Implementing family-based behavioral treatment in the pediatric primary care setting: Design of the PLAN study

Abstract

Family-based behavioral treatment (FBT) is an evidence-based treatment for pediatric obesity. FBT has primarily been implemented in specialty clinics, with highly trained interventionists. The goal of this study is to assess effectiveness of FBT implemented in pediatric primary care settings using newly trained interventionists who might implement FBT in pediatric practices. The goal is to randomize 528 families with a child with overweight/ obesity (≥85th BMI percentile) and parent with overweight/obesity (BMI ≥ 25) across four sites (Buffalo and Rochester, New York; Columbus, Ohio; St. Louis, Missouri) to FBT or usual care and obtain assessments at 6- month intervals over 24 months of treatment. FBT is implemented using a mastery model, which provides quantity of treatment tailored to family progress and following the United States Preventive Services Task Force recommendations for effective dose and duration of treatment. The primary outcome of the trial is change in relative weight for children, and secondarily, for parents and siblings who are overweight/obese. Between group differences in the tendency to prefer small immediate rewards over larger, delayed rewards (delay discounting) and how this is related to treatment outcome is also evaluated. Challenges in translation of group-based interventions to individualized treatments in primary care settings, and in study implementation that arose due to the COVID-19 pandemic are discussed. It is hypothesized that the FBT intervention will be associated with better changes in relative weight for children, parents, and siblings than usual care. The results of this study can inform future dissemination and implementation of FBT into primary care settings.

1. Introduction

Childhood obesity is a prevalent disorder that exhibits racial, ethnic and geographic disparities [1,2]. It is associated with increased blood pressure, cholesterol and triglycerides, and insulin resistance that persist into adulthood [1-3]. Childhood obesity is also associated with psychological changes that negatively impact quality of life [4,5]. Beginning in infancy [6], and extending through adolescence [7], gene-environment interactions contribute to the increased risk of a child developing obesity. Parents with obesity may also suffer from cardio-metabolic disease, initiating a cycle of obesity and cardio-metabolic disease that impacts both generations. Given the relationship between child and parental obesity [6-9], and evidence that parents arrange family eating and exercise environments and model behaviors [9-12], it seems logical that targeting parent and child can positively impact child weight control.

Family-based treatment (FBT) is an evidence-based treatment for childhood obesity [13-15] that targets children and their parents. FBT has shown significant reductions in relative weight at 10 year follow-up in specialized (non-primary care) settings [16,17]. There are strong relationships between child and parent outcomes in FBT [8,9], which persist throughout 5 year follow-up [18]. FBT’s simultaneous treatment of child and adult obesity is more efficacious and cost-effective than treating the child and parent separately by pediatrician and primary care doctors [19]. As FBT catalyzes changes in the home, weight change can generalize to siblings [20].

FBT has primarily been studied in specialty clinics, while the majority of children in primary care fail to receive evidence-based care [21-24]. Primary care offers an ideal setting for FBT [25] by capitalizing on the relationship between primary care physicians (PCPs) and families [26], and reducing fragmented care that can occur through multiple providers and offices. FBT may have cost savings [19] and health benefits [27] that extend beyond weight regulation.

The Primary care pediatrics, Learning, Activity, and Nutrition (PLAN) with Families study is a multi-site, individually randomized group treatment (IRGT) trial [28]. It is designed to compare the effectiveness of FBT plus usual care (FBT + UC) compared to a usual care (UC) over 24 months in 528 families with a 6 to-12-year-old child and parent or legal guardian, both with overweight/obesity, and their 2 to- 18- year-old sibling(s) with overweight/obesity. This publication describes the study rationale and protocol, including design, aims, intervention, and measures.

2. Methods

2.1. Study design

Families are randomized 1:1 to one of two groups; FBT + UC or UC and are followed for a 2-year period. The FBT + UC intervention has an inherently nested structure that is implemented within each site by newly trained interventionists called coaches, who are responsible for multiple families. Because this basic feature of the design implies clustering in one arm of the trial but not in the other, the study is an Individually Randomized Group Treatment (IRGT) trial [28], also known as a partially clustered trial. The trial is funded by the National Heart, Lung, and Blood Institute (U01HL131552) and is registered at clinicaltrials.gov (NCT02873715). The study is approved by the University at Buffalo (UB) Institutional Review Board (IRB). A single IRB is used with the participating sites at the University of Rochester, Nationwide Children’s Hospital in Columbus, Ohio, and Washington University in St. Louis, Missouri establishing reliance agreements with UB’s IRB. Each of the four sites include a variety of primary pediatric care practices from which families are recruited. Data Coordinating Center (U01 HL131639) is located at Washington University in St. Louis.

The goal is to recruit equal numbers of families across the four sites with multiple primary care pediatric practices at each site. All adult participants are provided written informed consent and parental permission for their child(ren), and all children age 7 years and older provide assent for participation. Assessments of height and weight occur every 6 months during the trial (0, 6, 12, 18 and 24 months).

2.2. Specific aims

The primary outcome measure is change in percent over BMI for children during the 24-month study. The primary hypothesis is that children randomized to FBT + UC will show greater reduction in percent over 50th percentile BMI change in comparison to children randomized to UC. Secondary aims include evaluation of between group change in parent weight and sibling percent over BMI; between group differences for participating parent and child in delay discounting (DD) [29,30] and how changes in DD are related to changes in percent over BMI, and if parental inconsistency [31] and environmental enrichment [32,33] are predictive of child percent over BMI outcomes.

