The Role of Parents in Adolescent Obesity Treatment: Results of the TEENS+ Randomized Clinical Pilot Trial

Abstract

Background:

The optimal role for involving parents in adolescent obesity treatment is unknown.

Objective:

To demonstrate that two parent approaches within adolescent obesity treatment are distinct, as evidenced by differential parent outcomes, and determine the preliminary efficacy of each approach on adolescent weight loss.

Methods:

Adolescent/parent dyads (N=82; mean adolescent age=13.7±1.2y) participated in TEENS+, a 4-month behavioral weight loss treatment. Participants were randomized to: 1) TEENS+parents as coaches (PAC; parent skills training), or 2) TEENS+parent weight loss (PWL; adult behavioral weight loss). Assessments occurred at 0, 4 (post-treatment), and 7-months. Within- and between-group repeated measures general linear mixed models examined change in parent weight (Δkg; primary outcome); parenting, feeding, weight control strategies, home environment, and adolescent body mass index (ΔBMI; secondary outcomes).

Results:

PWL parents had greater 4-month weight losses (Δkg_0–4m=−5.14±4.87kg) compared to PAC (−2.07±3.89kg; between-group p<.01). Key constructs differed between groups as expected. Both groups yielded significant within-group adolescent ΔBMI_0–4m (PWL: −0.97±1.38kg/m² vs. PAC: −0.93±1.42kg/m²; ps<.01); during maintenance, adolescents in PWL had ΔBMI_4–7m increases (+0.41±1.07kg/m²; p=.02) while PAC did not (+0.05±1.31kg/m²; p=.82).

Conclusion:

Parent treatments were distinct in implementation and both yielded significant 4-month adolescent weight loss. Differential weight patterns were observed during maintenance, favoring PAC, warranting further exploration.

INTRODUCTION

One-in-five U.S. adolescents ages 12–19 years have obesity, a condition associated with significant health risks.¹ Clinical guidelines consistently emphasize the importance of involving parents in adolescent obesity treatment;^2,3 yet, evidence is limited regarding how to involve parents optimally.^4–6 In a recent review, only five clinical trials had experimentally manipulated parents’ role in adolescent obesity treatment; many of these trials lacked details on parent-related behavioral strategies, parent outcomes, or relations between parent and adolescent outcomes.⁶ Normative developmental factors, such as opposition to authority, role transformations, and increasing personal responsibility,^2,7 make optimizing parental involvement in treatment particularly challenging during adolescence. Thus, rigorous investigations of the optimal role of parents in adolescent obesity treatment are urgently needed.

Previous adolescent obesity treatments involving parents typically include building parent skills and self-efficacy to support adolescents’ weight management via role modeling, positive reinforcement, authoritative parenting, home environment change, and enhancing communication skills.^6,8 These treatments generally yield reductions in adolescent body mass index (BMI)^6,9–14 and have identified parent variables that might influence adolescent weight loss. For example, authoritative parenting, characterized by both responsiveness (i.e., being attuned and supportive to foster autonomy) and demandingness (i.e., providing structure to cultivate responsibility), is associated with improved adolescent weight outcomes, although the lack of longitudinal investigations limits conclusions.^15,16 Authoritative feeding (e.g., providing choice within parameters) is also associated with younger children’s improved weight outcomes.^15,16 These associations also appear to exist among adolescents, with greater parental monitoring and restriction associated with higher adolescent BMI.^16,17 However, most research is cross-sectional.¹⁶ Longitudinal investigations that promote authoritative parenting and feeding within adolescent obesity treatment are needed to identify potential benefits of this approach, compared with alternative parent approaches, on adolescent weight loss.

In family-based obesity treatment for younger children, parent weight loss is consistently associated with child weight loss.¹⁸ Emerging evidence has also identified these relations among adolescents;^11,13 yet mechanisms to explain them remain unclear. For example, parent BMI reduction was the only independent predictor of adolescent BMI reduction at the end of a 4-month obesity treatment.¹¹ In another trial, parents who lost more weight during treatment had adolescents with a greater percent weight loss at 12-month follow-up, suggesting that engaging parents in weight loss efforts might help with adolescent weight loss maintenance.¹³ Importantly, parent weight loss was not directly targeted in these prior trials. Directly targeting parent weight loss might yield greater adolescent weight loss, as parents engaged in concurrent weight management might make greater changes to the shared environment and serve as powerful role models; yet this approach has not been tested directly.⁴

Given this emerging evidence, two potential approaches to engage parents in adolescent obesity treatment warrant further investigation: 1) parent skills training based on authoritative feeding and parenting, and 2) concurrent parent behavioral weight loss. We conducted a pilot randomized clinical trial (RCT) of these parent approaches within an evidence-based adolescent behavioral obesity treatment program (TEENS+). Specifically, adolescents and parents were randomized to one of two 4-month treatments, matched on contact: 1) TEENS+Parents as Coaches (PAC), or 2) TEENS+Parent Weight Loss (PWL).

