Protocol for the Imagine HEALTH Study: Guided imagery lifestyle intervention to improve obesity-related behaviors and salivary cortisol patterns in predominantly Latino adolescents

Abstract

Innovative lifestyle interventions are needed to reduce type 2 diabetes risk in adolescents. This report describes the protocol of the Imagine HEALTH cluster randomized control trial, that tests an intervention based in Self-Determination Theory (SDT) and uses lifestyle education combined with the mind-body, complementary health modality of guided imagery (GI), to address obesity prevention and treatment in predominantly Latino adolescents. The primary aim is to determine the unique effects of each of the three major components of the 12-week lifestyle intervention (lifestyle education, stress reduction guided imagery, and lifestyle behavior guided imagery) compared to control on primary outcomes of physical activity (accelerometry), dietary intake (3-day recall), and stress biomarker levels (salivary cortisol). Secondary aims assess changes compared to controls in psychosocial outcomes (stress, well-being, depression), diabetes-related metabolic outcomes (adiposity, insulin resistance), maintenance of outcome changes for one year post-intervention, and SDT-based mediation of intervention effects. The development and rationale for each of the intervention components, study design, and outcome measurement processes are described. Adolescent participants recruited from four urban schools are cluster randomized by school into one of four arms of the 12-week (3-month) intervention, followed by 6 months of maintenance and 6 months of no contact. Outcome measures are assessed at the end of each period (3-, 9-, and 15-months). Results to date show successful recruitment of 97% of the target study population. Future results will demonstrate the effects of this integrative intervention on primary and secondary outcome measures in adolescents at risk for lifestyle-related metabolic disease.

1. Introduction

The 2013–2014 NHANES estimated that 20.5% of U.S youth, and 22.8% of Hispanic/Latino youth, between 12 to 19 years of age were obese [1]. Among overweight and obese Latino youth, there is a high prevalence (~32%) of pre-diabetes [2–4] and the metabolic syndrome (~30 %), both associated with insulin resistance[5]. Furthermore, persistent pre-diabetes and metabolic syndrome are both associated with increasing diabetes risk over time [6, 7]. As in adults, insulin resistance, rather than body fat per se, is the primary patho-physiological factor leading to metabolic disease risk[8, 9]. Therefore, interventions that target reductions of insulin resistance through lifestyle behavior changes are key in the prevention of type 2 diabetes (T2D) in overweight adolescents.

Previous interventions to prevent and treat childhood obesity have been limited in several areas, including the relative paucity of interventions in minority, high-risk populations[10, 11], lack of interventions that are based in health behavior theory[12], emphasis on cognitive-based interventions that do not take developmental issues adequately into account[13, 14], and failure to address the metabolic outcomes of obesity. While chronic stress has been linked both to the development of obesity, and the development of obesity-related morbidities, previous intervention strategies have not addressed this issue, nor have they utilized complementary mind-body modalities to reduce chronic stress as part of an integrative approach to obesity prevention and/or treatment.

In response to these considerations, we previously developed and tested the 12-week Imagine HEALTH (Healthy Eating Active Living Total Health) pilot lifestyle intervention for obese Latino adolescents[15]. This multicomponent, integrative intervention had two major components: 1) lifestyle education, and 2) individual Interactive Guided Imagery^SM designed to reduce stress and promote healthy lifestyle (physical activity and eating) behaviors. Results demonstrated that compared to controls, IGI^SM acutely reduced the stress biomarker salivary cortisol across a 45-minute span, reduced sedentary behavior, increased moderate physical activity, and marginally decreased caloric intake after the 12-week intervention[15]. These findings led to the randomized controlled trial (RCT) described in this report.

We herein describe the protocol of the Imagine HEALTH cluster randomized control trial that uses lifestyle education combined with the complementary-integrative health modality of guided imagery (GI) to address obesity prevention and treatment in predominantly Latino adolescents. Our primary aim is to determine the unique effects of each of the three major intervention components of a 12-week intervention (lifestyle education, stress reduction guided imagery, and lifestyle behavior guided imagery) on obesity-related lifestyle behavior (physical activity and dietary intake) and stress biomarker (salivary cortisol) outcomes. We hypothesize that at the conclusion of the 12-week intervention, the guided imagery arms of the intervention will show greater improvement in these outcomes than lifestyle education alone. Our secondary aim is to determine if changes in outcomes seen after 12-weeks of intervention will be maintained for an additional 12 months. Additional aims of the study seek to determine intervention effect size estimates for secondary outcomes (insulin resistance, body fat, psychosocial outcomes), and to explore the role of Self-Determination Theory constructs (SDT) in mediating lifestyle behavioral change.

2. Methods

2.1. Theoretical Considerations underlying the Imagine HEALTH intervention

2.1.a. Theory-based Approach

The conceptual model for the intervention is illustrated in Figure 1. In contrast to most lifestyle interventions in adolescents which do not identify the active theoretical constructs being targeted by the intervention, we utilized SDT[16] as the underlying theoretical basis of the Imagine HEALTH intervention. SDT is a general theory of human motivation addressing both cognitive and affective factors involved in human motivation and highlights the importance of the social context in facilitating or thwarting autonomous human motivation. This approach to human motivation posits three innate human psychological needs that form the basis for optimal self-motivation: competence, relatedness, and autonomy[17]. Fulfillment of these three needs is necessary to facilitate psychological growth and integration, as well as for personal well-being[18–20]. Autonomous motivation, performance of behaviors with either intrinsic motivation (i.e. for inherent satisfaction) or with extrinsic motivation with high degrees of internal regulation, has been linked to performance of healthy behaviors[17], weight loss maintenance [21], and increased physical activity in adults [22–26] and adolescents[27–29]. The intervention described in this paper is designed to promote the three major SDT components: relatedness, competence, and autonomy. In so doing, it facilitates a sense of choice, volition, and freedom from excessive external pressure as well as focusing on affective determinants of behavior, thus making it developmentally appropriate for adolescents [30, 31].

Figure 1:

Conceptual Model

2.1.b. Combining obesity prevention and treatment in a single intervention.

Typically, interventions to prevent obesity and interventions to treat obesity in children have been considered separately [32, 33]. However, careful consideration of pediatric guidelines shows that many, if not most, treatment modalities are the same or similar for obesity prevention and treatment approaches [34]. Our school-based intervention includes both overweight/obese and normal weight youth, and thus necessarily utilizes elements of both treatment and prevention.

2.1.c. Promoting health and wellness (versus targeting body weight goals).

The Imagine HEALTH intervention approach explicitly targets whole health and well-being. The process (lifestyle behaviors), not the result (weight loss) is emphasized throughout the intervention, and issues relating to insulin resistance and weight loss are viewed in this context. The curriculum intentionally avoids setting program goals based on weight loss, instead focusing on the process of developing healthy lifestyle habits. Weight loss goals would be inappropriate for normal weight participants in the program, and the risks of restrictive eating practices for weight loss in adolescents have been well-documented [35, 36]. The few past interventions sharing this philosophical approach have shown psychological, eating behavior, and body composition benefits [37, 38]. This approach assumes that by improving stress and obesity-related health behaviors (our primary outcomes), improvements in secondary metabolic outcomes (insulin resistance, body fat) will ensue.

2.1.d. Rationale for stress reduction as intervention target: Moving beyond “eat less, exercise more”.