2.3. Participant recruitment

The goal is to recruit 528 families with at least one child who is at or above the 85th BMI percentile for their age and sex [34] and has at least one parent or legal guardian who has a BMI at or above 25 [35] willing to participate. Recruitment has been completed, and 452 families (86%) are randomized and are participating. Only one child per family is targeted for treatment to maintain independence between participants in each group. If there is more than one child with overweight/obesity in the family aged 6–12 years, we target the oldest of these children as the primary participant, similar to previous studies [20]. By targeting only one weight-eligible child per family for participation, we can determine whether FBT has the cost-effective advantage of benefiting non-targeted 2–18 year-old siblings with overweight/obesity. In previous clinical trials, 43.2% of families with one child with obesity had at least one sibling with obesity between 2 and 18 years-of-age. Based on this number, we planned to recruit 228 siblings with overweight/obesity, and 185 (81%) are participating. Families can enroll up to 4 siblings. Recruitment strategies are designed to include at least 30% minority families to reflect the ethnic group distribution at participating sites to facilitate the generalization of the results to the broader population. The distribution of child racial/ethnic minorities across groups by sex is shown in Table 1. See Appendix Table 1 for characteristics of the practices within each site.

Table 1.

Racial/ethnic minority status of target children and parents by sex

	Not Hispanic or Latino				Hispanic or Latino				Unknown/Not Reported
	Child		Parent		Child		Parent		Child		Parent
	Female	Male	Female	Male	Female	Male	Female	Male	Female	Male	Female	Male
American Indian	1	0	1	0	0	0	0	0	0	0	0	0
Asian	3	5	7	0	0	0	0	0	0	0	0	0
Native Hawaiian	0	0	0	0	0	0	0	0	0	0	0	0
Black	77	44	116	6	1	1	0	0	0	0	0	0
White	122	118	227	52	8	10	11	1	0	0	0	0
Multi-race	16	14	5	0	5	4	2	0	0	0	0	0
Unknown	1	0	1	0	4	7	10	2	4	7	8	3
Total	220	181	357	58	18	22	23	3	4	7	8	3

Families are recruited from pediatric practices in several ways. Each practice has children who meet the age and BMI percentile criteria flagged in their charts, and flagged families who are interested in the study are screened for the rest of the eligibility criteria, as listed in Table 2. We recruited from Practice-Based Research Networks (PBRN) at each site that encourage research in clinical practices. First, at pediatric visits, forms designed to acquaint families with the study are introduced by the primary care provider, and the families sign up for the study during their visit. In addition, flyers and postcards direct families to a website that sends their contact information to the recruitment specialist. Their contact information is entered into a database, and they are contacted by for screening. Second, pediatric practices send letters to families with children that meet eligibility criteria (i.e., age and BMI percentile), inviting them to participate by calling the study team or visiting the study website to complete eligibility questionnaires and schedule a phone screening. Third, office staff, nurses, and pediatricians who work with families during well-child visits prompt families about the study, and if interested, “warm handoffs” are made to the co-located coaches in the clinic setting. Families in both groups meet with an assigned coach in person or over the phone to learn more about the study before deciding to be screened for eligibility.

Table 2.

Inclusion and Exclusion Criteria

	Parent	Child
Inclusion	BMI at or above 25	BMI at or above the 85th percentile for age and sex
	Able to read and comprehend materials written in English at an eighth-grade level	Able to read and comprehend materials written in English at a third-grade level
	Parent be target child’s biological or adoptive parent or legal guardian	Child between the ages of 6 and 12
	Parent agrees to attend all treatment meetings	--
	Child resides with target parent at least 50% of the time	--
Exclusion	Concussion within 3 months of enrollment	Concussion within 3 months of enrollment
	Parent had or planning weight-related surgery within 2 years of enrollment	Child had or planning weight-related surgery at any time
	Medications affecting growth (e.g. insulin, thyroid hormone) or weight-affecting medications started/changed within 6 months of enrollment	Medications affecting growth (e.g. insulin, thyroid hormone) or weight-affecting medications started/changed within 6 months of enrollment
	Medical condition altering nutritional status or intestinal absorption (e.g. insulin-dependent diabetes), affecting growth (i.e. genetic or metabolic disease), chronic medical conditions including Type 1 diabetes, heart disease or heart failure, HIV or AIDS, muscular dystrophy, renal diseases, hypothyroidism	Medical condition altering nutritional status or intestinal absorption (e.g. insulin-dependent diabetes), affecting growth (i.e. genetic or metabolic disease), chronic medical conditions including Type 1 diabetes, heart disease or heart failure, HIV or AIDS, muscular dystrophy, renal diseases, hypothyroidism
	Severe restriction of diet due to allergies or religious beliefs that would inhibit family from reasonably following the Traffic Light Eating Plan	Severe restriction of diet due to allergies or religious beliefs that would inhibit family from reasonably following the Traffic Light Eating Plan
	Unmanaged/active psychiatric conditions meeting full DSM-5 criteria or impairing symptoms, including significant mood disorders, eating disorders or substance abuse disorders	Unmanaged/active psychiatric conditions meeting full DSM-5 criteria or impairing symptoms, including significant developmental delays, intellectual disabilities, or Autism Spectrum Disorder
	Disability that inhibits ability to walk or do minimal physical activity	Disability that inhibits ability to walk or do minimal physical activity
	Parent pregnant or planning on becoming pregnant in next 2 years	--
	Planning to move away from the area within 2 years	--

Since the pediatrician typically does not have information on parental health, we collect the parent’s medical and psychiatric history during the screening process, to ensure that parents meet study eligibility. Screening is conducted in person or over the phone with either a coach or recruitment specialist. Table 2 lists the screening measures and assessments collected during the study.

