Optimizing an Extended Care Intervention to Promote Weight Loss Maintenance: Protocol for the OPT-X Multiphase Optimization Strategy Factorial Trial

Abstract

Background:

Weight loss maintenance (WLM) remains a limitation of behavioral obesity interventions, as long-term adherence is burdensome and diminishes over time. Simplified recommendations and/or alternative delivery methods that target evidence-based WLM strategies may reduce burden to improve long-term outcomes.

Methods:

This protocol describes a multiphase optimization strategy (MOST) trial using a 2x2x2x2 factorial design to optimize a 12-month extended care program for weight loss maintenance. Adults with obesity who achieve at least 5% weight loss during an initial 4-month behavioral weight loss program are randomized to receive a core maintenance program plus zero to four “minimally disruptive” intervention components: reduced food variety, home-based resistance training, buddy training and support, and acceptance and commitment therapy workshops.

Results:

The primary outcome is weight loss maintenance at 12 months following randomization. Secondary outcomes include treatment adherence, treatment burden, behavioral maintenance constructs, and hypothesized mediators and moderators of intervention effects.

Discussion:

This study explores unconventional but evidence-based treatment strategies for WLM and applies a ‘minimally disruptive medicine’ paradigm to behavioral maintenance. These findings have the potential to inform best practices for the provision of WLM interventions, improve longterm outcomes for individuals accessing weight management services, and strengthen the overall public health impact of available interventions for disease prevention and management.

1. Background

The promotion of long-term weight loss maintenance (WLM) remains one of the greatest challenges of behavioral obesity management. Although behavioral interventions for obesity achieve significant and clinically meaningful initial reductions in body weight that help prevent and manage numerous chronic conditions,^1–3 individuals typically regain at least half of their lost weight in the first year after treatment.^4–8 Over time, only 20% of individuals successfully keep off lost weight.^4,5,8 Extended care programs (i.e., continued intervention delivery and support beyond the initial treatment period) can improve long-term weight outcomes by promoting continued reductions in energy intake, increased energy expenditure via physical activity (PA), and use of behavioral strategies to facilitate social support and motivation.^6,9–13 Despite the documented benefits of extended care, challenges to long-term WLM persist. Sustaining behaviors required for WLM is often difficult because of competing demands, behavioral fatigue, and waning motivation, leading to declining adherence and weight regain.^4,5,8,14,15 Strategies to alleviate the burden of WLM and encourage sustained adherence are needed.

Simplified recommendations and/or alternative delivery methods that target evidence-based WLM behaviors strategies (e.g., strategies for caloric restriction that reduce the effort of meal planning, convenient and accessible PA, digital behavioral support) may minimize the effort required by treatment and strengthen individuals’ capacity to persist in WLM behaviors. Such strategies align with “minimally disruptive medicine” (i.e., care designed to advance patient priorities while minimizing treatment burden)^16,17 and, when coupled with intentional targeting of theoretical drivers of behavior maintenance,¹⁸ represent a promising approach that may favorably shift the burden-capacity balance and support sustained treatment adherence. This trial, Optimizing an Extended Care Intervention to Promote Weight Loss Maintenance (OPT-X), evaluates four “minimally disruptive” WLM strategies, including 1) reduced food variety (RFV), 2) home-based resistance training (RT), 3) buddy training and support, and 4) acceptance and commitment therapy (ACT) workshops. These components have empirical support for weight management and are designed to enhance behavioral maintenance with less burden. To efficiently and rigorously evaluate these WLM strategies, this extended care trial utilizes the multiphase optimization strategy (MOST) framework with a factorial experimental design.¹⁹

2. Methods

2.1. Conceptual Model

A comprehensive systematic review of 100 published theories identified five foundational domains explaining behavior change maintenance: 1) maintenance motives, defined as satisfaction with behavior or its outcome, or behavioral congruence with self-identity and values; 2) habits, defined as automatic behaviors supported by relevant cues or triggers; 3) self-regulation, defined as monitoring of behavior to match goals and effective strategies to overcome barriers; 4) resources, defined as psychological and physical assets (e.g., skills or equipment) leveraged to facilitate behavior; and 5) contextual influences, defined as social and physical environments supportive of behavior.¹⁸ Our conceptual model integrates this behavioral maintenance framework with the minimally disruptive medicine paradigm and proposes that minimally disruptive treatments for WLM that target maintenance-relevant domains may favorably shift the burden-capacity balance to promote sustained behavioral adherence and WLM (Figure 1).

Figure 1.

Conceptual model of sustained adherence and WLM. The model proposes that minimally disruptive intervention strategies improve long-term adherence by reducing treatment burden and enhancing capacity through effects on maintenance motives, self-regulation, habits, contextual influences, and resources, resulting in improved weight loss maintenance.

2.2. Overview of Study Design

The OPT-X trial is delivered in a hybrid format and includes a 4-month weight loss induction phase followed by a 12-month randomized extended care program. Participants achieving ≥5% weight loss during the induction phase will be randomized into the extended care program, which includes four intervention components (RFV, home-based RT, buddy training and support, and ACT workshops) (Supplemental Figure 1).^{16–18,20,21} To enhance study efficiency given multiple intervention components, this study utilizes a multiphase optimization strategy (MOST) framework with a factorial experiment.¹⁹ Each candidate component has two levels (yes/no) resulting in 16 intervention conditions (Table 1). With 272 randomized participants allocated equally across 16 experimental conditions, approximately 17 participants are expected per condition. Participants will be enrolled in four sequential recruitment cohorts (approximately 68 randomized participants per cohort) to facilitate scheduling and delivery of group-based intervention activities. Within each cohort, participants who meet eligibility criteria for the extended care phase will be randomized to one of the 16 experimental conditions defined by the 2x2x2x2 factorial design. Group-based activities during the extended care phase (i.e., ACT and RT workshops) will include 8-12 participants per workshop. Optimization for this study prioritizes identifying an intervention that achieves clinically meaningful WLM with lower-burden components. Data are collected at four timepoints, including at baseline [4 months prior to randomization (i.e., Month -4)], at randomization (Month 0), and follow up (Months 6 and 12).