Study Aims

The primary aim of this pilot was to determine whether PAC and PWL were distinct treatments, as evidenced by differential impact on key variables consistent with each approach. Specifically, we examined the within- and between-group effects of each treatment on: 1) parent weight change (kg; primary outcome); 2) parenting variables (e.g., child feeding, authoritative parenting, parental modeling and use of weight control strategies); and 3) the home food environment, from baseline to post-treatment (4-month). Our primary hypothesis was that at 4-months, PWL parents would have greater reductions in weight and improvements in weight control behaviors, compared to PAC, and PAC would have greater positive impacts on parenting and feeding variables, compared to PWL. Although both parent approaches had theoretical and empirical support, this pilot was critical to first demonstrate that the parent treatments were distinct in implementation, as evidenced by differential effects on parent weight (primary aim) as well as parenting and home environment variables, thus supporting the internal validity of each approach and setting the stage to subsequently evaluate their impact on adolescent weight outcomes in a fully-powered trial.

Towards that end, the second aim was to explore preliminary within-group efficacy of TEENS+PAC and TEENS+PWL on adolescent BMI change from baseline to 4-months. We also evaluated adolescent BMI change during a no-treatment maintenance phase, from 4- to 7-months. We hypothesized that adolescents in both treatments would demonstrate significant 4-month BMI reductions. Although not powered to detect between group differences in adolescent BMI, we sought to determine whether each approach warranted additional study in a fully-powered trial, evidenced by adolescent BMI reductions and participant satisfaction.

METHODS

Participants and Recruitment

Participants were recruited through pediatrician referrals, direct mailings, and community events. Interested families completed a screener via telephone, a secure website, or in-person. Eligible families were invited to a group orientation in which study procedures were described, eligibility confirmed, and written consent/assent obtained. Upon consent, baseline measures were completed, which included final eligibility screening (i.e., medical evaluation, depression and eating disorder screenings). Eligible adolescents were ages 12–16 years, with BMI≥85^th percentile for age and sex¹⁹ who primarily resided in the participating caregiver’s home. Exclusion criteria included: 1) a clinical eating disorder (anorexia or bulimia nervosa; recent compensatory behavior) or significant psychopathology (e.g., suicidality, psychosis), 2) inability to follow the protocol due to a physical or developmental disorder (e.g., mobility or severe autism), 3) change in medications that could impact weight (e.g., antidepressants) or use of atypical antipsychotics or steroids within the past 3 months; 4) diabetes mellitus; and 5) medical conditions known to impact weight or eating (e.g., Prader Willi Syndrome, bariatric surgery). Parents were eligible if they were ≥18 years with BMI≥25 kg/m², with the same exclusion criteria as those applied for adolescents; however, parents with diabetes were eligible if their dose of diabetes medications was stable over the past 3 months. Adolescents and parents received medical clearance prior to participation.

Eligibility screening was completed by 224 families (91% online); 101 families attended orientation and consent visits. Ultimately, 82 dyads (n=82 adolescents, n=80 parents [2 sibling pairs]) were randomized. (Figure 1).

Figure 1.

CONSORT

Note: participants who completed at least the primary outcome at 4m and 7m (fasting height and weight measurements) are included in these numbers.

Design

This RCT, conducted in 2016–18, included 2 study groups (TEENS+PAC; TEENS+PWL) undergoing 4-months of treatment delivered via weekly in-person group sessions, with a 3-month maintenance phase (during which no intervention contact was provided). Assessments were obtained at baseline (0), post-treatment (4-month), and follow-up (7-month). Adolescent-parent dyads were stratified by adolescent sex and randomized (1:1) using a computer-generated variable block randomization schedule to TEENS+PAC or TEENS+PWL. The study coordinator informed participants of their group night (day of the week only) following completion of baseline measures; study group (PAC or PWL) was revealed by behavior coaches at the first treatment session. All study activities took place at an academic medical center in Virginia (NCT#02586090). This study was approved by the Institutional Review Board of Virginia Commonwealth University.

Interventions

Theoretical Foundation.

The parent and adolescent interventions were informed by our prior work^9,20,21 and grounded in Social Cognitive Theory (SCT), via its emphasis on interactions among environmental, personal and behavioral factors, capitalization on social learning, and skill building to enhance self-efficacy for weight management.²² The group format was designed to enhance peer support, with strategies taught to modify one’s environment (physical, emotional, cognitive, and social) and response patterns to support goal attainment. Core strategies that form the basis of behavioral weight control interventions^23,24 were integrated into all sessions. The interventions were further informed by Self-Determination Theory through use of motivational interviewing (MI), highlighting autonomy, evoking participants’ reasons for change, and building self-efficacy.²⁵ Importantly, the two parent treatments had distinct foci: PAC focused on use of theoretically-guided parenting strategies to support adolescents’ weight loss efforts; PWL used a theoretically-guided behavioral approach focusing on parents’ own weight loss.

Tailoring for Inclusivity.

TEENS+ served a sociodemographically diverse sample with respect to race/ethnicity (~50% non-Hispanic Black) and income (including many families from lower socioeconomic backgrounds). The interventions were thus designed with particular sensitivity to Black cultural values and traditions.²⁶ This included modifying traditional meals with healthy adaptations, acknowledging and incorporating racial differences in body image ideals, and recognizing extended kinship networks.^27,28 The treatments also acknowledged unique pragmatic challenges of single parent families and families from lower-income backgrounds by focusing on feasible behavioral strategies, emphasizing accessible and free activity options, teaching healthy and budget conscious food shopping strategies, and recognizing relevant policies (e.g., school meals) that might impact diet.

TEENS+ Adolescent Intervention.