2.4. Randomization

Each family that meets eligibility criteria is scheduled for orientation, where they are provided an overview of the study, and their heights and weights are measured to confirm eligibility. After the orientation, interested families review and sign consent and assent forms, and baseline assessments are administered. Families complete delay discounting tasks, and parental consistency and environmental enrichment measures, after which they are randomized. Families in both groups are contacted by their assigned coach who informs them of the treatment group to which they were randomized.

The PCPs are kept blinded to family group assignment. They continue their usual care approach and are instructed to avoid talking about the program with families to limit their bias during regular UC pediatric visits. The assigned coach follows up with randomized families either to schedule an FBT + UC family’s first FBT session or to instruct a UC family to continue to see their PCP as usual and plan the 6-month follow-up visit. Collection of the primary weight and height data is blinded.

To maintain blinding of the PCP, the PCP is not informed about the referred families’ eligibility status, participation, or randomization group. This information is not in the patient’s chart, and all PLAN program scheduling occurs outside of the PCP clinic scheduling. Families do not check in with clinic staff when they arrive, and when possible, families from either group come to the practice for follow-up data collection by blinded assessors. In addition, families are instructed not to discuss program issues with the PCP. At the end of the study, notification of patient progress will be sent to the PCP, which can be entered into clinic records. We recognize that the PCP may try and deduce group assignment if a family loses significant weight and looks different, or if a family inadvertently shares that information. However, there may be families in the usual care group who also do well, so PCP guesses about group membership may not always be accurate.

3. Study intervention

3.1. Intervention – UC

Usual Care is the comparison group reflecting the standard of care treatment recommendations by the Expert Committee Recommendations for Assessment and Treatment of Obesity [36] and the American Academy of Pediatrics [36]. PCPs are encouraged to schedule follow-up visits with families according to their standard practice of following of Stage 1 Prevention Plus guidelines [36]. While PCPs have little exposure to FBT, they receive information regarding methods of implementation and efficacy of FBT based on previous research at the beginning of the study.

3.2. Intervention – FBT + UC

FBT is a rigorously tested, multicomponent intervention that targets diet, activity, behavioral skills, parenting, and social facilitation in children and their parents. The treatment includes: 1) the Traffic Light Eating Plan, which utilizes RED, YELLOW, and GREEN labels for food to guide families toward the goal of consuming healthy low energy dense, high nutrient dense foods; 2) the Traffic Light Activity Program, also utilizes RED, YELLOW, and GREEN labels for different levels of caloric expenditure to increase physical activity and reduce sedentary behaviors; 3) a variety of behavioral techniques, including stimulus control, self-monitoring, goal setting, problem solving, finding behavioral substitutes for highly reinforcing food, improving positive parenting, providing reinforcers for behavior change and weight loss; and 4) facilitation of support in the family and peer environments to optimize the durability and generalizability of improved eating and activity habits across multiple social and environmental contexts (e.g., home, school, with friends) [37,38].

During each treatment meeting, parents and children are weighed and then have an individual family meeting with their coach. This includes: 1) education on diet, physical activity and behavior change techniques; and 2) review of diet and activity self-monitoring and 3) coaching on overcoming any barriers to adherence with the weight-loss behaviors. FBT is mastery-based to adapt the dose of treatment to individual participant needs, consistent with previous research using mastery criteria for FBT [39]. While adaptable, treatment dose targets are within the United States Preventive Services Task Force (USPSTF) recommendation of 26 or more hours of individualized, intensive intervention [40,41]. The minimal projected treatment includes 8 weekly visits, 8 biweekly visits, and 6 monthly visits for the first year, and 4 quarterly visits for the second year of intervention. All FBT families start on a weekly schedule followed by spacing of the sessions to biweekly, monthly, and quarterly schedules based on their uptake of positive behavior changes and weight loss. This mastery-based schedule allows for more attention devoted to families who need more time to absorb the content of the program and master behavioral changes.

3.2.1. Mastery system

Use of a mastery system, shown to improve outcomes of weight control intervention [39], recognizes variability in the rate-of-change for eating, physical activity, and parenting behaviors among participants. Thus, the system introduces behavior change information and goals at an individualized pace, with the skills necessary for weight control taught sequentially, and with basic skills acquired before more complex or intensive changes are included. Mastery refers to knowledge about eating, physical activity and behavior change skills and the demonstration of behavior change in these areas. Mastery of knowledge is determined by scores on quizzes based on recommended module readings, while mastery of eating, physical activity and behavioral skills are determined by meeting specified behavioral goals for at least two weeks consecutively. Behavioral skills are systematically built on previous skills that are learned, or scaffolded, to move progressively toward meeting behavioral goals. Knowledge and behavioral goals previously mastered are periodically reviewed during later weeks to ensure maintenance of these skills over time.