Adolescents have experienced an increase in anxiety over the past few decades [39], suggesting that today’s children may be exposed to increased numbers of stressors than in the past. This is supported by data from the recent Stress in America study by the American Psychological Association, which surveyed a nationally representative sample of 1018 youth aged from 13–17 years old, and found that at least one third of American youths had felt irritable, nervous, anxious, or overwhelmed due to stress in the past month [40]. Minorities also experience higher levels of stress [41]. Socioeconomic and immigration-related factors can be stress-inducing in Latinos [42], and Latino Los Angeles residents are especially susceptible to these forms of stress, given that Latino individuals are 13 times more likely to live in low-income areas than whites[43]. In addition, the high school years can be a vulnerable time [44, 45], and 25.3% of high school students in Los Angeles report being “very stressed” on a daily basis[46].

Chronic life stress has been implicated as a factor in obesity-related disease risk through two mechanisms. First, stress promotes obesity through its effects on eating behaviors, resulting in the ingestion of higher calories, more fat, and calorically dense snack foods [47–51]. Secondly, studies in primates [52, 53] and human adults [54–60] suggest that chronic stress, via neuroendocrine mechanisms producing hypothalamic-pituitary-adrenal (HPA) axis dysregulation, can result in a “pseudo-Cushingoid” obesity phenotype characterized by visceral adiposity, insulin resistance, and metabolic syndrome [56, 61–64]. There is also an extensive literature in adults supporting the relationship between HPA dysregulation and multiple other adverse health outcomes (e.g. development of depression among mentally healthy individuals[65–68] and cancer diagnosis and disease progression[69]). Our own prior research confirms links between stress biomarkers and obesity-related disease risk in overweight/obese adolescents. For example, obese Latino youth with metabolic syndrome have higher morning serum cortisol levels [70], and higher serum cortisol predicts future deterioration of insulin sensitivity [71]. Additionally, a blunted diurnal salivary cortisol pattern predicted increased carotid intima media thickness (IMT), a marker of early subclinical peripheral vascular disease [72]. Utilizing mind-body interventions to treat dysregulated HPA axis activity is a key component of this intervention.

2.2. Protocol approval process

This study was reviewed and approved by the Internal Review Board (IRB, #HS-13–00836) of the University of Southern California Health Science Campus, and went through a rigorous independent regulatory monitoring process, inclusive of annual site visits, commissioned by the study sponsor. Informed consent was obtained from the parents/guardians of all participants, while the participants below the age of 18 signed a youth assent document.

2.3. Study Design Overview

The overall study design is shown in Figure 2. Inclusion and exclusion criteria are outlined in Table 1. Participants receive 12 weeks of intervention, followed by a six-month maintenance program, and then six months of no contact. The intervention is being conducted in three waves of intervention at four schools, each wave consisting of 12 weeks during the spring semester of each year for three years (January through May 2015–2017). This eliminated potential seasonal and winter holiday confounding effects on primary outcomes. The recruitment goal for each of the three waves was 20 students for each of the four intervention arms (n=80 per wave overall).

Figure 2:

Study Design Overview

Table 1:

Inclusion and Exclusion Criteria

Inclusion Criteria	Exclusion Criteria
1) Male and female; age 14–17, in second or third year (i.e. sophomore or junior year) of high school at time of study entry (consent), with stated intention to complete high school to graduation.	1) Serious chronic illness or physical, cognitive, or behavioral disability
2) Overweight/obese and normal weight.	2) Taking medication known to effect body composition (e.g. prednisone, stimulant medications for attention deficit disorder)
	3) Prior diagnosis of clinical eating disorder or psychiatric disorder
	4) Lack of English fluency
3) Agreement to attend up to 3 after-school classes per week for the 12 weeks of the program.	5) Participation in previous school-based “council” programs
	6) Participation in formal weight-loss programs in preceding 3-months
	7) Pregnancy
	8) Sibling or other household member enrolling for the intervention
	9) Participation in after school sports or other extracurricular activities, including more than one AP class *

^*This exclusion was added to the protocol after the first wave of intervention, despite the risk of introducing bias into the study population. In the first wave, we observed significantly poorer adherence to intervention attendance than anticipated, primarily due to conflicts in after-school scheduling among students with multiple extracurricular activities and commitments.

Participants were randomized by school into each of 4 arms of the intervention: 1) No-intervention “Control”; 2) “Lifestyle”, consisting of 12-weeks of twice weekly 75-minute sessions of didactic and experiential education relating to healthy eating and physical activity practices; 3) “Stress Reduction Guided Imagery” (SRGI), which received the same lifestyle education plus an additional once weekly 75-minute group stress reduction GI session for 10 sessions; 4) “Lifestyle Behavior Guided Imagery” (LBGI), which received the same lifestyle education, plus four weekly 75-minute group stress-reduction GI sessions, followed by six subsequent weekly sessions of lifestyle behavior GI. The final 2 weekly sessions for both GI arms contained imagery content that reviewed, summarized, and integrated the entire 12-week program. A summary of the curriculum content of the two guided imagery intervention arms is shown in Table 2.

Table 2:

Imagine HEALTH Guided Imagery Curriculum

	Stress Reduction Guided Imagery		Lifestyle Behavior Guided Imagery
Week	Session Title	GI Content	Session Title	GI Content
1	Stress Reduction 1: Relaxation Breathing	Mindful focused breath	Stress Reduction 1: Relaxation Breathing	Mindful focused breath
2	Stress Reduction 2: Relaxation Breathing and Progressive Muscle Relaxation (PMR)	Following focused breath, relaxation of muscle groups in conjunction with breath, progressing from head to toe	Stress Reduction 2: Relaxation Breathing and Progressive Muscle Relaxation (PMR)	Following focused breath, relaxation of muscle groups in conjunction with breath, progressing from head to toe
3	Stress Reduction 3: Relaxing Place Image	Following focused breath and PMR, exploration of an image of a place that represents just comfort and relaxation	Stress Reduction 2: Relaxing Place Image	Following focused breath and PMR, exploration of an image of a place that represents just comfort and relaxation
4	Stress Reduction 4: Relaxing Place Image; Conditioned Relaxation	Repeat of relaxing place imagery, followed by a second imaginal exploration of relaxing place after only 3-focused breaths (conditioned relaxation)	Stress Reduction 4: Relaxing Place Image; Conditioned Relaxation	Repeat of relaxing place imagery, followed by a second imaginal exploration of relaxing place after only 3-focused breaths (conditioned relaxation)
5	Stress Reduction 5: Conditioned Relaxation; Self-led imagery	Repeat of 3-breaths to relaxing place; Students begin to lead each other in relaxing place imagery exercise	Lifestyle Behavior 1: Hunger-Fullness Image	Imagery of a symbol that represents the state of fullness and hunger, to use as an aid in portion control
6	Stress Reduction 6: Self-led imagery	Repeat of 3-breaths to relaxing place; Students lead each other in relaxing place imagery exercise	Lifestyle Behavior 2: Healthy Eating Image	Imagery of oneself eating healthily
7	Stress Reduction 7: Inner Advisor Image	Imagery of, and dialogue with, an Inner Advisor figure to obtain guidance to meet life stressors.	Lifestyle Behavior 3: Inner Advisor Image	Imagery of, and dialogue with, an Inner Advisor figure to obtain guidance to eat more healthily
8	Stress Reduction 8: Self-led imagery; Rewriting imagery scripts in own words	Participants rewrite relaxing place imagery script in their own words and lead each other in the exercise	Lifestyle Behavior 4: Physical Activity Image	Imagery of oneself participating in physical activity
9	Stress Reduction 9: Inner Advisor Image	Imagery of, and dialogue with, Inner Advisor figure to obtain guidance to improve stress management practices	Lifestyle Behavior 5: Inner Advisor Image	Imagery of, and dialogue with, an Inner Advisor figure to obtain guidance to increase physical activity
10	Stress Reduction 10: Inner Warrior Image	Imagery of, and dialogue with, an Inner Warrior figure to elicit ways to overcome stressful life challenges	Lifestyle Behavior 6: Inner Warrior Image	Imagery of, and dialogue with, an Inner Warrior figure to elicit ways to overcome challenges to eating healthy and being active
11	Integration 1: Future Life Imagery	Imagery of life in future living according to the principles of healthy living learned in program	Integration 1: Future Life Imagery	Imagery of life in future living according to the principles of healthy living learned in program
12	Integration 2: Self-Compassion Imagery Program graduation	Imagery of self-compassion as continue forward with goals of developing healthy lifestyle behaviors	Integration 2: Self-Compassion Imagery Program graduation	Imagery of self-compassion as continue forward with goals of developing healthy lifestyle behaviors