All adolescents, regardless of treatment assignment, participated in TEENS+. Adolescents were in groups according to their parents’ treatment; each group included a single sex to maximize perceived similarity among group members (aligned with SCT²²) and enhance comfort with discussions about weight and related topics, consistent with preferences expressed by this population.²⁹ Groups were led by two trained behavior coaches (psychology doctoral trainees or similar and registered dietitians; masked to study hypotheses). Over 16 weeks, adolescents participated in weekly 1-hour group visits and 1-hour supervised exercise sessions, occurring on the same evening. Adolescents also attended one 30-minute individual visit per month. Two of these individual visits were with their behavior coach who used an MI-informed approach to explore adolescents’ values, reasons, and strategies for change, to enhance autonomy and engagement, and build self-efficacy for behavior change. The remaining two individual visits were with their dietitian who used an MI-informed approach to guide dietary modifications to yield an energy deficit and help adolescents meet their stated weight loss goals. Parents and adolescents were invited to attend a single cooking class (combined with all parents and adolescents in a treatment arm) during treatment.

Manualized sessions followed a behavior therapy approach, including guided goal-setting and self-monitoring, identifying barriers and solutions, contingency management, stimulus control, setbacks, and relapse prevention. Weight was assessed weekly prior to group during brief, individual check-ins with the behavior coaches. Participants kept food and physical activity (PA) logs. Coaches reviewed logs and provided weekly personalized feedback in an MI-consistent manner, using a self-regulation framework to teach adolescents strategies to adjust their eating activity patterns to reach their weight management goals. This process of healthy self-regulation of energy balance behaviors includes: self-monitoring (of diet, PA, and weight), self-evaluation (comparing self-monitoring data to goals), and self-reinforcement and/or problem-solving and setting new strategies (reacting to self-evaluation).³⁰

The TEENS+ dietary intervention (see Raynor et al.³¹) was designed to yield a caloric deficit via adding low calorie, nutrient dense foods (“Go Foods”) while remaining within a prescribed calorie goal, that aimed for a 1–2lb/week weight loss. Individual calorie (1200–1400kcal/day for girls; 1500–1800kcal/day for boys; based on sex and baseline height) and Go Food (based on baseline food records) goals were adjusted throughout treatment as needed (i.e., based on activity level and rate of weight change), with Go Food goals increased as previous goals were met. This evidence-based approach was based on prior research which found that, within the context of a calorie goal, targeting an increase in healthy foods vs. reduction in unhealthy foods was associated with greater BMI reduction.³² Importantly, interventions targeting an increase in fruits and vegetables in the absence of a calorie goal did not result in a reduction of fat and sugar.³³ Thus TEENS+ focused on adding Go foods (and adolescents must reduce higher energy foods to remain within calorie goals). Evidence-based lessons provided education about energy balance, macronutrients, and high-risk eating behaviors associated with obesity (e.g., sugared beverage intake and meals away from home). Personalized weekly strategies were collaboratively set to help adolescents achieve dietary goals.

The TEENS+ exercise intervention was designed to help adolescents achieve U.S. Department of Health and Human Services guidelines of ≥1 hour/day of PA, with a focus on increasing moderate to vigorous activity and incorporating ≥3 days/week of resistance training.³⁴ Exercise progressions were provided that gradually increased PA duration, frequency, and intensity. Personalized weekly strategies were collaboratively set to help adolescents achieve these targets, with a focus on realistic, sustainable strategies. Adolescents could exercise ≥1x/week at the program gym (on their group night and other “open gym” sessions, by treatment group); at the YMCA (parent and adolescent memberships provided); or other locations. Exercise sessions were designed to expose participants to a range of physical activities, build competency and self-efficacy in a supportive environment, and help adolescents meet PA targets.

Parent Interventions.

While adolescents were in their treatment groups, parents participated in their assigned intervention (PAC or PWL), matched on contact, and led by trained, supervised behavior coaches who were masked to study hypotheses. See Table 1 for a detailed comparison of the parent interventions, briefly described below.

Table 1.

Distinctions between the two parent interventions in the TEENS+ behavioral weight loss randomized clinical trial: Parents as Coaches (PAC) and Parents Weight Loss (PWL)^a