A study website provides information to families about FBT in the form of downloadable manuals for the Traffic Light Eating Plan and Activity Program, tools for cooking, ideas for getting more physical activity, and tips for improving positive parenting skills. Quizzes to assess mastery of educational materials are implemented on the study website, with multiple versions of quizzes on each unit available with the recognition that some people will acquire the information more slowly than others. Families are provided traditional paper and pencil self-monitoring tools through the form of a habit book to self-monitor eating and activity behaviors, and after the skill of self-monitoring is acquired, families can choose to continue their habit books or switch to technology-based recording. Progress information on chapter quizzes and eating, physical activity, and parenting goals are coordinated in an online family dashboard. Families have access to the family dashboard for feedback, and coaches have access to the dashboard to assess patient progress, assist with problem solving, and to communicate with families to structure solutions. The website also contains password protected sections that are for internal use by study personnel. This section, which is only available to study staff who have a password, is a repository for study documents and a communications hub for the study, such as regular reports prepared for Data Safety Monitoring Board (DSMB) and steering committee meetings, the forms and datasets manual which facilitates the matching of variables in a dataset with items on a form, lists of certified personnel, and instructions on the use of the data entry system. The website does not contain protected health information or participant data used for analysis of outcome measures.

3.3. Dietary and activity goals

Participants in FBT are instructed to follow the Traffic Light Eating Plan, which categorizes foods into the colors of the traffic light: RED, YELLOW and GREEN [42,43]. Categorization is based upon a food item’s energy, fat, and added sugar content in relationship to foods that provide the same nutritional value to maximize nutrient density. Examples of the Traffic Light Eating Plan are presented in Appendix Table 2. In general, only non-starchy vegetables are considered GREEN foods. YELLOW and RED foods are determined by caloric cut-offs per serving for each food group (vegetables; starchy vegetables; fruit; grains; dairy; protein; soups; and condiments, dressings, and other ingredients). All foods categorized as fats, oils, added sugars, and others such as high sugar foods including fruit juices are considered RED foods. Furthermore, modified foods categorized as fats, oils, added sugars, and others (i.e., reduced-fat cookies), are also considered RED foods, as these foods generally have low nutrient density, do not shape healthy eating because they mimic the taste of RED foods, and can decrease consumption of healthier foods. Combination foods comprised of several individual ingredients (i.e., pizza, hamburger, lasagna), are considered RED foods if one of the individual foods in the combination food is a RED food and contributes one or more servings of RED food. Participants are instructed to consume between 1200 and 1500 kcal per day, with an ultimate goal of ≤2 servings of RED foods per day and ≥ 5 servings of GREEN foods per day, which is shaped over time. Participants are instructed to eat the portion size and number of servings from the differing food groups as recommended in the 2015–2020 Dietary Guidelines for Americans [44]. Families develop weekly meal plans to assist them with achieving these dietary goals, shaping them toward the ultimate goals. Energy goals are adjusted as needed for each individual participant for weight loss and maintenance.

The Traffic Light Activity Program provides information on metabolic equivalent values or multiples of resting metabolic rate (METS) for a wide range of physical activities. The activities are given a color code based on whether they are sedentary (less than 3 METS, RED), moderate (3–4.9 METS, YELLOW) or hard (5 to 6.9 METS, GREEN), or very hard or fitness boosters (≥ 7 METS, GREEN). The activity goal is to increase the amount of moderate- to vigorous-intensity physical activity (MVPA) by families to meet recommended guidelines (≥ 60 min/day [children] or 150 min/week [adults], structured as ≥30 min/day at least 5 times/ week) to assist with forming daily habits [45]. Similar to the diet, physical activity is gradually shaped (setting smaller goals that lead up to the ultimate goal) to meet the targeted goals. Participants are encouraged to complete ≥10 min/day 5 times/week at a moderate-intensity activity level starting at week 2, and then increase physical activity 10 min/day (child) and 5 min/day (adult) each week until they have reached the intervention goal. Children are encouraged to engage in developmentally appropriate play activities and family-based physical activities (e.g., go for a family walk or a bike ride) and parents are also encouraged to do brisk walking and other aerobic activities.

3.4. Behavioral strategies

A variety of behavioral skills are used in FBT [46], including self-monitoring, positive reinforcement, stimulus control, modeling, pre-planning and problem solving, goal setting, social facilitation, and relapse prevention. Families self-monitor their daily intake throughout the study until mastery is met, including recording the type and amount of all foods and beverages, the calories, and the number of GREEN and RED foods. While paper habit books are encouraged for use at the beginning of treatment, families can choose to self-monitor in online applications such as MyFitnessPal throughout treatment or until a healthy lifestyle is maintained. Physical activity is monitored in terms of RED, YELLOW and GREEN categories, and minutes of physical activity. Families are also trained to weigh themselves at home throughout the treatment, up to once a day for parents and a twice a week for children.

Parents are trained to use praise and positive reinforcement to shape and maintain their child’s healthy behaviors. Either on paper or on the PLAN website, children earn points for completing dietary goals, activity goals, and achieving weight loss that are redeemed for parent-delivered reinforcers. These reinforcers are chosen from a reinforcement menu that lists types of activities or privileges that the child can earn (e.g. spending time with friends, sleepovers, family or friend based physical activities, play games with parents). Points are often used as conditioned reinforcers to develop or maintain positive behavior change. Food and material gifts are not encouraged as reinforcers.

Modifying the environment to reduce cues for unhealthy eating and sedentary behaviors, while increasing cues for GREEN foods and physical activity are encouraged. Participating adults are encouraged to make changes in the same behaviors as their children, and to model these healthy behaviors and healthy attitudes about behavior change [47].

