
Parent age (10-year increments)	0.21 (−2.76 to 3.17)	1.49	−0.07 (−2.51 to 2.37)	1.23	−0.15 (−1.08 to 0.79)	0.47
Total minutes spent on the DK website	−0.01 (−0.02 to 0.01)	0.01	−0.022 (−0.03 to −0.006)^a	0.01	0 (−0.00 to 0.01)	0.00
Days between clinical visit 1 and 2	−0.01 (−0.03 to 0.01)	0.01	0.011 (0.002 to 0.03)^b	0.01	0 (−0.005 to 0.009)	0.00
Parent educational level (some high school or less)	1.0 (ref)		1.0 (ref)		1.0 (ref)
Parent educational level (high school degree or GED certificate)	−2.14 (−6.41 to 2.14)	2.16	0.04 (−3.55 to 3.63)	1.82	0.29 (−0.92 to 1.50)	0.62
Parent educational level >high school (some college or more)	−1.95 (−7.92 to 4.02)	3.00	−1.28 (−4.98 to 2.43)	1.89	1.08 (−0.65 to 2.80)	0.87
Parent perception of weight (correct weight)	1.0 (ref)		1.0 (ref)		1.0 (ref)
Parent perception of weight (slightly overweight)	4.94 (−2.56 to 12.44)	4.79	−1.24 (−6.35 to 3.86)	2.60	−1.08 (−3.57 to 1.41)	1.25
Parent perception of weight (very overweight)	1.19 (−6.91 to 9.29)	4.12	−1.11 (−7.37 to 5.16)	3.19	−0.5 (−3.40 to 2.40)	1.47
User language (English)	3.78 (−0.41 to 7.97)	2.13	1.67 (−2.01 to 5.36)	1.88	−1.44 (−2.94 to 0.05)	0.76
Constant	−3.37 (−21.56 to 14.82)	9.24	5 (−10.94 to 20.94)	8.09	−1.91 (−8.87 to 5.06)	3.50

Missing data are shown to be missing completely at random for the following covariates: parent baseline (15%) and follow-up BMI measures (51%), food insecurity status (8%), follow-up FNPA score (42%), follow-up child BMI_p95 (19%), and time between baseline measures and follow-up visit (19%). All models were performed using multiple imputation mixed-effects multivariate linear regression.

p < 0.003.

p < 0.02.

%BMI_P95, BMI as a percent of the 95th percentile; CI, confidence interval; GED, General Education Development; Ref, reference category; SE, standard error.

Discussion

DK, a pilot, e-health family-based program for addressing pediatric excess weight in primary care, was successfully implemented in multiple primary care centers, with improvement in healthy family routines among hard-to-reach, high-risk vulnerable populations for childhood excess weight (Table 6). Reduction in parent BMI was significant only for parent BMI among DK users, but time on the website predicted improvement in child %BMI_p95 in the regression model, suggesting intervention benefit.

Given the increasing prevalence of and disparities within pediatric OW/OB,⁹ DK shows potential for addressing a vexing public health and clinical challenge that has both proximal and distal health consequences for patients and health care systems.

Compared with other studies using the FNPA tool as a primary outcome, this study showed a lower score at baseline and a smaller change in score over time.^37–39 Whereas previous research focused predominantly on non-Hispanic White populations with healthy weight, this study population included only racial and ethnic minority participants with excess weight.^38,40 Importantly, the FNPA tool has not been validated in exclusively racial and ethnic minorities.