	Parents as Coaches	Parent Weight Loss
Theoretical Framework	• grounded in Social Cognitive Theory; focused on making changes to the family environment to support family-level healthy eating and activity	• grounded in Social Cognitive Theory; focused on making changes to one’s environment to support healthy eating and activity
	• infused with the spirit of Motivational Interviewing via its focus on autonomy support and eliciting participant-derived strategies and goals that are personally-relevant, aligned with their values, and perceived as sustainable	• infused with the spirit of Motivational Interviewing via its focus on autonomy support and eliciting participant-derived strategies and goals that are personally-relevant, aligned with their values, and perceived as sustainable
Treatment Structure/Content	• 16-weeks; 60 min group sessions	• 16-weeks; 60 min group sessions
	• led by trained behavioral coaches	• led by trained behavioral coaches
	• 8 behavioral parent skills training lessons (parent only) & 8 nutrition education sessions (combined with adolescents )	• 16 parent-only behavioral weight loss lessons (including 8 nutrition education sessions)
Treatment Focus	• adolescent weight management via parent skills training based on authoritative parenting approach	• parent weight loss via lifestyle intervention focused on modifications to diet, activity and behavior
	• taught how to help their adolescent meet dietary goals via meal planning, preparing meals at home, and home environment changes	• personalized diet goals based on initial weight (e.g., 1200–1800 kcal, 30% kcal from fat; adjusted as needed based on progress)
Dietary Approach	• problem-solved around challenges to adherence (e.g., schedules, food preferences, cost and stress)	• weekly weights obtained and adjustments made to address rapid, unsafe weight loss if needed
	• encouraged to role model healthy eating	• taught problem-solving skills to address challenges to adherence (e.g., access, cost, stress, and social influence)
	• emphasized the importance of all family members making dietary changes	• taught to make adjustments to dietary intake to maintain weight loss after the program
	• parents learned practical (free and low-cost) strategies to increase family PA and to support their adolescent meeting their PA goals	• personalized PA progression to shape behavior and promote self-efficacy with the end goal of ≥250 minutes/week of moderate intensity PA
Physical Activity (PA) Approach	• behavioral goals focused on increasing opportunities for adolescents to engage in PA and reducing / setting limits on screen time	• multiple short bouts of PA encouraged (e.g., ≥10 min) to minimize potential barriers
	• psychoeducation and problem-solving around overcoming family and adolescent barriers to PA	• psychoeducation and problem-solving around challenges to PA (e.g., time, motivation, access)
	• parents kept personal logs of family-based change behaviors (e.g., meal planning, preparing meals at home, Go food and exercise opportunities for their teen) and self-monitored progress	• parents kept personal logs including detailed food records and minutes/type of PA
Self-Monitoring, Reporting and Feedback	• coaches reviewed logs weekly & provided written feedback to affirm participant progress and provided personalized suggestions to assist with goal attainment	• coaches reviewed logs weekly & provided written feedback to affirm participant progress and provided personalized suggestions to assist with goal attainment
	• parents reviewed adolescent logs and assisted as needed	• parents taught to self-weigh several times per week (daily recommended) and use this information to evaluate their progress and make adjustments in their behavior as needed
	• parents met briefly with coach individually before group to facilitate personalized feedback within self-regulation framework / brainstormed barriers	• parents met briefly with coach individually before group to obtain weigh and facilitate personalized feedback within self-regulation framework / brainstormed barriers
Core Behavioral Strategies	• parents taught core behavioral / parenting strategies to help implement family changes and support their adolescents’ weight management	• parents taught core behavioral and cognitive skills to help implement personal changes for weight loss

Note:

^aAll adolescents participated in the TEENS+ behavioral weight loss intervention as described in the Methods.

Parents as Coaches:

PAC taught parent skills training based on authoritative parenting and feeding strategies to support adolescents’ weight management. PAC emphasized parent role modeling, providing Go Food and exercise opportunities, and making home environment changes to foster adolescents’ healthy eating and exercise behaviors. Parent weight management was not specifically addressed, and parent weight was not monitored. PAC included 16 weekly sessions, alternating between 8 parent skills training sessions (while adolescents attended their own groups), and 8 nutrition education sessions (which parents attended with adolescents). Parents self-monitored key parenting behaviors, and coaches provided weekly personalized written feedback within a self-regulation framework that taught parents to identify connections among their behaviors and their adolescents’ weight management, and subsequent weekly weight change.

Parent Weight Loss.

In PWL, parent and adolescent behavioral weight loss treatment was concurrent, but independent. Parents were provided with a weight loss goal of 1–2lbs/week, personalized calorie, fat and PA goals, and instructions to self-monitor key information (e.g., weight, dietary intake, and PA). PWL emphasized consuming Go Foods to meet calorie goals, with psychoeducation provided regarding energy density and diet quality. Parent weight was measured weekly throughout the study duration and feedback provided. Parents received training in core behavioral weight loss strategies (e.g., goal-setting, stimulus control) and techniques to help them meet their own diet and PA goals.³⁵ They received personalized feedback within a self-regulation framework to identify how their eating and exercise behaviors were associated with weight change. PWL did not address how to support adolescents’ weight management directly, although emphasized that their concurrent engagement in these behaviors should be helpful for their adolescents.

Retention Protocols

To enhance engagement, raffle tickets reinforced parent and adolescent attendance and log completion, consistent with behavior theory. Raffles were drawn at variable intervals and the winner selected a prize (e.g., items ≤$10). Group incentives (e.g., water bottles) were provided monthly. Coaches contacted participants after a missed visit and provided materials via email. Families also received payments for follow-up assessment visits ($40 at 4-months, $50 at 7-months). Integration of MI was also designed to enhance retention, consistent with prior studies.^36,37

Training and Fidelity Monitoring

Coaches were masked to study hypotheses and trained following the Operations Manual, which included arm-specific training in adolescent behavioral weight loss, parent skills training, and/or adult behavioral weight loss. All coaches were trained in culturally sensitive intervention delivery.²⁷ Coaches conducted mock treatment sessions and received detailed feedback prior to intervention initiation. All intervention sessions were audiotaped and reviewed in weekly supervision. Coaches completed fidelity checklists after each session, and trained independent assessors reviewed a random sample of 10% of audiotapes and completed fidelity checklists to assess protocol adherence.

Measures

Assessments were conducted by masked assessors at 0, 4 (post-treatment), and 7 (follow-up) months, unless otherwise noted. Surveys were completed using the secure online data capture platform, REDCap (Research Electronic Data Capture).³⁸

Sociodemographics.