Families are encouraged to pre-plan and use problem solving to make it easier to engage in healthy behaviors, particularly around holidays, vacations, parties/functions or eating out. During each treatment session families make plans to address potential challenging situations. Families are encouraged to use goal setting for both short and long-term goals. Families are taught the importance of short- and long-term goal setting for successful behavior change and shaping is used for changing moderate to vigorous physical activity since rapid changes in this behavior increases risk of injury. Families are taught social facilitation and are encouraged to create and maintain social connections. These connections are encouraged to be ones that foster healthy behaviors and support community efforts that create environments for healthy eating and activity. Families are encouraged to have regular brief family meetings to review and reinforce daily behavior change, problem solve barriers to behavior change and use pre-planning for future events.

3.5. Adaptations of treatment from specialty clinics to pediatric primary care clinics

FBT is an evidence-based treatment that has been implemented in a mixed (group + individual) format [48], almost exclusively in specialty clinics, with highly trained personnel. The translation of this treatment to pediatric primary care centers required modification of the existing treatment format from a mixture of group and individually based sessions to solely individual-based sessions. Pediatric primary care centers seldom see children in groups. The same FBT material that was used in the mixed treatment format was used for individual treatment as well, but there are several differences in implementation. In traditional FBT, both individual and group aspects of treatment are available. Group sessions are delivered to parents and children separately to allow learning of age-appropriate information about behavior change, diet and exercise. Parent and child dyads are also seen in individual family case management sessions to generalize behavior change principles for that family. In the individual treatment format used in this study, the coach spends time on the psychoeducational components of behavior change, diet and exercise, as well as on case management for the family. Therefore, families are not able to take advantage of the group format, which includes parents and children learning from each other, rather than directly from the coach. This exchange of information among parents or children during group sessions can strengthen information provided by the group leader and lead to new insights. However, there are also major advantages of the individual approach, which include the ability to be very flexible in scheduling, as well as a focus on an individual family rather than a group of families, and less space required for individual rather than group treatment in an already tight clinical setting.

3.6. Coach training

Given that the study is designed to test whether FBT could be implemented in primary care settings, the decision was made to implement FBT with coaches who were not familiar with FBT prior to participating in this study. This approach was to parallel the idea that providers would experience a learning curve with FBT content such as the behavior change methods used, or the Traffic Light Eating Plan and Activity Program. Importantly, it ensures capability of translation of the program as coaches can have a variety of backgrounds and would be readily available for hire by practices if this approach were generalized. All PLAN coaches possess at least a bachelor’s degree and may be pursuing an advanced degree in a related field of study such as dietetics, social work, or clinical psychology. Coaches complete IRB required trainings (i.e. mandated reporter training, HIPPA, Good Clinical Practice training) as deemed necessary per site requirements. The study specific training included a 30-h initial training certification process with a combination of education, interactive role play, and simulations focused on introducing FBT and childhood obesity, parenting skills, nutrition, and social facilitation maintenance. See Appendix Table 3 for training details. To conduct FBT sessions, coaches had to demonstrate mastery of relevant material by scoring at or above 80% on all required quizzes. PLAN coaches then conducted FBT with two pilot families prior to treating their first study family. These sessions were audio recorded for parents who agree to audio recording and reviewed for competency by the Training and Fidelity Core (TFC) at Washington University in St. Louis, MO. Following four consecutive competency ratings, the PLAN coaches are assigned study families. FBT session recordings are uploaded to a secure website for review by the TFC to maintain treatment fidelity. If a session is rated by a supervisor within the TFC as less than competent, the coach undergoes increased supervision and training until a session is rated competent. This process continues throughout treatment.

3.7. Treatment Fidelity

Recordings of each FBT session are reviewed by a FBT supervisor and discussed with the coach. Supervision is by study staff experienced in FBT. The TFC provides oversight of supervisors to maintain consistency across sites. Treatment fidelity is evaluated by weekly random audits of audio recordings using a fidelity rating checklist (Appendix Table 4) that is shared with the coach during supervision to discuss areas in which they excelled and areas that need improvement. Additional remediation may be provided if needed. Supervisors meet as a group weekly to discuss areas of concern; they also perform quarterly fidelity rating calibrations to ensure that consistency is maintained across sites and supervisors. At completion of the study, audits will be conducted by trained TFC raters who were not involved in training the coaches.

3.8. Assessments (Table 3)

Table 3.

Screening and Assessment Measures Timeline.

				Months
Domain	Measure	Respondent	Measure Target	Screening	0	6	12	18	24
Study Overview and General Eligibility	Initial Eligibility	Parent	Parent/Child	X
	Eligibility Screen	Parent	Parent/Child	X
Medical and Psychiatric Eligibility	Patient Health Questionnaire	Parent	Parent	X
	Questionnaire on Eating and Weight Patterns	Parent	Child	X
	Pediatric Symptoms Checklist	Parent	Child	X
	Kiddie Schedule for Affective Disorders and Schizophrenia	Parent	Child	*X
	Structure Clinical Interview for DSM Disorders – Research Version	Parent	Parent	*X
Demographics and Motivation	Clinical Interview	Parent/Child	Parent/Child	X
	Additional Demographics	Parent	Parent/Child	X
Baseline Predictors	Parental Consistency	Child	Parent	X
	Environmental Enrichment	Parent	Parent/Child	X
Delay of Gratification	Delay Discounting Task	Parent/Child	Parent/Child	X		X		X
Relative Weight	Height and Weight	Parent/Child	Parent/Child	X	X	X	X	X
Participant Safety	Adverse Event Questionnaire	Parent	Parent/Child		X	X	X	X
Participant Attitudes	Client Satisfaction Questionnaire	Parent/Child	Parent/Child					X
rimary Care Provider Measures	Provider Demographics	Provider	Provider	X
	Obesity Treatment Knowledge	Provider	Provider	X
	Evidence-Based Practice Attitude Scale	Provider	Provider	X
	Attributes of FBT	Provider	Provider			X
	Intended Adoption	Provider	Provider					X

Note –

^*Structured interviews are administered when indicated by scores on screening instruments.