Moreover, while research has shown that an increase of 10 points in the FNPA score corresponds to a decrease of 0.17 BMI units, research has not demonstrated what change in points is required for an observable behavioral change.^31,41 Additionally, the lower baseline scores may reflect less healthy behaviors, and the smaller change may reflect the intensity of program use.³⁹

Among DK users, the improvements in FNPA scores and parent BMI are notable because DK use was objectively tracked. The association between child %BMI_p95 and time on the website using multivariable regression suggests that parents who used the website implemented healthy behaviors for their children. Importantly, OW/OB BMI trajectories for this age group were magnified during the COVID-19 pandemic when this study was conducted, so any reduction may indicate a benefit, especially compared with the increase in child %BMI_p95 for this population during this time period.^37,42

Provider adherence to expert guidelines for weight management within primary care encounters may be poor.¹⁵ One study showed that primary care providers can motivate families to use self-guided tools to achieve short-term weight improvement.⁴³ Technology-assisted delivery of weight management within primary care has been limited to hybrid group programming⁴⁴ or low-quality telehealth engagement or mobile alerts that found minimal or no effect.²² DK is innovative as all components are digital and the educational component is self-paced and self-directed. However, given the intervention's novelty, the expected dosage, usage, and acceptance could only be surmised based on the intervention's needs assessment.²⁷

The relationship between exposure time and child %BMI_p95 demonstrates the importance of intervention dosage, as recommended by USPTF guidelines,¹⁶ and raises the question of whether e-health interventions and in-person interventions require the same intervention dosage. While designed to be a 12-hour intervention based on the information ascertained in the needs assessment,²⁷ this intervention in this population during this time period was a lower-intensity intervention as participants opted to use the DK website, on average, for ∼2.5 hours compared with the recommended 26 hours.

The significant findings are noteworthy at a low-intensity intervention because new research demonstrates how high-intensity interventions create barriers for patient engagement.⁴⁵ e-Health programs such as DK may improve engagement by allowing patients to engage on their own time and device in their own space and personalized manner. Importantly, a follow-up implementation science evaluation, perhaps using a standardized framework such as RE-AIM (reach, effectiveness, adoption, implementation, and maintenance),^46,47 would help identify the impact of and need for tailoring DK for diverse populations and settings.

Study Limitations and Strengths

There are several study limitations. A lower than expected proportion of participants used the website and/or did not return for a follow-up visit. Additionally, the follow-up clinical measures were collected at visits with varying follow-up time. COVID-19 may be an explanation for high attrition, given patient concerns about other health issues, and also an explanation for low use, given competing parent demands during a tumultuous time.

Furthermore, given the survey length, survey completion was relatively low, a documented phenomenon,^48,49 although completers and noncompleters did not differ in baseline characteristics. However, non-COVID-19 clinical weight management programs commonly face high dropout rates.⁵⁰ Moreover, without a control group, the relative impact of the intervention is unknown, an important question especially during COVID-19 when a comparison group could have had higher than typical BMI change.^14,37

In addition, because parent weight was self-reported, social desirability bias may be present and help explain why parent BMI may have significantly dropped and child %BMI_p95 did not. Finally, the small sample size prevented subgroup analyses, which limits the generalizability of our findings.

Study strengths include the entirely virtual format of the intervention, which proved necessary during the COVID-19 pandemic and provided an immediate solution to treat pediatric OW/OB. Importantly, we successfully reached high-need populations who are disproportionately affected by OW/OB and typically lack access due to program availability and/or barriers to access (63% of participants used the DK website).

Aligned with the Obesity Chronic Care Model, DK was integrated into the health care system so that the diagnosis and physiological measurements took place in a clinical setting, while the intervention was home based and accessible online. This format is scalable and aligns with patient preferences. Thus, DK proved to be a response to common barriers of language, internet, time, and transportation for users.

Conclusions

DK, an e-health pilot intervention, demonstrates that e-health interventions may be feasible for some low-income predominantly Hispanic families and may have impact among users in terms of improving parenting behaviors and reducing BMI and child %BMI_p95, even during the COVID-19 pandemic. Larger controlled studies are needed to assess effect size and conduct subgroup analyses, as well as to assess the integration of additional implementation outcomes, to understand low uptake and for effective translation to other settings and populations.