At baseline, parents reported parent and adolescent age, sex, race, and ethnicity; household income; and insurance status.

Anthropometrics.

Parent and adolescent height and weight were measured after a 12-hour fast in light clothing to the nearest 0.1cm and 0.1kg using a wall-mounted stadiometer (Seca 213) and electronic digital scale (Scale-Tronix 5125-X), respectively. The average of three measures was used. Parent and adolescent BMI (kg/m²) and adolescent BMI percentile (using Epi Info Software³⁹) were calculated.

Parenting Style.

Adolescents completed the 16-item Authoritative Parenting Index⁴⁰ (API), a widely used measure with strong psychometric properties. The API assessed the participating parents’ general parenting style across two domains: Responsiveness (sum of 9 items; e.g., “[parent] makes me feel better when I’m upset”) and Demandingness (sum of 7 items; e.g., “[parent] has rules that I must follow). Adolescents indicated how much each item is similar to their parent, ranging from 1 (“not like”) to 4 (“just like”). Subscales and total scores were calculated. Internal reliabilities ranged from 0.82–0.90, with the maternal demandingness subscale α=0.69.

Parent Feeding Practices.

Parents completed the Child Feeding Questionnaire (CFQ; Adolescent version)¹⁷ to evaluate beliefs, attitudes and practices related to feeding. For this study, the Restriction (8 items; e.g., “I have to watch out that my teen does not eat too many of his/her favorite foods”) and Monitoring (4 items, e.g., “How much do you keep track of the snack food that your teen eats”) factors were examined. Parents responded on a scale ranging from 1 (“disagree”/“never”) to 5 (“agree”/“always”), with higher scores suggesting greater restriction or monitoring, respectively. The CFQ has been validated in a multi-ethnic sample of adolescent parents.¹⁷ Internal reliabilities were adequate for both restriction (α=0.78) and monitoring (α=0.94).

Parent Weight Control Behaviors/Role Modeling.

The Weight Control Strategies Scale (WCSS)⁴¹ assessed parents’ use of healthy weight control practices (for their own weight management), and was also used as a measure of parent modeling, on three domains: dietary choices (10 items), self-monitoring (7 items), and PA (6 items). Parents reported how frequently they engaged in the target behavior, from 0 (never) to 4 (always). The mean of each subscale and the total score (overall mean) were calculated (range 0–4); internal reliabilities were adequate (α=0.80–0.93). Adolescents completed the 4-item diet modeling subscale of the Family Experiences Related to Food Questionnaire (FERFQ; mother [(α=0.66] or father [α=0.69] version based on the participating parent) to assess perceptions of parental modeling of dieting behaviors,⁴² consistent with its prior use in this manner.¹⁰

Home Food Environment.

Parents completed the Home Food Inventory (HFI)⁴³ to assess the types of foods and beverages available in the home. The HFI includes 190 items; parents indicated whether each item was present in their home (yes [1], no [0]). The obesogenic food availability score was calculated (based on 71 items), with lower scores suggesting a less obesogenic environment. The HFI has demonstrated validity, with the obesogenic food availability score significantly associated with both parent and adolescent energy intake in validation samples.

Exit Survey.

Parents completed exit surveys at 4-months, assessing: intervention likes/dislikes; thoughts about duration, frequency, and number of sessions; perceived benefits and barriers to implementing the intervention goals; overall satisfaction; and suggestions for improvement.

Statistical Analyses

Analyses were performed with SPSS v26.0. A priori power analyses revealed that n=67 adolescent/parent dyads would provide 80% power to detect significant (α=0.05, 2-sided) between-group difference of −2.8 to −3.8kg for parent weight loss (primary outcome) at 4-months. Variance and effect size estimates were based on our pilot work engaging parents as helpers in adolescent behavioral weight loss, as well as prior adult behavioral weight loss trials.^44,45

Descriptive statistics explored differences in attrition at 4- and 7-months by study group and participant demographics (adolescent race, sex, age and baseline BMI; parent race, weight, and income), using independent t-tests and chi-square analyses for continuous and categorical variables, respectively. One outlier on the primary outcome (parent weight change; Δkg_0–4m) was identified, with 4-month weight loss >3 standard deviations above the mean. To retain all participants, winsorization was applied by identifying the next closest value for Δkg_0–4m and applying this degree of change to the outlier’s 4-month weight value.⁴⁶

Primary analyses for parent and adolescent weight outcomes used an intent-to-treat (ITT) approach to compare parent weight change and adolescent BMI change between the two groups, wherein the non-completers were assumed to have returned to baseline. These results were compared to completer analyses (for individuals with height and weight data at 4- and/or 7-month; see Figure 1). This approach to imputation was used given the relatively few studies investigating adolescent weight loss outcomes, consistent with prior studies.^47,48 For Aim 1, repeated measures general linear mixed models (GLM) were used to examine within and between-group differences in parent weight change (kg) from baseline to 4-months (post-treatment), controlling for parent race.⁴⁹ Between-group models included a study group by time interaction to examine changes in parent weight from baseline to 4-months. Data were then stratified by study group to examine within-group patterns of time predicting parent weight change. Similar repeated measures GLMs also examined within- and between-group effects on parenting and psychosocial constructs (API, WCSS, HFI, FERFQ, CFQ) from baseline to 4-months, using completers only.