3.8.1. Screening

A variety of variables are assessed during screening either to gather demographic information or to determine family eligibility for the study. Race and ethnicity, socioeconomic status, and current physical and mental health problems are measured during screening. Family income, parent educational level, and racial/ethnic background are obtained using a standardized questionnaire. This questionnaire is adapted for this study using MacArthur’s Research Network Socioeconomic Status and Health [49] and the Barratt Simplified Measure of Social Status [50]. Medical history of the parent and child is obtained by parent interview. Disordered eating and other psychopathology are assessed at screening using a two-step approach. First, parents complete the Questionnaire on Eating and Weight Patterns – Parent (QEWP-5-P) [51] and Patient Health Questionnaire (PHQ) [52] and the Pediatric Symptom Checklist (PSC-35) [53] for their child. Second, individuals with elevated scores are followed up with a structured interview. The Structured Clinical Interview for DSM-5 Research Version (SCID-5-RV) [54] and The Kiddie Schedule for Affective Disorders and Schizophrenia (KSADS) [55] is used for adults and children respectively during screening to assess the participants’ psychiatric history. If appropriate, referrals are made for children and parents at one of the local Office of Mental Health-licensed community mental health centers. Families with a current diagnosis of a co-morbid psychiatric disorder such as anxiety or depression are eligible to participate if the treatment they are receiving is not related to eating or weight. Parents and children with an eating disorder (i.e., anorexia nervosa, bulimia nervosa, binge eating disorder) are excluded from participation.

3.8.2. Baseline measures

Parental Inconsistency is assessed using child-report on 3 items from the validated parenting style inventory (ESI) [31] that ask how often the parent follows through with earned rewards/punishments and a sum score calculated with higher values representing larger inconsistency. Program staff were available to provide help for the younger children. When necessary, program staff read program materials to younger children, including consent forms and measures. In general, if a child can read and understand the consent form, they complete the parental consistency measures without help, though it is available. If a child needs the consent form to be read to them, the parental consistency measure is read to them. In no instance did parents assist in completion of this measure. Parents worked on a separate task while the children complete this questionnaire. Environmental Enrichment is assessed at baseline using items from the HOME scale [32,33], a scale to assess physical activity and sedentary equipment in the home [56], and the Land Use Mix-Diversity subscale from the Neighborhood Environment Walkability Scale (NEWS) [57]. These items describe the availability of fruits and vegetables, electronics, physical activity equipment, cognitively stimulating items in the home, and the proximity of local community facilities to the home and have demonstrated acceptable test-retest reliability, as well as construct and predictive validity [32,33]. Items from these scales are combined to create a global index of environmental enrichment that has been shown to predict weight loss success during FBT [58]. Demographic predictors such as age, gender, race/ ethnicity, and SES are also assessed using the standardized questionnaire administered during screening adapted for this study [49].

3.8.3. Treatment outcome measures

Assessments of targeted child and parent height and weight are conducted at baseline, 6, 12, 18 and 24 months. Primary outcome is change in percent over BMI normalized to child age, sex and height. Secondary outcome variables measured include parental weight/BMI change, sibling change in relative weight, changes in parent and child DD. We calculate change in percent over BMI (for the targeted child and sibling(s)), z-BMI, weight, and BMI at each assessment time point from participants’ heights and weights. z-BMI are standardized scores based on CDC growth charts for BMI values, which are also used to create BMI percentiles for the diagnosis of overweight and obesity [34,59,60].

Weight is assessed throughout treatment in the FBT + UC group as part of the goal setting and reinforcement system. The plan was for weight to be measured using a state-of-the-art SR Instrument (Tonawanda, NY) Portable scale with wireless transmission of weight that collects and transmits height and weight directly to a data file. This was designed to reduce potential measurement bias and errors in observing and recording weight [61]. However, recurrent problems in data transmission were noted and since November of 2019, weight data is obtained by assessors blind to treatment condition and recorded manually into the database. Height is measured by trained blinded reliability within 0.3 cm.

Delay discounting (DD) is measured because the inability to delay gratification is a trans-disease process related to obesity [62,63]. Inability to delay gratification predicts childhood weight gain [64,65], poorer response to FBT, [58] and children with obesity are more likely to choose a small immediate versus a larger delayed food reward [66-68]. Likewise, adults who discount the future are more likely to be obese [69,70], purchase ready-to-eat foods than healthier foods [71], and have increased weight [69,70,72]. DD is assessed at baseline, 12 months, and 24 months. Participants complete the DD task at 2 monetary amounts ($50 and $100). At each amount, they are asked using computerized assessments whether they would choose a smaller amount of the hypothetical reward available immediately or the larger amount available later [29]. The magnitude of the immediate commodity is adjusted [30] until it is subjectively equivalent to the later larger amount. These indifference points are obtained at 6 delays (1 day, 1 week, 1 month, 3 months, 6 months, and 1 year). The derived discounting parameter, k, describes the decline in the present value of a reward with delay to its receipt, with higher k values indicating lower degrees of discounting the future. The DD task takes approximately 5 min to complete.