However, a simple solution to childhood weight management is unlikely. Therefore, sustainable scalable interventions that target vulnerable children and that show even modest benefits may contribute to multilevel changes that impact childhood OW/OB.

Impact Statement

We provide a quantitative evaluation of the efficacy of Dynamo Kids!, a pilot, self-paced, customized e-health intervention informed by the Obesity Chronic Care Model, which was implemented in public primary care sites for parents of pediatric patients aged 6–12 years who have overweight and obesity (OW/OB).

Footnotes

Acknowledgment

The authors would like to acknowledge Mr. Lionel Santibáñez for his editorial assistance. The University of North Carolina (UNC) at Chapel Hill Connected Health for Applications and Interventions Core services received support from the National Institutes of Health (grant no. P30DK056350; PI, Mayer-Davis and Shaikh) to UNC Nutrition Obesity Research Center and National Cancer Institute (grant no. P30CA16086; PI, Earp) to the UNC Lineberger Comprehensive Cancer Center. Agency for Healthcare Research and Quality R24 (grant no. HS022418; PI, Halm) provided a pilot award for formative interviews. Institutional support from Children's Health, Children's Medical Center of Dallas for Web site creation.

Authors' Contributions

J.S.Y. drafted the full manuscript, developed the analysis plan, was responsible for manuscript edits, and prepared the manuscript for submission. He also served as a project team member in all phases, from development through evaluation. M.A.A. served as a coprincipal investigator and was responsible for project ideation, implementation, and evaluation. She also provided feedback on the final manuscript. F.D.A. served as a subject matter expert who approved the analysis plan and provided feedback on the final manuscript. C.A.G. and S.E.M. served as subject matter experts who provided feedback on the final manuscript. S.E.B. is the principal investigator and was responsible for project ideation, implementation, and evaluation. She provided significant feedback and insight throughout the manuscript writing process and on the final manuscript.

Ethical Approval

This study was approved by the Institutional Review Board at UTSW and the Committee for the Protection of Human Subjects at UTHealth School of Public Health.

Funding Information

This study was partially supported under a pilot funding mechanism from the AHRQ-sponsored patient-centered outcomes core facility at UT Southwestern (Grant Number: 5R24HS022418-05)

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

References

Khaodhiar

, McCowen

, Blackburn

. Obesity and its comorbid conditions. Clin Cornerstone, 1999; 2(3):17–31.

Nuotio

, Laitinen

, Sinaiko

, et al. Obesity during childhood is associated with higher cancer mortality rate during adulthood: The i3C Consortium. Int J Obes (London), 2022; 46(2):393–399.

Yach

, Stuckler

, Brownell

. Epidemiologic and economic consequences of the global epidemics of obesity and diabetes. Nat Med, 2006; 12(1):62–66.

Acosta

, Streett

, Kroh

, et al. White paper AGA: POWER—Practice guide on obesity and weight management, education, and resources. Clin Gastroenterol Hepatol, 2017;15(5):631–649. e610.

Benjamin

, Virani

, Callaway

, et al. Heart disease and stroke statistics—2018 update: A report from the American Heart Association. Circulation, 2018; 137(12):e67–e492.

Pi-Sunyer

. The medical risks of obesity. Postgrad Med, 2009; 121(6):21–33.

Ben-Sefer

. The childhood obesity pandemic: Promoting knowledge for undergraduate nursing students. Nurse Educ Pract, 2009; 9(3):159–165.

Simmonds

, Llewellyn

, Owen

, et al. Predicting adult obesity from childhood obesity: A systematic review and meta-analysis. Obes Rev, 2016; 17(2):95–107.

Cunningham

, Hardy

, Jones

, et al. Changes in the incidence of childhood obesity. Pediatrics, 2022; 150(2):e2021053708.

10.

Fryar

, Carroll

, Ogden

. Prevalence of overweight, obesity, and severe obesity among children and adolescents aged 2–19 years: United States, 1963–1965 through 2015–2016. NCHS Health E-Stats, 2020.