For Aim 2, data were stratified by study group to examine within-group differences in adolescent BMI from baseline to post-treatment (ΔBMI_0–4m) and post-treatment to follow-up (ΔBMI_4–7m), using repeated measures GLM. Although not powered to detect changes in adolescent weight outcomes, the between-group differences in adolescent BMI at both timepoints were also explored. Note that for weight and BMI outcomes, GLM p-values and effect sizes (partial n²) are reported, with mean±SD change scores presented for interpretation.

RESULTS

Participants, Retention, and Attendance

Please see Table 2 for participant demographics. At 4-months, 85% of randomized dyads were retained; 81% were retained at 7-months. There were no differences in retention by study group, or adolescent or parent demographics (ps>.05). However, adolescents not retained at 4-months (n=12: 38.6±8.4 kg/m²) had higher baseline BMIs, compared with adolescents who were retained (n=70: 34.2±6.6 kg/m²; p=.046). There were no differences in 7-month retention by adolescent baseline BMI (p=.13). Parents and adolescents each attended an average of 71% of group sessions and adolescents attended an average of 82% of individual sessions, with no differences in attendance by study group (ps>.05).

Table 2.

Baseline demographics, weight, and body mass index of adolescents and parents enrolled in the TEENS+ pilot, presented by study group

	Adolescents (n=82)		Parents (n=80)
	PAC (n=40)	PWL (n=42)	PAC (n=39)	PWL (n=41)
Female, n (%)	26 (65.0)	26 (61.9)	34 (87.2)	34 (82.9)
Race, n (%)
African American/Black	21 (52.5)	17 (40.5)	20 (51.3)	16 (39.0)
White	18 (45.0)	23 (54.8)	19 (48.7)	21 (51.2)
Asian	1 (2.5)	2 (4.8)	1 (2.6)	2 (4.9)
Native American	1 (2.5)	1 (2.4)	1 (2.6)	0
Other	2 (5.0)	3 (7.1)	0	2 (4.9)
Hispanic, n (%)	2 (5.0)	2 (4.9)	0	2 (4.9)
Insurance, n (%)
None	--	--	0	2 (4.9)
Medicaid	--	--	7 (17.9)	10 (24.4)
Private Insurance	--	--	32 (82.1)	29 (70.7)
Annual income, n (%)
<$10,000	--	--	1 (2.6)	1 (2.4)
$10,000–$19,999	--	--	4 (10.3)	3 (7.3)
$20,000–$29,999	--	--	6 (15.4)	2 (4.9)
$30,000–$39,999	--	--	2 (5.1)	4 (9.8)
$40,000–$49,999	--	--	1 (2.6)	1 (2.4)
$50,000–$74,999	--	--	9 (23.1)	14 (34.1)
$75,000–$99,999	--	--	6 (15.4)	7 (17.1)
$100,000–$149,999	--	--	7 (17.9)	9 (22.0)
≥$150,000	--	--	3 (7.7)	0 (0.0)
Age (yrs), mean (SD)	13.6 (1.1)	13.8 (1.3)	45.1 (7.2)	43.4 (6.8)
Weight (kg), mean (SD)	93.5 (16.7)	95.1 (21.8)	97.1 (19.6)	99.6 (20.4)
BMI (kg/m2), mean (SD)	34.7 (6.2)	35.0 (7.7)	36.1 (7.2)	36.3 (7.4)
BMI percentile, mean (SD)	98.4 (1.4)	98.3 (1.5)	--	--

Note: n=1 parent and n=2 children declined to provide race information; participants could select more than 1 racial category; thus, percentages do not total 100%.

PAC=Parents as Coaches

PWL=Parent Weight Loss

BMI=Body Mass Index

Parent Outcomes

Within- and between-group changes in parent weight and psychosocial outcomes were examined for PAC and PWL. (Table 3). Parents in PWL had significantly greater 0–4 month weight losses (Δ−5.14±4.87kg) compared to parents in PAC (Δ−2.07±3.89kg; p<.01). Completer analyses were similar (PAC: Δ−5.86±4.78kg; PWL: Δ−2.53±4.16kg; p<.01). Parents in both groups reported increased use of weight control strategies, with significantly greater use of weight control strategies (overall and self-monitoring) in PWL, compared to PAC. Further, adolescents in PWL were significantly more likely to perceive that their parents were “dieting,” based on the FERFQ (p=.02). Both PAC and PWL parents reported significant improvements to the home food environment and decreased use of restrictive feeding practices. Adolescents in PAC reported increases in authoritative parenting while adolescents in PWL reported decreases (p=.08, between groups), with significant group differences in the parental responsiveness domain of this construct showing increased responsiveness in PAC, with no change for PWL (p=.01).

Table 3.