2.4. Rationale for Intervention Components

2.4.a. Rationale for the non-intervention control arm.

The typical student at the study schools gets little direct attention to lifestyle health programs, so that inclusion of this non-intervention control in our study design allows clear documentation of the effects of our other 3 arms above current school-based “standard of care”. Since recent epidemiological data (e.g. www.rwjf.org.healthpolicy) suggest a flattening of the prevalence of childhood obesity, but we do not know if that is true specifically in our highly at risk population, we felt it critical to have a true, non-intervention control to make full conclusions as to the effects of our intervention. Finally, since we will be including a non-obese group, it will be critical to compare normal weight intervention youth with a comparable group that did not receive intervention

2.4.b. Rationale for the lifestyle only arm

The pediatric obesity literature, and our clinical experience in this population, stress the importance of lifestyle education as the first approach in the prevention and treatment of obesity as the first approach [34]. Including this arm of intervention will allow us to determine the effects of this lifestyle education alone compared to non-intervention in this high school population at high risk for obesity-related morbidity.

2.4.c. Rationale for the two guided imagery arms

This design seeks to determine whether adding GI designed to motivate physical activity and healthy eating leads to greater lifestyle behavioral effects than stress-reduction GI alone. We chose to use group GI over individual GI because it is not practical or affordable to use individual GI delivery for such large numbers of students. In addition, a group intervention is preferred, indeed necessary, to maximize the underlying theoretical SDT concept of relatedness.

2.4.d. Rationale for maintenance program and no-contact period

There were several reasons to include the 6-month maintenance program and subsequent 6-month no-contact period. At the conclusion of a previous 4-month lifestyle intervention we conducted in obese Latino adolescents, we saw no short-term improvements in adiposity or insulin resistance outcomes [73], but after a subsequent 8-month maintenance phase, insulin sensitivity improved and dietary intake of added sugar decreased. These data [74] support our conclusion that the 12-week Imagine HEALTH pilot intervention was inadequate to fully assess the effects of the intervention, and longer-term follow-up is needed. This is also consistent with the overall literature in pediatric and adolescent obesity interventions, which generally have a minimum follow-up period of 6–12 months [10, 12, 75, 76]. It is also important to note that behaviour change occurs within a complex adaptive system with multiple components[77]. Therefore, studies with multiple longitudinal follow-up measures may provide the opportunity to examine non-linear behavioral processes otherwise challenging to assess. Finally, the 6-month no-contact phase will allow us to detect the ability of participants to sustain changes beyond the period of direct intervention.

2.4.e. Rationale for after-school delivery of intervention and collaboration with Los Angeles Unified School District (LAUSD).

We collaborated with four LAUSD public high schools to maximize access to the study population. LAUSD is one of the largest public school districts in the country, and successful collaboration with LAUSD could lay the groundwork for a future effectiveness trial, when access to even more student participants will be needed. We met early in our planning phase and followed the suggestions of district-level school administrative leaders to make such early key decisions as delivering the program in the after-school period (vs during actual school time), targeting primarily 11^th graders (for both developmental reasons and because there are fewer high school dropouts among students once they matriculate for the 11^th grade), and confirming the specific choice of schools to participate. The four schools had an average of 1310 students in attendance (range 661–2064), serve predominantly lower-SES (79% eligible for free lunch) Latino student populations (93% Latino, 5% Asian, 1% African American), and are in the East Los Angeles area within a 2-mile radius of our research offices.

2.5. Development and Delivery of Intervention Components

2.5.a. Lifestyle Curriculum

To develop and deliver the lifestyle curriculum, we collaborated with a community non-profit organization, SOSMentor, who had extensive past experience delivering after-school lifestyle programs in LAUSD high schools. We merged their lifestyle curriculum with our previously used Imagine HEALTH pilot lifestyle curriculum. The resulting Imagine HEALTH Lifestyle Curriculum used for this study fully encompasses health-promoting nutrition and physical activity practices consistent with consensus pediatric recommendations [34, 78, 79], emphasizing modification in quality of carbohydrate intake [73, 80, 81], as well as key concepts of “Intuitive Eating” [82–84]. a non-diet approach to healthy eating that bears many similarities to mindfulness-based eating[15]. The final lifestyle curriculum also incorporates an active mentoring component previously used by SOSMentor, whereby the high school participants receive the lifestyle education in the first 6 weeks of the program, and then deliver the intervention to students of a partnered, geographically proximate elementary school during the final 6-weeks of the program. This “train-the-trainer” approach reinforces the lessons and helps build students’ confidence, leadership, and health advocacy skills.

The primary instructors for the lifestyle component of the program were nutrition interns in a masters degree dietetics program of a local collaborating university (Pepperdine University). Other voluntary adult mentors (e.g. volunteer university students or interested community members) served as assistant instructors in the lifestyle classes chosen utilizing SOSMentor’s previous model of identifying and training such staff. The physical education component was taught by a professional trainer. The lifestyle curriculum was taught in twice weekly after-school sessions of 75-minutes each over the course of the 12-weeks. The total of 24 lessons included 12 weekly sessions of nutrition education, 10 sessions of physical activity, and 1 session each for orientation and integration at the beginning and end of the program respectively.

2.5.b. Lifestyle curriculum training and fidelity

All primary lifestyle health educators were fully trained by the subcontractor SOSMentor in accordance with their established procedures. In addition, each underwent a 2–3-hour training with our registered dietician consultant to ensure they understood and could teach the key concepts and approaches of the intuitive eating portion of the curriculum. To ensure fidelity of the Lifestyle curriculum delivery, SOSMentor staff provided on-going training and assistance to teachers and interns in the field, and weekly training and support emails were sent to educators to review upcoming curriculum and use of program materials. Weekly meetings were held with all health educators throughout the duration of the intervention, and experiences and problems with delivery of the program content were discussed.