11.

Barlow

. Expert committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: Summary report. Pediatrics, 2007; 120(Suppl. 4):S164–S192.

12.

Ogden

, Fryar

, Hales

, et al. Differences in obesity prevalence by demographics and urbanization in US children and adolescents, 2013–2016. JAMA, 2018; 319(23):2410–2418.

13.

Kelder

, Springer

, Barroso

, et al. Implementation of Texas Senate Bill 19 to increase physical activity in elementary schools. J Public Health Policy, 2009; 30(1):S221–S247.

14.

Lange

SJKL

, Freedman

, Kraus

, et al. Longitudinal Trends in Body Mass Index Before and During the COVID-19 Pandemic Among Persons Aged 2–19 Years—United States, 2018–2020. CDC: USA; 2021.

15.

Cygan

, Reed

, Lui

, et al. The Chronic Care Model to improve management of childhood obesity. Clin Pediatr (Phila), 2018; 57(6):727–732.

16.

US Preventive Services Task Force; Grossman

, Bibbins-Domingo

, Curry

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

17.

Epstein

, Paluch

, Roemmich

, et al. Family-based obesity treatment, then and now: Twenty-five years of pediatric obesity treatment. Health Psychol, 2007; 26(4):381–391.

18.

Janicke

, Steele

, Gayes

, et al. Systematic review and meta-analysis of comprehensive behavioral family lifestyle interventions addressing pediatric obesity. J Pediatr Psychol, 2014; 39(8):809–825.

19.

Wilfley

, Staiano

, Altman

, et al. Improving access and systems of care for evidence-based childhood obesity treatment: Conference key findings and next steps. Obesity, 2017; 25(1):16–29.

20.

Johnson

, Gupta

, Mackey

, et al. “We feel like we are in it alone”: A mixed-methods study of pediatric primary care barriers for weight management. Child Obes (In press).

21.

Shreve

, Scott

. Adequately addressing pediatric obesity: Challenges faced by primary care providers. South Med J, 2017; 110(7):486–490.

22.

Hammersley

, Jones

, Okely

. Parent-focused childhood and adolescent overweight and obesity ehealth interventions: A systematic review and meta-analysis. JMIRx Med, 2016; 18(7):e203.

23.

Golan

. Parents as agents of change in childhood obesity—From research to practice. Int J Pediatr Obes, 2006; 1(2):66–76.

24.

Hand

, Birnbaum

, Carter

, et al. The RD parent empowerment program creates measurable change in the behaviors of low-income families and children: An intervention description and evaluation. J Acad Nutr Diet, 2014; 114(12):1923–1931.

25.

Yee

, Pfeiffer

, Turek

, et al. Association of the family nutrition and physical activity screening tool with weight status, percent body fat, and acanthosis nigricans in children from a low socioeconomic, urban community. Ethn Dis, 2015; 25(4):399.

26.

Von Elm

, Altman

, Egger

, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies. Ann Intern Med, 2007; 147(8):573–577.

27.

Yudkin

, Nelson

, Quirk et al. A needs assessment of parents and healthcare staff in low-income, diverse populations [in spanish]. Revista Salud Bosque, 2022; 12(1):1–19.

28.

Dietz

, Solomon

, Pronk

, et al. An integrated framework for the prevention and treatment of obesity and its related chronic diseases. Health Aff (Millwood), 2015; 34(9):1456–1463.

29.

Barlow

, Yudkin

, Nelson

, et al. Dynamo Kids!/¡Niños Dinámicos! A website for pediatric primary care providers to offer parents of children 6–12years old with overweight and obesity: Website development and protocol for pilot study. J Pediatr Health Care, 2023; 37(1):17–24.

30.

Tucker

, Howard

, Guseman

, et al. Association between the Family Nutrition and Physical Activity Screening Tool and obesity severity in youth referred to weight management. Obes Res Clin Pract, 2017; 11(3):268–275.