3.8.4. Participant level descriptors

For primarily descriptive purposes, participant adherence to FBT is monitored by 1) session attendance and 2) completion of a laboratory constructed questionnaire to obtain self-reported adherence to diet, physical activity and behavioral skills. Child and parent acceptability of FBT + UC or UC is assessed at the end of treatment using the Client Satisfaction Questionnaire [73], a general 8-item scale meant to assess client/patient satisfaction with services with known good psychometric properties. For the parent, an additional four items from a questionnaire developed by Ede et al. [74] are added to assess acceptability of receiving behavioral obesity treatment in the primary care setting. General participant demographics such as gender, race/ethnicity, etc. are also assessed.

3.9. Adaptations due to COVID-19

On March 16th, 2020, approximately 28 months into the study, it was necessary to modify treatment and assessment procedures to shift from in-person contact to remote contact as a result of the COVID-19 pandemic. This has been accomplished using phone and secured video conferencing for all treatment sessions during the pandemic. Training of coaches to implement treatment remotely and to conduct needs assessments on unique familial circumstances during COVID was accomplished within two weeks after in-person visits were interrupted.

After initiation of remote treatments, it was also necessary to initiate remote assessments. We developed a protocol for remote collection of outcome measures, which included electronic consent, assessment checklists for remote assessors to use, a tutorial video and family instruction sheet, obtaining HIPAA-compliant video conferencing accounts, and electronic participant payment. This took from approximately March 16th to June 1st to get final approval and initiate remote assessments. Digital home weight scales, carpenter’s squares, and metal tape measures are mailed to families to conduct weight and height measurements as similarly as possible to prior assessment procedures. The participating parent is led by a blinded assessor via phone or video conferencing to measure and report parent, child and siblings’ heights and weights at home. The blind assessor guides and supervises the measurements and reports heights and weights in the database. Studies on the validity of weight measurement of scales sent to families and reliability of the height measures were implemented. A sample of 12 Etekcity scales (Anaheim, CA) were validated in comparison to a standardized 50-pound calibration weight and calibrated digital scale (SR Instruments, Tonawanda, NY). Six of the scales measured 50 pounds, while the other six 50.2 pounds. Families are emailed instructions and a tutorial video prior to the assessment so they could review the procedure.

Given the reliance on parents or caregivers to provide measurements, the height equipment and protocol was verified by remeasuring 40 children’s heights 1 to 2 weeks after their first measurements. To accomplish this, ten families at each site were randomly chosen for remeasuring and were consented and compensated with $10 after participation. The intraclass correlation coefficient (ICC) between the two measurements and bootstrapped 95% confidence interval for the ICC is 0.995 (95% CI, 0.991, 0.999).

3.10. Statistical analysis plan and power analysis

Initial analyses involve Chi-Square and t-tests to compare baseline values of key variables across groups for outcomes. The analytic strategy comparing between group differences in child percent overweight reflects the fact that this is an IRGT trial involving clustering in only one group with 24-months as the primary endpoint. Thus, the primary analysis is analysis of covariance with the 24-month outcome as the dependent variable and baseline value of percent overweight as the covariate. Independent variables are the treatment assignment, site and coach nested within the site in the FBT group. Because this is an IRGT, different covariance structures are evaluated and used in the two treatment arms. Additional analyses adjust for study site, sex and race of the participating child, family income, educational level of the participating parent, whether there are one or two parents in the household. Sensitivity analyses include all time points (baseline, 6, 12, 18, and 24 months) and the use of a mixed model repeated measures analysis of variance (ANOVA) with the same nesting and covariance structure that are employed in the primary analysis. Since the primary time point is 24 months, these sensitivity analyses involve an assessment of the statistical contrast that compares the change from baseline to 24 months in the FBT + UC group with corresponding changes in the UC group. We use intention-to-treat analyses that include all randomized families. We also perform a sensitivity analysis not including the families that drop out before providing their six-month data.

In addition to evaluation of treatment effects on children, we are also testing the null hypothesis that there is no between-group difference in weight change among participating parents and siblings. The same analytic plan will be used to assess between group differences in relative weight for the participating parent and sibling and changes in DD for the participating parent and child. For analysis of sibling change, additional analyses will also include sibling age and whether the sibling is older or younger than the participating child. We may explore additional analyses adjusting for the possible impact of the Covid-19 pandemic, which required modification of both the administration of treatment as well as the collection of outcome data.

Further analyses will include bivariate linear mixed models to evaluate the association between changes in DD and changes in child and parent relative weight, as well as the relationship between baseline parental inconsistency and environmental enrichment and changes in relative weight of the child.

Exploratory analyses to best characterize change for the sensitivity analyses, and to determine whether significant variability existed in participants’ (both children and their parent) rates of change and initial values, are conducted using a fixed and random effects approach to growth modeling. All models are estimated under missing data theory using all available data and robust maximum likelihood estimation. This strategy is a modern method of modeling with missing data that makes use of all available data points [75]. Treatment differences by sex, race, and ethnicity are included as covariates.

Because a non-linear relationship is anticipated between time and all outcomes (e.g. change from 0 to 6 months, 6 to 12 months, and 12 to 18 months followed by either stability or a slight return toward baseline values), unconditional piecewise growth is modeled using the following random parameters (along with their respective variances): intercept (reflecting individual differences in baseline values), slope 1 (reflecting individually varying rates of change during 0 to 6 months), slope 2 (reflecting individually varying rates of change from 6 to 12 months), and slope 3 (reflecting individually varying rates of change from 12 to 18 months).