2.5 c. Guided Imagery Curriculum

Group GI was delivered in the context of the facilitated group process known as “council”. The conduct of council involves the following principles and processes [85]:

1. The group sits in a circle so all members can be easily seen and heard, minimizing hierarchical structure.

2. Only one person speaks at a time. A designated personal object (the “talking piece”) is utilized to focus the discussion, clearly identifying who holds the floor to speak.

3. Group members speak with brevity and clarity, intentionally, “from the heart”.

4. When not speaking, all others listen actively, “from the heart”. Cross-talk is not permitted.

5. All communications are held in confidentiality outside the sessions

We had multiple rationales for using council to deliver the group GI. First, preliminary data from a separate study of ours suggested that council may be ideal to maximize the effects of a group process according to the concepts of SDT, our underlying theoretical model of behavior change [86]. Second, we wanted to replicate the quiet, therapeutic counseling setting used for delivery of individual GI in our pilot studies[15, 87], which maximized participants’ ability to have an inward-directed experience. Similarly, the council format allows participants to quickly enter a mindful state of awareness, enabling them to readily experience the inward-directed process of guided imagery. Third, the LAUSD Council in Schools program has been successfully using council to promote scholastic achievement and other benefits in high schools, middle schools, and elementary schools for many years, and LAUSD was therefore familiar with and accepting of this process. The specific schools participating in this project had not previously used council or participated in the LASUD Council program, except for one school which had a small council program in restorative justice involving a tiny minority of its 2000-plus student body.

During each group GI council session, group members would generally first “check in” with each other, speaking briefly of events since the last meeting, their current state of emotion, and give a brief report of their home guided imagery practice from the preceding week. Group GI exercises were then conducted using pre-written scripts written by the PI. The specific imagery content of each session was based on our pilot Imagine HEALTH study, with minor adaptations, and are summarized in Table 2. Modifications of the standardized processes of individual Interactive Guided Imagery^SM (IGI) [88] were utilized to deliver the GI, consisting of “Foresight” (introduction of the upcoming imagery exercise), “Insight” (the actual imagery experience), and “Hindsight” (debriefing and grounding of the imagery experience). In IGI, participants’ personalized images are used to promote health through several standardized, yet adaptable, techniques including, but not limited to, relaxation/stress reduction, working through resistance to behavioral change, and empowerment through the activation of self-derived insight and inner resources. The GI allows the subject to bypass cognitive processes to obtain insight through understanding of symbolic image content and the affect raised by the image. The facilitator’s goal is to enable the participants to engage their own images, which have salience to their specific health issues, in order to either promote physiological changes (such as stress reduction) or develop self-directed insights and intrinsically motivate health-promoting behavior change. Following the group GI, participants debriefed their imagery experiences in the council format.

Participants were instructed after each weekly session to practice the imagery exercises taught in each week’s session for at least 10 min daily. Subjects were given written copies of the GI scripts, and were Emailed audio file recordings of the GI exercises to help them with home practice.

2.5.d. Guided imagery council curriculum training and fidelity

The GI-council intervention curriculum was adapted from the prior pilot intervention curriculum[15] by the PI. This adapted curriculum included standardized content that directed conduct of the council and the text of the GI scripts delivered in each session, and was used to deliver the intervention in all 3 waves. Prior to each of Waves 2 and 3, minor wording and structural changes in the curriculum were made to ease the delivery of the intervention by the facilitators without any substantive changes in the essential curriculum content. GI scripts remained identical throughout the intervention.

All GI facilitators completed Level One certification training in IGI (http://www.acadgi.com/certification/), including training in all GI techniques utilized in this intervention. All staff involved with GI council delivery additionally received a 2-day training in the goals, objectives, and methods of council, led by council consultant (JP) and the PI (MW), just prior to the onset of intervention Wave 1, with re-training one-day “boosters” prior to beginning each subsequent intervention wave.

The PI delivered the GI-council curriculum to both GI arms in the first wave of intervention, during which the other four facilitators underwent final training by participating as co-facilitators. These 4 facilitators then served as the primary GI council facilitators for the 2 subsequent waves of the intervention, 2 facilitators for each GI arm. All facilitators were highly educated and skilled, and included an academic social worker, an integrative physical therapist, an expressive arts therapist with a MA in psychology and extensive group facilitating experience in schools, and a long-standing AGI-certified guided imagery practitioner. The pairing of co-facilitators was determined by the PI to optimally balance the experience and skills of the facilitator pairs, and they were randomly assigned to SRGI or LBGI arms for the conduct of intervention Waves 2 and 3.

To ensure fidelity of the intervention during Waves 2 and 3, the PI conducted weekly video conferences with all GI council facilitators, during which that week’s GI council session was discussed, and any deviation from program fidelity was immediately identified and corrected. Curriculum for the upcoming week’s intervention session was then previewed to prepare intervention staff to deliver the content and meet the designed session goals. All GI-council intervention sessions were videotaped and selectively reviewed by the PI or other project staff. Finally, delivery of the pre-determined goals of each session was quantified by study staff, either in real-time during the sessions or upon review of session videos.

2.5.e. Maintenance Curriculum

Upon completion of the 12-week intervention, participants received a monthly 75-minute after-school Lifestyle Education class or GI council session for 6 months, taught by the same facilitators. Monthly lifestyle class content reviewed key constructs of the lifestyle lessons taught in the initial 12-week program. Monthly GI council sessions included a brief (10–15 minute) review of the key lifestyle constructs, followed by review and practice of the GI exercises from the initial 12-week program.

2.5.f. Parent program

Immediately following consenting as described below, a 15-minute bilingual group class was delivered to parents and participants together, describing the basic content of the lifestyle education class, including description of the main dietary and nutritional curricular elements. Parents of participants in the active intervention arms were subsequently mailed a quarterly newsletter during the intervention phase containing information from the curriculum, providing for their education and encouraging them to support their child’s process in the program. Finally, the parents received a “Students Teaching Parents” component at the end of the 12-week intervention, in which participants shared information they had learned in the lifestyle education curriculum.

2.6. Recruitment and consenting processes

Recruitment staff visited 10^th and 11^th grade classrooms of required courses (e.g. History and English) to present program features to a broad cross section of students at participating schools. Interested prospective student participants were given IRB-approved flyers, explanations of the program, and filled out a participant application questionnaire. Participant applications were reviewed by study staff, and those who potentially met inclusion criteria were then pre-screened by telephone. Following successful pre-screening, prospective participants and their parents attended a group information and consenting visit at their respective high school campus, conducted by bi-literate research staff (English and Spanish). During the group consenting visit, research staff fully described the program, determined final exclusion/inclusion criteria, and answered all questions. Once the parent and subject verbally demonstrated understanding and agreement, the parental consent and youth assent forms were signed.

2.7. Program Adherence

Adherence was assessed by taking attendance to all intervention sessions. Adherence was defined on the individual participant level as attendance by the individual participant to at least 9 of 12 classes of each component of the intervention (i.e. 75% of lifestyle classes and 75% of GI council sessions). For the purposes of the intent to treat analysis, we defined adherence at the group level as 80% of participants attending at least 75% of classes. Adherence was monitored, and measures for improving adherence were discussed and implemented when needed. For example, participation in after-school sports or more than 1 AP class was added as an exclusion criteria for Waves 2 and 3 for this reason, as multiple after-school commitments clearly adversely affected adherence during Wave 1.