Baseline to post-treatment (4-month) within- and between-group changes in key parent and adolescent variables for TEENS+PAC and TEENS+PWL

	Parents as Coaches (PAC) (n=39 parents / 40 adolescents)a			Parent Weight Loss (PWL) (n=41 parents / 42 adolescents)b			Between Groupi
	Baseline	4-month	p	Baseline	4-month	p	p	Partial n2
Parent Measures
Weight (kg)c	97.10±19.62	95.03±19.68	<.01	99.59±20.36	94.45±21.30	<.01	<.01	.11
Parent Feeding Practicesd
Restriction	3.71±0.53	3.04±0.95	<.01	3.50±0.82	3.17±0.81	0.04	.14	.04
Monitoring	3.53±0.90	3.80±0.94	.09	3.20±1.10	3.45±0.93	.18	.92	<.01
Obesogenic Home Food Environmente	19.74±8.97	11.71±7.74	<.01	17.57±7.99	11.63±10.15	<.01	.34	.02
Weight Control Strategiesf
Total Score	1.23±.56	1.90±.75	<.01	1.40±.81	2.58±.72	<.01	.02	.09
Dietary Choices	2.34±.59	2.93±.67	<.01	2.40±.78	2.82±.76	.02	.42	.01
Self-Monitoring	0.64±.62	1.48±1.10	<.01	0.79±.84	2.94±.99	<.01	<.01	.29
Physical activity	0.91±.97	1.51±1.09	.01	1.23±1.18	2.35±.95	<.01	.11	.04
Adolescent Measures
BMI (kg/m2)	34.74±6.24	33.82±6.69	<.01	34.97±7.74	34.00±8.00	<.01	.89	<.01
Authoritative parentingg	50.13±7.23	52.19±7.56	.11	50.64±7.32	49.52±8.43	.38	.08	.05
Responsiveness domain	26.97±5.28	29.71±4.68	<.01	27.58±4.78	27.39±5.23	.80	.01	.12
Demandingness domain	23.16±3.16	22.48±4.49	.32	23.01±4.08	22.12±4.09	.22	.80	<.01
Parent diet modelingh	10.32±3.12	10.16±2.41	.78	9.06±2.88	10.82±3.69	<.01	.02	.09

Note: PAC=Parents as Coaches; PWL=Parent Weight Loss; All randomized participants were included in weight and BMI analyses, using an intent to treat approach; Only completers were included in analyses of psychosocial outcomes.

^aPsychosocial outcomes n=32 parents / 31 adolescents

^bPsychosocial outcomes n=33 parents / 33 adolescents

^canalyses controlled for parent race

^dChild Feeding Questionnaire

^eHome Food Inventory obesogenic food availability score- higher scores suggest more obesogenic environment

^fWeight Control Strategies Scale

^gAuthoritative Parenting Index (adolescent report)

^hFamily Experiences Related to Food Questionnaire, Diet Modeling subscale (father or mother version used as appropriate based on the participating parent)

ⁱEstimates of treatment effects on primary and secondary outcomes obtained using repeated measures general linear models; the study was powered on parent weight change.

Adolescent weight loss

TEENS+PAC and TEENS+PWL both yielded significant within-group adolescent ΔBMI_0–4m of −0.93±1.42kg/m² and −0.97±1.38kg/m², respectively. (Table 3). Among completers, results were similar (PAC=1.12±1.50kg/m²; PWL= 1.16±1.44kg/m²). Examination of ΔBMI_4–7m during the maintenance phase, however, suggests the potential of PAC to confer better adolescent weight loss maintenance. Specifically, within-group analyses indicated that adolescents in PWL had a significant ΔBMI_4–7m increase (+0.41±1.07kg/m²; p=.02), while PAC did not (+0.05±1.31kg/m²; p=.82). This between group difference corresponded to a small effect size (partial η² =.02; p=.18; see Figure 2). Completer analyses were similar (PWL: +0.47±1.20 kg/m²; PAC: +0.07±1.52 kg/m²; p=.25).

Figure 2.

Change in adolescent BMI in TEENS+PAC and TEENS+PWL at post-treatment (ΔBMI_0–4m) and after a 3-month no-contact maintenance phase (ΔBMI_4m-7m), with standard error bars presented.

Fidelity and Participant Satisfaction

Fidelity monitoring indicated that TEENS+PAC and TEENS+PWL were delivered as intended, with no aberrations from the protocol. Program satisfaction was high, with no differences between groups, based on exit surveys. On a scale of 0–10, with 10 being the most satisfied, parents were very satisfied with both PAC (9.1±1.0) and PWL (9.5±0.8). Almost all (97%) parents in both groups agreed/strongly agreed that they enjoyed attending groups, and 90% of PAC and 100% of PWL parents agreed/strongly agreed that they would recommend other families participate in the program. The most common barriers to attendance were competing priorities related to, family (42%), school (39%), or work (34%), highlighting the need to minimize participant burden when possible. Lastly, 97% of parents reported that the program addressed their concerns, with no differences between study groups.

DISCUSSION

This pilot trial sought to demonstrate that two approaches to involving parents within adolescent obesity treatment—parent skills training and concurrent parent behavioral weight loss—were distinct and differentially impacted key outcomes, consistent with their focus. It also sought to determine the preliminary efficacy of each parent approach on adolescent weight outcomes, evidenced by within-group adolescent BMI change. Both parent and adolescent outcomes demonstrated that the approaches were distinct in implementation. Specifically, PWL parents had significantly greater weight losses during treatment than PAC parents. Note that parent weight losses in PAC are similar to those achieved when parents were involved as helpers in prior adolescent BWL trials (where parent weight was not directly targeted).^11,13 Parents in both groups reported increases in self-monitoring, improved dietary choices and increased PA behaviors, consistent with the treatments targeting role modeling (PAC) or parent weight loss (PWL); as anticipated, the magnitude of these behavior changes, overall and for self-monitoring, was significantly greater in PWL. Further, only adolescents in PWL perceived increases in their parents’ “dieting” behaviors, suggesting that the increased use of weight control strategies observed in PAC likely reflects parent role modeling of healthy behavior changes. As anticipated, adolescents in PAC perceived increases in authoritative parenting (particularly parental responsiveness) not observed in PWL, further supporting the successful manipulation of the parent treatments. Collectively, these results suggest that PAC and PWL are distinct in implementation as evidenced by differential outcomes on key variables.