31.

Peyer

, Bailey-Davis

, Welk

. Development, applications, and refinement of the family nutrition and physical activity (FNPA) child obesity prevention screening. Health Promot Pract, 2021; 22(4):456–461.

32.

Laurson

, Eisenmann

, Welk

. Body fat percentile curves for U.S. children and adolescents. Am J Prev Med, 2011; 41(4 Suppl. 2):S87–S92.

33.

Kuczmarski

, Ogden

, Guo

, et al. 2000 CDC growth charts for the United States: Methods and development. Vital Health Stat 11, 2002;(246):1–190.

34.

Little

, Rubin

. Statistical Analysis with Missing Data. Vol. 793. John Wiley & Sons: Hoboken, NJ; 2019.

35.

Martin

, Ferris

. Food insecurity and gender are risk factors for obesity. J Nutr Educ Behav, 2007; 39(1):31–36.

36.

Hemmingsson

. Early childhood obesity risk factors: Socioeconomic adversity, family dysfunction, offspring distress, and junk food self-medication. Curr Obes Rep, 2018; 7(2):204–209.

37.

Tucker

, Eisenmann

, Howard

, et al. FitKids360: Design, conduct, and outcomes of a stage 2 pediatric obesity program. J Obes, 2014; 2014:370403.

38.

Christison

, Daley

, Asche

, et al. Pairing motivational interviewing with a nutrition and physical activity assessment and counseling tool in pediatric clinical practice: A pilot study. Child Obes, 2014; 10(5):432–441.

39.

Chiong

, Gray

, Roy

. Development of a family-based nutrition program rooted in food parenting literature. Fam Consum Sci Res J, 2020; 49(1):67–83.

40.

Bailey-Davis

, Kling

SMR

, Wood

, et al. Feasibility of enhancing well-child visits with family nutrition and physical activity risk assessment on body mass index. Obes Sci Pract, 2019; 5(3):220–230.

41.

Ihmels

, Welk

, Eisenmann

, et al. Prediction of BMI change in young children with the family nutrition and physical activity (FNPA) screening tool. Ann Behav Med, 2009; 38(1):60–68.

42.

Lange

, Kompaniyets

, Freedman

, et al. Longitudinal trends in body mass index before and during the COVID-19 pandemic among persons aged 2–19years - United States, 2018–2020. MMWR Morb Mortal Wkly Rep, 2021; 70(37):1278–1283.

43.

Rhee

, Herrera

, Strong

, et al. Guided self-help for pediatric obesity in primary care: A randomized clinical trial. Pediatrics, 2022; 150(1):e2021055366.

44.

Wilson

, Sweeney

, Law

, et al. Web-based program exposure and retention in the families improving together for weight loss trial. Ann Behav Med, 2019; 53(4):399–404.

45.

Zoellner

, You

, Hill

, 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.

46.

Gaglio

, Shoup

, Glasgow

. The RE-AIM framework: A systematic review of use over time. Am J Public Health, 2013; 103(6):e38–e46.

47.

Glasgow

, Harden

, Gaglio

, et al. RE-AIM planning and evaluation framework: Adapting to new science and practice with a 20-year review. Front Public Health, 2019; 7:64–64.

48.

Liu

, Wronski

. Examining completion rates in web surveys via over 25,000 real- world surveys. Soc Sci Comput Rev, 2018; 36(1):116–124.

49.

Farhat

, Sharma

, Abrams

, et al. Kamp K'aana, a 2-week residential weight management summer camp, shows long-term improvement in body mass index z scores. J Pediatr Gastroenterol Nutr, 2016; 62(3):491–494.

50.

Skelton

, Goff

Jr ., Ip

, et al. Attrition in a multidisciplinary pediatric weight management clinic. Child Obes, 2011; 7(3):185–193.

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

0.04 MB

0.05 MB