To promote participant attendance and adherence to intervention visits, assigned study staff texted and/or called participants to remind them to attend the weekly intervention sessions and monthly maintenance classes. Participants were also texted/called by study staff when they did not arrive within 10 minutes of the intervention session start time. For intervention Waves 2 and 3, the responsibility for these calls and texts was assigned to a single, young adult, masters-level, staff person, who worked to develop rapport with participants and to role model behavior and improve adherence with the intervention sessions. If participants missed any two consecutive classes, they were texted/called the day following the second missed class. If a participant could not be contacted on the first phone call/text attempt, staff conducted a second phone call the following day. If a participant remained unresponsive, every effort was made to contact the participant’s parents. To remind participants of their ongoing participation, study staff mailed participants birthday and holiday cards.

2.8. Compensation and incentives

Participants received $20 cash compensation for completing each consent and measurement visit upon returning all measurement equipment, up to $100 total. Participants were also eligible for an additional $10 gift card if they submitted at least 2 complete days of saliva samples. Travel reimbursements of $10 per visit were given to families who provided their own transportation. Other non-cash incentives were offered (e.g. movie tickets, raffle prizes, gift cards) to encourage participant attendance to sessions. Scholastic incentives were also offered in the form of service learning credits that could be used to meet graduation requirements. In an effort to improve adherence/attendance after Wave 1, participants in Waves 2 and 3 who attended at least 75% of classes received a letter of recognition signed by the PI stating the degree of their participation, which they could use for future college or work applications. Students in the non-intervention Control arm received a copy of the lifestyle class curriculum materials at the completion of the 12-week intervention period.

2.9. Concomitant Interventions

There were no specific allowed, required, or prohibited interventions. A checklist was administered at each measurement visit that asked about participation in concomitant interventions since the last, preceding measurement visit.

2.10. Adverse Events

An adverse event (AE) was defined as any untoward medical occurrence in a subject during participation in the clinical study, including any sign, symptom, abnormal assessment (laboratory test value, vital signs, electrocardiogram finding, etc.). A serious adverse event (SAE) is any AE that results in one or more of the following outcomes: 1) death; 2) a life-threatening event; 3) inpatient hospitalization or prolongation of existing hospitalization; 4) persistent or significant disability/incapacity; 5) an important medical event based upon appropriate medical judgment. AEs were labeled according to severity: “mild” if it did not have a major impact on the participant, “moderate” if it caused the participant some minor inconvenience, and “severe” if it caused a substantial disruption to the participant’s well-being.” AEs were also categorized according to the likelihood that they were related to the study intervention, labelling them as definitely unrelated, definitely related, probably related, or possibly related to the study intervention. Expected risks of the study were considered to be minimal.

2.11. Independent Monitoring Committee

The Independent Monitoring Committee (IMC) for this study was comprised of four expert scientists who were not associated with this research project, and who were qualified to review the patient safety data generated by this study because of their unique expertise and qualifications in the areas of biostatistics, childhood obesity, conduct of clinical trials, and school-based CAM lifestyle interventions.

The IMC met yearly to review: (1) an Annual Report that included a list and summary of AEs, and whether AE rates were consistent with pre-study assumptions; (2) reason for dropouts from the study; (3) whether all participants met entry criteria; (4) whether continuation of the study is justified on the basis that additional data are needed to accomplish the stated aims of the study; and (5) conditions whereby the study might be terminated prematurely. The Annual Report and signed recommendations and comments from the chair of the IMC were sent to the IRB and NCCIH sponsor.

2.12. Six-week pilot of preliminary intervention and measurement procedures

The delivery of the intervention was pilot tested in a sample of 17 students in a high school of similar demographics to our study high schools. Entry criteria were the same as for the main study except that students from any grade were invited to participate. We delivered an abbreviated 6-week intervention, consisting of 2 lifestyle education classes and 1 GI council session each week. The aims of the pilot were primarily to work out implementation issues before beginning the full intervention. During this pilot, we delivered selected GI council sessions to ensure that adaptations made in the GI curriculum (e.g. Stress reduction, Inner Advisor and Inner Warrior imagery) would be presentable and deliverable in group format rather than in the individual GI format of our prior report[15]. We also worked out logistical issues such as procedures for home practice of guided imagery, measurement visits, and the subsequent week-long ambulatory assessment. Outcome measures were obtained as described below for the full study, except that no blood measures were obtained. After post-intervention measurements, a group exit interview with the participants was held to enable us to determine acceptability and feasibility of critical elements of the program.

Our experience during this 6-week pilot of the intervention procedures and methods generally supported the assumptions and decisions we had made in making our original adaptation of the curriculum. The program curriculum was then finalized for use in the full intervention. The pilot experience also supported the acceptability and feasibility of doing this intervention in three after-school sessions weekly, the use of particular guided imagery components and exercises delivered in the group setting, and the overall approach and logistics of our complicated outcome measurement assessment procedures.

2.13. Outcome Measures

Primary and secondary outcome measures were assessed at baseline, 3-months (immediately post-intervention), 9-months (after 6-month maintenance program), and 15-months (after 6-month no-contact period). In addition, intra-intervention and descriptive outcomes were assessed. The specific outcome measures are fully described in Table 3.

Table 3:

Outcome Measures

Outcome	Measure/Instrument	Description
Primary Outcomes *
Physical Activity	Accelerometer	The trial-axial Actigraph accelerometer model GT3X+ was used to collect the primary, objective outcome measure of time spent in physical activity. The accelerometer is a well-validated mode of objective physical activity measurement and has been utilized in large scale national studies such as the NHANES [98]. Participants are instructed to wear the accelerometer on the wrist of their non-dominant hand at all times, except during water-related activities (e.g. showering, swimming, etc). Participants are provided with a paper- and-pencil log to record the beginning and the end of non-wear time. The accelerometer is set to collect data at 30-sec epochs. Upon retrieving the accelerometers, the raw data will be immediately downloaded to the ActiLife software version 6.0 for processing. Both wear time validation and categorization of time spent in physical activity in various intensities will be calculated using algorithms developed for wrist-worn accelerometers.
Physical Activity	Physical Activity Recall	Self-reported physical activity data was collected to supplement the accelerometry outcome, using a modified version of the 3-Day Physical Activity Recall, which has been well validated against accelerometry data in adolescents [95]. This yields additional information as to pattern and type of physical activity. For each measurement visit, participants were required to have at least 2 days of activity recall interviews, simultaneous to salivary cortisol sampling. The metabolic equivalence of intensity level (MET) of each day is derived from interview data in accordance with the Compendium of Physical Activities. [99]
Dietary Intake	Dietary Recall	Dietary intake was collected using multiple pass dietary recall interviews using the Nutrition Data System for Research (NDS-R version 2014–2017). The NDS-R-assisted interview protocol is a well-validated 24-hour recall dietary assessment method [94] that uses a standardized protocol based on the “multiple pass” method, which was developed and tested by the USDA for use in the 1994–1996 CSFII in an effort to limit the extent of under-reporting. This approach has previously been validated against total energy expenditure by doubly labeled water in children[100]. For each measurement visit, participants were required to have at least 2 days of activity recall interviews, simultaneous to salivary cortisol sampling.
Diurnal Salivary Cortisol Pattern	Commercially available ELISA (Salimetrics, Inc; inter-assay CV = 3.75% [high], 6.41% [low])	During the week-long data collection in the free-living environment, participants collected saliva samples at 3 time points per day (upon awakening, 30 minutes post-awakening, and near bed time) for three consecutive weekdays. Participants were instructed not to eat, drink, smoke, or rinse their mouth for at least 5 minutes prior to saliva sample collection. Salivary samples were obtained using the Salivette system (Sarstedt, Newton, NC) by placing a cotton swab to passively absorb saliva [101]. Salivette samples were kept in subjects’ home freezers until retrieved by study personnel.
Secondary Outcomes *
Stress Measures
Perceiv ed Stress	Perceived Stress Scale	17-item modified version [102] of the original 14-item Perceived Stress Scale [103], which assesses perception of stress in the preceding month and has good psychometric properties (μ = 0.68).
Fasting Serum cortisol	Commercial ELISA assay (Alpco) Inter-assay CV = 3.8% [high], 8.1% [low]	Cortisol was measured in a fasting serum sample collected on the morning of each measurement visit.
Well-Being & Psychological Measures
Self-esteem	Self Description Questionnaire (SDQ) General Self subscale [104]	The SDQ was constructed specifically for adolescent populations and has high validity and internal consistency, with Chronbach’s alpha = .86 in our study population.
General Well-Being	Psychological General Well-Being Index, positive well-being construct [105]	4-item Likert scale, which addresses how the individual feels about their “inner personal state,” has been used with participants as young as 14 years old with high validity and reliability[106].
Global Well-Being	Arizona Integrated Outcomes Scale [107]	2-item visual acuity scale in which subjects are asked to rate their overall sense of well-being, taking into account their physical, mental, emotional, social, and spiritual condition for the past 24-hours and the past month. Higher scores on this scale has been associated with beneficial health markers in college students [107].
Mindful ness	Mindful Attention Awareness Scale for Children [108]	15-item Likert scale self-report instrument,validated in adolescents [109]
Intuitive Eating	Intuitive Eating Scale [110]	23-item Likert scale instrument has been linked to health markers in college populations. Wording of 2 items underwent minor adaptation for use with adolescent study population.
Emotion al Eating	Dutch Eating Behavior Questionnaire (DEBQ)[111]	13-item Likert scale Emotional Eating subscale of the DEBQ (μ=0.83)[111].
Depress ion	Center for Epidemiological Diseases Depression Scale (CES-D)[112]	20-item scale assesses depression symptoms
Adiposity & Metabolic Measures
Body Mass Index (BMI)	Calculated as weight (kg)/height (m2).	Height and weight are measured in triplicate using a wall-mounted stadiometer (height) and clinical medical balance (weight). BMI percentile and Z-score (standard deviation score) are determined using Epi Info software[113].
Body composition	Percent body fat and lean tissue	Bioimpedance analysis (BIA), using the Tanita (Arlington Heights, IL) TBF-310GS Body Composition Analyzer/Scale[114]
Insulin Resistan ce	Fasting Glucose (YSI analyzer; glucose oxidase method). Fasting Insulin (EMD Millipore Luminex xMAP multi-analyte platform)	Insulin resistance is estimated using the Homeostatic Insulin Resistance Model (HOMA-IR), based on a single fasting blood sample (HOMA-IR = fasting glucose (mmol/L) X fasting insulin μu/ml) / 22.5). HOMA correlates with insulin sensitivity in overweight Latino youth (r= −.57, p<.05) [115].
Fasting lipids	Roche Cobas clinical enzymatic analyzer	Measured to determine diagnoses of metabolic syndrome according to standard pediatric definitions[116, 117].
Waist circumf erence	Flexible tape measure	Measured in duplicate using a tape measure at the level of the iliac crest per NHANES methodology [118] to determine diagnosis of metabolic syndrome according to standard pediatric definitions [116, 117].
Blood pressure	Clinical automated blood pressure cuff (Welch Allyn)	Measured in duplicate using an appropriately sized cuff in a seated position with left arm resting on table at heart level, to determine diagnoses of metabolic syndrome according to standard pediatric definitions [116, 117].
Hemogl obin AIC	High performance liquid chromatography performed on Tosoh clinical analyzer	Measured to determine diagnoses of pre-diabetes and metabolic syndrome according to standard pediatric definitions [116, 117]
Inflamm atory Cytokin es	Interleukin-6 (IL-6) and Tumor Necrosis Factor - alpha (TNF-a) (EMD Millipore Luminex xMAP multi-analyte platform)	Measured as an assessment of stress-related inflammation [119]
Self-Determination Theory measures
Compet ence	Perceived Competence for Participating in Regular Physical Activity and Healthy Eating (PCS[120])	8-item scale assesses perceptions of competence related to physical activity and healthy eating behaviors.
Autono my Support	Health Care Climate Questionaire[121]	12-item scale assessing autonomy support for eating and for exercise from health care providers (6 items for each behavior)
Autono mous motivation	Treatment Self-Regulation Questionnaire [122]	30-item scale measuring motivation for eating and for exercise (15 items for each behavior) across the spectrum of motivation (e.g. amotivation, controlled motivation, and autonomous motivation).
Related ness	Group Cohesion Scale[123]	10-item scale validated in lower income Latino youth
Intra-Intervention Measures **
Acute Salivary Cortisol Change	Commercially available ELISA (Salimetrics, Inc; inter-assay CV = 3.75% [high], 6.41% [low] )	Salivary cortisol was collected pre- and post-intervention sessions (approximately 75 minutes apart) during weeks 3 and 4 of the program to assess acute change in salivary cortisol in response to group stress reduction guided imagery.
Acute Stress Mood	Measured using a visual analog scale adapted from the Profile of Mood States[124]	Was collected pre- and post-session (approximately 75 minutes apart, simultaneously with salivary cortisol collection) during weeks 3 and 4 of the program to assess acute change in stress mood in response to group stress reduction guided imagery.
Perceiv ed Stress	Perceived Stress Scale	17-item modified version [102], of the original 14-item Perceived Stress Scale [103], which assesses perception of stress in the preceding month and has good psychometric properties (μ = 0.68).
Descriptive Measures **
Guided Imagery Home Practice	Self-report instrument developed by investigators	Brief 2-item self-report survey administered at the start of each guided imagery session (Sessions 2–12) inquiring about the number of days and the amount of time per day participants practiced guided imagery during the preceding week
Progra m Diffusion	Self-report instrument developed by investigators	A 2-questions survey that assesses the degree to which program content may have been shared between participants in different arms of the intervention, administered at the completion of the program.
Demogr aphics	Self-report instrument developed by investigators	Completed at the time of consent and enrollment, this survey assesses age, sex, self-reported race/ethnicity, parental education level, occupation, and household income
Accultur ation	Acculturation, Habits and Interests Multicultural Scale for Adolescents (AHIMSA) scale[125].	Completed at the time of consent and enrollment, this 8-item, multicultural, multidimensional measure developed specifically for adolescents provides measures of assimilation, separation, integration and marginalization
Comple mentary and alternati ve medicin e (CAM) use	Complimentary Alternative Medicine Survey[126]	CAM use was measured at each outcome visit using an adapted version of the survey described by Wilson.
Social Connect edness	Keyes’ Social Well-Being Scale[127]	7-item Likert scale with good psychometric properties

^*All primary and secondary outcomes were measured at Baseline (pre-intervention), 3-months (post-intervention), 9-months (post-maintenance), and 15-months (post-no contact period)

^**Intra-intervention and Descriptive measures were assessed at times described

2.14. Measurement Visits: Procedures for outcome measurements

Measurement visits were conducted at the USC Diabetes and Obesity Research Institute (DORI) laboratory on weekend mornings at each of the measurement timepoints. Average time for the in-lab data collection and training visit was 2–3 hours across the intervention.