Outcomes related to parent feeding behaviors were less clear. PAC used an authoritative approach to feeding, tailored for adolescents. This included avoiding over-restriction of unhealthy foods, setting limits while allowing adolescents to make choices within these limits, and fostering increased adolescent responsibility over food choices. There was no content related to parent feeding in PWL. However, parents in both groups reported reductions in restrictive feeding practices. PWL parents were taught to avoid over-restriction with respect to their own dietary intake, as well as the importance of psychological flexibility and mindful eating and enjoyment of all foods within moderation. Thus it is possible that this generalized to their adolescents. It is also possible that the decreased availability of unhealthy foods in the home, as evidenced by the home food inventory scores, led to less use of restriction in PWL (as there was less unhealthy food to restrict). Future research should investigate these potential explanations. Of note, the relation between parent feeding practices and adolescent obesity is not well-understood, particularly within longitudinal studies or diverse adolescent samples. Within a cross-sectional sample of youth (with a range of BMIs), parental monitoring and restriction were positively related to adolescent BMI.¹⁷ Similarly, when both parents and adolescents had overweight or obesity, parents were most likely to use restriction.⁵⁰ Greater use of restrictive feeding practices was also associated with higher adolescent BMI z-scores in Black families with adolescents ages 12–19 years.⁵¹ Yet, adolescent age appears to be an important consideration, with increased oversight of food choices (monitoring) associated with higher BMI z-scores among younger, but not older adolescents. Additional research to clarify the optimal parent feeding strategies for adolescents with obesity that considers the heterogeneous and fluid nature of this developmental period is needed.

Both parent approaches yielded significant adolescent weight losses at the end of the 4-month treatment. Obesity treatments targeting similarly diverse adolescent populations have reported varied effects (e.g., ΔBMI: +0.5;⁵² −0.1;⁵³ −1.3¹² kg/m²), which are typically lower than those achieved in mostly White adolescent populations (e.g., −1.5kg/m²).⁴⁷ Thus, both PAC and PWL are viable approaches to treatment. However, differences began to emerge during maintenance, suggesting the potential for PAC to confer better weight loss maintenance. Although effect sizes were small 3-months after treatment, there is potential that these differences, if maintained, would become increasingly robust over time. Findings suggest the potential benefit of a parent skills training approach designed to increase parents’ use of authoritative parenting, which could create an emotional climate that fosters sustained weight losses in adolescents. This is consistent with previous research suggesting that authoritative parenting might play a protective role related to adolescent overweight.⁵⁴ Findings contrast somewhat with the lower adolescent weight losses (ΔBMI=−0.43kg/m²) observed in a pilot behavioral weight control treatment targeting parental modeling and adolescent-parent communication, compared with a treatment without this component (ΔBMI=−1.28kg/m²).¹⁰ The authors suggested the potential for unintended family conflict related to targeting weight-related communication directly. Future research is needed to identify mechanisms of action in PAC and guide parent intervention targets within adolescent obesity treatment.

Limitations include the small sample size, only powered to detect between group differences on parent weight. Study strengths include its focus on adolescents, the role of parents, and inclusion of both adolescents and parents in treatment; this design recognizes the need for parental involvement and capitalizes on adolescents’ desire to be included in decision-making regarding their weight management treatment.⁵⁵ The sample’s sociodemographic diversity is also a strength, with over 50% of the sample from minoritized backgrounds for whom disproportionately higher rates of obesity are observed. Adolescent obesity treatment is understudied,¹⁴ and most research investigating the role of parents has focused on younger, predominantly White, children.^8,56 The study also had high retention, comparable to that achieved in primarily White adolescent samples from higher socioeconomic backgrounds.⁴⁷ However, consistent with prior investigations,⁴⁸ adolescents with higher baseline BMIs were more likely to discontinue treatment (although there were no group differences in BMI at baseline); thus additional retention efforts and/or alternative treatment models for adolescents with more severe obesity might be warranted.

To our knowledge, this is the first study to directly compare parent skills training with alternative, empirically-supported parent approaches and examine their potential impact on adolescent outcomes. Results suggest that PAC and PWL are feasible and distinct in implementation, evidenced by differential outcomes on key constructs. Further, both treatments yielded significant adolescent weight loss, thus represent viable approaches to adolescent obesity treatment. Yet, differential patterns were observed during maintenance, favoring the parent skills training approach. These findings warrant further exploration with a larger sample and longer follow-up and lay the foundation to explore potential mechanisms for their impact on adolescent weight loss and maintenance in a fully-powered trial. Future research could also examine the potential benefit of a combined approach (parent skills training plus parent weight loss). Moreover, given heterogeneity of this developmental period, a precision medicine approach might be indicated, in which the optimal parent approach is based on adolescent phenotypes (e.g., age, sex, autonomy, preferences, executive function, parent/child relationship). Results of this line of research have the potential to advance science and clinical practice regarding the optimal parental approach within adolescent obesity treatment.