Although it is expected that randomization equalizes group characteristics on average, a secondary analysis will compare baseline differences by group for sex and race of the participating child, family income, the educational level of the participating parent, and number of parents in the household, and adjust for these covariates if they are different between groups and are related to child or parent outcomes.

The sample size for this study of children (and parents) per arm is based on a two-sided test at the 0.05 level of significance and power 0.90 and is generated using 1000 simulations.

A major advantage of recruiting patients from primary care practices is that changes in height and weight can be assessed from medical records as needed if a family discontinues study participation or is unwilling to complete a study measurement at the 24 month time point as we will request. Recognizing that personnel at the pediatric primary care setting may not take heights and weight using the same standardized protocol that we use [76], an ancillary study will establish the relationship between site recorded heights and weights and study recorded heights and weight.

3.11. Monitoring safety issues

Adverse events (AE) and serious adverse events (SAE) for children and parents are assessed using a structured adverse event surveillance form. AEs and SAEs are assessed by blinded assessors during structured assessment visits for both FBT + UC and UC groups. In addition, families in FBT + UC may report AEs and SAEs during treatment sessions to an unblinded coach. In these cases, the families’ immediate needs are addressed, and they are reminded to report the AE or SAE to the blinded assessor at their next assessment visit. All AEs and SAEs are reviewed by site PIs, and those that are coded as related or possibly related to the study are reported to the IRB in accordance with their requirements. As a reliability check, all AEs and SAEs are reviewed with attention to expectedness of the event and relatedness to the study by at least two members of the Safety Monitoring Committee. In addition, all AEs and SAEs are categorized into the type of adversity: injury, doctor visit, common pediatric event, pregnancy, hospitalization, or mental health. This categorization facilitates the adverse event reporting, reviewing, and coding process and is useful for safety and data analyses. It should be noted that the FBT + UC group has significantly more contact with their coach and the study team in comparison to the UC group due to the design of the intervention, and thus may have more AEs reported in the FBT + UC group than the UC group. A modification to have blind assessors, not the family’s coach, review the survey for completeness at the assessments to better equate reporting from both groups.

4. Discussion

While FBT has been shown to be an efficacious intervention when delivered in a mixed format (group + individual) in specialty clinics, the majority of children with obesity are not seen in specialty clinics but rather are seen in their pediatric primary care clinician’s office. Thus, it would be advantageous to evaluate the effectiveness of FBT as implemented using an individualized approach to treatment in pediatric primary care.

The current trial evaluates FBT from the perspective of a family-based problem that is shared by the targeted child, a parent with overweight/obesity, and in many families, a sibling with obesity. One of the strengths of FBT is the fact that the same intervention can affect excess weight in both the child and parent, as well as the weight of a sibling with obesity who is never seen, but who benefits from changes in parental behavior and the shared family environment.

The translation of FBT from specialty clinic to primary care represents unique challenges due to differences in treatment implementation. In specialty clinics FBT is primarily implemented in a mixed format, while in primary care treatments are primarily implemented individually. There is not a large database on individual versus group treatment in children, but research suggests a benefit for mixed group and individualized treatment [48]. In FBT, both the parent and child are treated, while in pediatric practice pediatricians do not usually actively intervene on parental behavior or health. This may make implementation of treatment challenging for pediatricians who are not used to or who feel uncomfortable intervening with parents and children simultaneously.

Previous implementation studies in children [13,77] and adults [78,79] have shown a reduction in treatment effectiveness when evidence-based treatment is implemented in effectiveness studies. PLAN was implemented in part during the COVID-19 pandemic. Thus, there were participants who were treated in person and those who had a hybrid of in-person plus remote treatment. The influence of these two treatment modalities on outcomes will be assessed. An important finding that there is no difference in outcomes between those receiving FBT treatment in person versus those with hybrid in person and remote treatment, would suggest that implementing FBT remotely in the future is likely to be feasible. However, the power to detect this difference will be limited.

The demonstration of successful between-group differences in child percent overweight change can justify more widespread implementation of FBT in primary care practices. Generalization of treatment effects to non-treated family members is a cost-effective strategy to promote family-wide intervention effects. Successful treatment outcome can lead to development of methods to train primary care pediatricians and professionals who work in these practices to implement FBT as a treatment for parents and children. A complementary approach is to increase the use of non-medical professionals and staff who are trained in methods of behavior change to implement obesity treatments in primary care. These professionals can also provide interventions for a wide variety of problems that pediatricians commonly deal with and parents are concerned about, including developmental concerns, parenting and adherence issues, smoking, eating disorders, medication adherence, and bullying. Given the important role of the pediatrician in child health care, and the reliance on the pediatrician’s opinion about the need for behavior change and weight control in the child and family, the ideal model may be to develop a close partnership between pediatricians and individuals trained to deliver dietary and behavioral teaching so that the family perceives the weight control program as integral to the pediatric practice. This study provides the opportunity to assess a variety of predictors of outcome, ranging from executive function and decision making (Delay Discounting), through parenting style [31], environmental enrichment [32,33,56], and neighborhood walkability [80].

The current study tests the effectiveness of an evidence-based treatment translated into primary care pediatric practice. This study builds on previous research implementing FBT in primary care pediatric practice with preschool children [25] to test effectiveness of FBT in pediatrician offices for preadolescent children, and these studies can provide a model for translation of evidence-based treatment to primary care. We acknowledge the challenges that can occur in modifying the format of an evidence-based treatment to fit into primary care settings, and the potential impact of COVID-19 on treatment effects. This study will assess a wide range of predictors of treatment outcome, as well as predictors of implementation in primary care settings.