Participants arrived following an overnight fast. Certified phlebotomists and trained research staff obtained fasting blood draw and body composition measures, after which breakfast was served. Surveys were then administered in an adjacent conference room space, whereby participants entered survey responses directly into the REDCap database platform using electronic tablet devices, taking about 20 minutes to complete all surveys. Staff conducting the measures were blinded to intervention condition to minimize bias during data collection points.

At the conclusion of the in-lab measurement session, each participant received training for the upcoming week-long data collection, including wearing Actigraph accelerometers [89], collecting 3-day home salivary samples, and responding to dietary and physical activity recall interviews. Each participant also received an assigned package with individually labeled data collection apparatus for primary outcomes and supporting documents for the following week, including accelerometer wear instructions, accelerometer wear-time log, instructions and tubes for saliva sample collection, a paper copy of the activity recall, and a food amount booklet for portion estimation.

The project staff also led a practice run-through of the salivary cortisol sample collection procedure. Saliva was collected at three timepoints (awakening, awakening + 30-minutes, and bedtime), on three weekdays of the ambulatory assessment week. Each participant was loaned an Android phone with an installed application, ZEMI, which was developed and validated [90] in-house specifically to facilitate the collection of salivary cortisol data at the designated times, and to avoid the common pitfalls in obtaining accurate timing of salivary collection in a free-living environment [91–93]. Study staff personalized the sampling alarm time for each participant so that during the subsequent week, ZEMI emits alarms that are personalized to the participants’ unique schedule. Participants initiate the saliva sample when the Zemi alarm sounds, upload a picture of the completed sample as prompted by ZEMI, and answer a brief a 5-item profile of mood state scale on a visual analog scale. ZEMI also emits additional reminder alarms twice in 10 minute intervals if the participant did not respond to the initial alarm and then becomes inaccessible 10 minutes after the second reminder alarm. Both the photos and the mood state responses are time-stamped and stored in a secure server.

In addition to salivary cortisol collection and accelerometry, self-reported physical activity and dietary behavior recall interviews were conducted on at least two week days during the week-long ambulatory assessment, coinciding with the days of salivary cortisol sampling. For each wave of measurement, approximately 5–10 bachelors or masters degree program nutrition students recruited from a neighboring academic institution served as interviewers for this data collection, and were trained by co-investigators (JD and CW) in previous day dietary data recall interview using the Nutrition Data System for Research NDSR[94] and the Physical Activity Recall (PAR)[95]. Interviewers conducted at least 10 practice interviews, and received a final evaluation before data collection was initiated. Interviewers also received a brief “booster” training prior to each measurement period (e.g. 3-month, 9-month, 15-month).

2.15. Data Management

In collaboration with Biostatistics and Bioinformatics Resources Program at the Southern California Clinical and Translational Science Institute (SC-CTSI), a (Research Electronic Data Capture) REDCap database[96] was developed and utilized for this study. REDCap is a secure, web-based application designed to support data capture for research studies, providing: 1) an intuitive interface for validated data entry; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for importing data from external sources [97]. REDCap allowed for data collection of surveys to be captured immediately and directly from participants via electronic tablets. Reports regarding the completeness of data and progress were monitored by the study statisticians through the course of the study to keep study staff, IRB, and IMC apprised of the creation of recruitment, retention, missing data, and adverse event reporting.

2.16. Statistical Considerations

Descriptive and exploratory analyses will be performed for all variables, including patient characteristics, clinical and psychological measures, assessment of attrition, missing data, and adherence. At this time, outliers will be determined and data transformations applied as needed. Intention to treat population will include all participants randomized within the schools. Per protocol, sensitivity population will include all participants randomized within the schools who completed at least 80% of instructional time. For this cluster randomized trial, all analyses will be nested with in schools to account for the variability shared within school environment. Statistical modeling of continuous outcomes will be accomplished through use of longitudinal linear mixed effects modeling with diagonal covariance matrices to account for repeated measures, while categorical outcomes will be compared using generalized estimating equations. Preliminary analyses of pre-randomization values of outcomes and participant characteristics will be compared across arms to assure that randomization was successful across intervention arms. Variables that are unequal at baseline will be included as covariates per protocol sensitivity analyses after ITT analyses have been completed. A conservative p < 0.20 will be used to determine if such sensitivity analyses are required. Arm effects will be assessed using omnibus hypotheses across all arms with the primary endpoint of 12-weeks, followed by post-hoc comparisons of arm effect. A priori covariates include baseline values, age, gender, ethnicity, BMI, and year in the study. To account for the possibility of different types of changes between participants who are overweight/obese vs normal body composition, we will add in BMI category (characterized as a dichotomous random effects indicator of normal vs overweight at baseline) as an effect modifier in our mixed effects modeling. This interaction of starting arm by treatment will allow for the possibility that the effects due to treatment received would be different based on starting values. Year in the study will be included to help account for a potential spillover effect, and if it is specific to any specific treatment arm. For these comparisons, α = 0.05.

Maintenance effects will involve comparing equivalence in change from post-intervention (3-months) through the final visit at 15-months post randomization. Thus, the primary endpoint for maintenance models will be one year after the intervention ended (15 month visit). These models will also be analyzed using longitudinal linear mixed effects models or generalized estimating equations utilizing all data collected through the study (pre-randomization, post-intervention, 9-month, and 15-month after post-intervention.) The limit of agreement for maintenance will be set at 20% across variables. If change is ≤20%, then we will find for the alternative hypothesis, no change.

We have also defined two a priori exploratory aims which follow the analysis plan set forth above. The first exploratory aim will examine the proportion of participants classified as showing “favorable outcome” in terms of body composition at post-test. This dichotomous variable will be defined as either: 1) improvement in body fat/BMI (in participants who are overweight or obese at entry); or, 2) maintenance of body fat/BMI (in participants who are normal BMI at entry). While we do not anticipate being powered to detect differences in BMI category, this characterization may be useful in powering a future trial with the focus of addressing weight loss in this population. It will also allow us to examine whether this mind-body modality is equally effective in generating improvements as maintaining healthy body fat across a heterogeneous population. For participants who are overweight or obese, any reduction of BMI-Z-score will be considered a favorable outcome, while a BMI-Z score that is the same or increased will be unfavorable. For normal weight participants, maintaining BMI-Z-score < 0.5 points over or within the healthy range is considered favorable, while a BMI Z-score that increases by 0.5 or more points or into the overweight/obese range is unfavorable.

The second exploratory aim will use structural equation models (SEM) in MPlus to model the mediating effects of psychosocial constructs on behavioral, and more distally, metabolic outcomes (see Figure 1 for conceptual model). A priori covariates include age and sex for all analyses, and body composition (BMI and/or body fat) for HOMA-IR. Model fit for SEM models will be assessed using multiple fit indices: the fit of the overall model will include an assessment of parsimony adjusted sample based indices (PNFI, PGFI, AIC), and population based criteria (RMSEA). Residuals will be examined for points of model misfit. The best model will be indicated by the convergence of goodness of fit and small residuals. The significance of hypothesized paths will be assessed by comparing the LRT of nested models, which approximates a χ² distribution (degree of freedoms of the model(df) = dfΔ of the models).

Sample Size and Randomization

Estimates of power and sample size were computed a priori for a cluster randomized trial including four schools. Sample sizes were estimated for effect sizes ranging from Δ = .7 – 3.0, following our preliminary findings[15]. We assumed intra-individual correlation across time to be moderate (0.65 – 0.80) and within school correlation as low (0.10 – 0.20). We determined that the proposed sample size of 60 per arm (N = 240 total) would be sufficient to detect effects in the target range for the primary aim, assumed an attrition rate of ≤ 20% at the end of the intervention (3 months) based on previous work within this population.

Treatment Assignment Procedures

The study biostatistician performed the randomization of schools and maintained the master randomization list, sharing with the research coordinator after baseline measurements were made at each wave. Study administrative staff were notified 2–3 weeks in advance of each wave of which schools were randomized to allow for staff coverage for the intervention. Participants were notified of assignment at first day of participation.

3. Results

Recruitment and enrollment.

In general, our recruitment and consenting scheme went as planned and described in Section 2.6 above. Results of our recruitment efforts are shown in Figure 3. Of the 965 students who were excluded at the pre-screening step, the main reasons for exclusion were incomplete or illegible screening form (n = 591, 61%), followed by being enrolled in a competing after-school activity at the time (e.g., AP prep courses or after school sports, n = 215, 22%), and unwillingness to commit to the time prescribed by the intervention (n = 83, 9%). Recruitment for Wave 1 (n=64) was lower than anticipated (target recruitment = 80 per wave) due to unavoidable time factors delaying onset of recruitment procedures. The recruitment deficit was mostly made up in Wave 2, with recruitment of 91 participants. The total number of participants recruited for the 3 waves of intervention combined was 232, representing 97% of the a priori recruitment goal of 240 total participants. The recruitment efficiency was 64.3%, with 232 consented out of 361 deemed eligible after prescreening procedures.

Figure 3:

Recruitment and Consent

4. Discussion & Conclusions

This randomized controlled intervention offers several innovations not addressed by prior lifestyle interventions for obesity prevention and treatment in adolescents. Most specifically, we have incorporated the mind-body modality of guided imagery as an integrative healing modality which has previously shown promising benefits in the area of improving physical activity and stress hormone (cortisol) reduction [15]. The emphasis on stress reduction in this intervention is based on the growing evidence linking chronic life stress to obesity-related metabolic disease, and has not, to our knowledge, been utilized in past obesity interventions on such a large scale. In addition, our emphasis on whole health and wellness as opposed to emphasizing weight loss, and the targeting of both obesity prevention and management, follows from our integrative philosophical approach as well as the demands of working with a basically healthy population of youth of all weights and sizes. The use of a clear, underlying theoretical basis for the intervention, Self-determination Theory, also sets this study apart from most past obesity interventions in youth. Finally, our work with predominantly Latino youth will add significantly to the pediatric obesity field as it relates to populations of youth suffering from significant health disparities, and for whom the evidence base for lifestyle interventions remains relatively sparse [11].

Our recruitment experiences thus far have demonstrated our ability to recruit urban adolescents for an after-school intervention that carries a significant degree of study burden for the participants. Our recruitment efficiency was quite high at 64%, indicating generally successful recruitment processes. In fact, 99% of those prospective participants who actually attended the group consenting visits were in fact consented, as the vast majority of the 129 eligible participants who did not ultimately enroll in the study were those who failed to show up for their scheduled consenting visit. Many of these no-shows were related to the inability to reach prospective recruits to confirm consenting visit times and locations due to changes in phone numbers or cancellations of phone plans between the time of school recruitment and consent. A key factor to our recruitment success was the utilization of young bilingual and bicultural recruitment staff to screen the participants and parents. Having cultural and linguistic sensitivity allowed our staff to develop a good rapport and provide reassuring explanations to parents of the participants. In addition, conducting pre-screening phone calls during the evenings and weekends to reach parents was crucial, as was offering evening and weekend group consenting times, locating recruitment sessions at their nearby school which was already familiar to them, and providing food during the consenting visits.

In designing the study, we were quite concerned about the degree of study burden on our participants, both due to the number of sessions per week and the number of outcome measurement visits of high intensity. Despite the need to attend 2 or 3 after school sessions, youth in our 6-week pilot phase of the program development felt this was not overly burdensome, encouraging us to proceed with this after-school design for the full intervention. Subsequently in our recruiting and consenting activities, we found that the majority of eligible youth were willing to commit to the 2- or 3-day after-school intervention, and that they seemed generally interested and intrigued by the study methods. Anecdotally during the recruitment and consenting process, and in line with our experiences in our prior work, we found this population of urban youth to be generally naïve with respect to their exposure to mind-body integrative modalities. Despite this, our pilot work has shown that both the council group process [86] and the guided imagery itself [15] have been well received in similar populations of urban youth that we have worked with, including in the 6-week pilot we conducted as part of the final stages to the development of this intervention.

It is important to note that this intervention requires a great deal of skill and training in the facilitators to deliver it properly. Given our multiple intervention facilitators, it is critical to maintain treatment fidelity across the length of the intervention. Therefore, we spent a considerable effort in developing processes that will ensure that each of the guided imagery treatment arms receives the intervention as it was intended.

This study with its randomized design aims to determine the effects on critical obesity-related lifestyle behaviors and stress biomarker measures, above non-treatment controls and lifestyle education only, of 2 different guided imagery curricula, one based solely on stress reduction guided imagery, and one with both stress reduction and behavioral guided imagery targets. This is important to determine, since our prior pilot study combined both stress reduction imagery and lifestyle behavioral imagery, the latter of which involves more complex imagery exercises and may prove more difficult in future dissemination efforts. Thus, it is important to know if stress reduction guided imagery alone, which is less rigorous and generally more easily deliverable, will be as beneficial or more beneficial than the behavioral imagery. In addition to the main outcomes, this study will provide new information on effects on psychosocial outcomes, including positive psychological constructs such as well-being. Finally, assessment of physiological consequences of obesity such as insulin resistance and metabolic syndrome, moves the study beyond the typical weight-based outcomes of most pediatric obesity-related interventions, and begins to address the true cost of obesity in terms of morbidity and actual metabolic health consequences. The magnitude of effects of the intervention on these key metabolic parameters will also help guide subsequent steps to develop a future trial aimed at significantly reducing obesity-related disease risk in obese youth.

In conclusion, we have described in detail the development of a 12-week lifestyle intervention that seeks to improve lifestyle behaviors and reduce markers of chronic stress in a minority population of adolescents at considerable risk for both obesity and obesity-related disorders. The described intervention is theory-based and innovative in many aspects. The RCT design is powered to determine efficacy outcomes on the primary measures of physical activity, dietary intake, and salivary cortisol patterns. Results will determine the next steps in the development of a fully-powered efficacy trial to reduce insulin resistance and diabetes risk in Latino adolescents.

Abbreviations:

T2D

Type 2 diabetes

SDT

Self-Determination Theory

GI

Guided imagery

IGI

Interactive Guided Imagery^SM

RCT

Randomized controlled trial