Table 1.

Conditions for OPT-X factorial experiment

	RFV	Home-based RT	Buddy Training & Support	ACT Workshops
1	Yes	No	No	Yes
2	Yes	No	Yes	Yes
3	Yes	Yes	No	Yes
4	Yes	Yes	Yes	Yes
5	Yes	No	No	No
6	Yes	No	Yes	No
7	Yes	Yes	No	No
8	Yes	Yes	Yes	No
9	No	No	No	Yes
10	No	No	Yes	Yes
11	No	Yes	No	Yes
12	No	Yes	Yes	Yes
13	No	No	No	No
14	No	No	Yes	No
15	No	Yes	No	No
16	No	Yes	Yes	No

RFV: reduced food variety; RT: resistance training; ACT: acceptance and commitment therapy

2.3. Study aims

This study has three aims:

1. Identify active intervention components that promote WLM and assemble an optimized intervention package that maximizes WLM.

2. Examine whether intervention-related improvements in habit strength, psychological resources, social support, and motivation explain (mediate) the effects of intervention components on WLM.

3. Explore whether participant characteristics, including age, race, and magnitude of initial weight loss, influence (moderate) the effectiveness of intervention components.

2.4. Study population

The study will be conducted at the University of Alabama at Birmingham in the United States. Participants are community-dwelling adults aged ≥18 years with obesity (i.e., body mass index ≥30.0 kg/m²). Table 2 summarizes all eligibility criteria. Participants will be primarily identified through the Informatics for Integrating Biology and Bedside (i2b2) research repository, which is used to identify potentially eligible individuals from electronic health record data.²² Contact with prospective participants will occur only through Institutional Review Board (IRB)-approved recruitment procedures designed to protect patient privacy. Additional recruitment strategies include institutional research registries, social media advertisements, and community outreach. Use of multiple recruitment approaches is intended to enhance participant representativeness. Those achieving a pre-specified weight loss criterion of ≥5% will be randomized to the extended care program (Supplemental Figure 1). Participants will receive compensation for completion of assessment visits. The protocol was approved by the University of Alabama at Birmingham IRB, and all participants provide written informed consent before study procedures. Potential risks include exercise-related injury, psychological discomfort during behavioral discussions, and loss of confidentiality; potential benefits include access to an evidence-based weight management program. Adverse events will be monitored throughout the study according to institutional requirements.

Table 2.

Participant eligibility criteria

Category	Criterion
Inclusion criteria	Age ≥ 18 years
	Body mass index ≥ 30 kg/m2
	Regular access to internet (to access Zoom videoconferencing)
	Willing to obtain medical clearance for exercise prior to enrollment (if indicated)
Exclusion criteria	Current or recent (past 6 months) use of prescription weight loss medications
	Weight loss > 10 pounds in the past 6 months (other than postpartum)
	Pregnancy or anticipated pregnancy
	Household member already participating in the study
	Current participation in another randomized research project
	Uncontrolled hypertension, defined as systolic/diastolic blood pressure > 160/100 mm Hg at screening
	Any of the following medical conditions:  Myocardial infarction or cerebrovascular accident in the past 6 months  Unstable angina in the past 6 months  NYHA Class III or IV congestive heart failure  Type 1 diabetes  Chronic lung diseases that limit physical activity  Advanced kidney disease requiring dialysis and/or special diet
	Inability to identify an adult friend or family member for extended care intervention phase
	Residing > 50 miles from the assessment visit location (University of Alabama at Birmingham, Birmingham, AL, USA)
	Planning to relocate out of the area in the next 18 months
	Inability to read or understand English
	Unwilling or unable to give informed consent or accept randomization assignment

2.5. Intervention

2.5.1. Initial weight loss intervention

Participants will complete an initial 4-month weight loss program consisting of 16 weekly, group-based sessions (60 minutes) supplemented with 1:1 contact (e.g., video call, phone call, or chat messaging) with trained interventionists. Healthie, a HIPAA-compliant digital platform accessible via a web browser or the accompanying mobile app, serves as the remote modality for the delivery of the intervention (Supplemental Figure 2; platform demo available at https://www.gethealthie.com/platform-overview). Session content is based on modified versions of the lifestyle interventions of the Diabetes Prevention Program²³ and Look AHEAD trial²⁴ used in previous protocols.^25–29 Initial sessions will focus on dietary modification, increasing moderateintensity PA (>200 min per week), and behavioral strategies such as goal-setting, self-monitoring, and problem-solving.^1,3,6,24 Participants will be encouraged to use Healthie to self-monitor dietary and PA behaviors. All participants receive a digital weight scale and food scale to assist with self-monitoring. Participants are informed that achieving ≥5% weight loss is a prerequisite for continuation into the extended care program but are encouraged to work toward a 10% weight loss goal.

2.5.2. Study personnel and interventionists

Intervention content will be delivered by trained interventionists with backgrounds in psychology, nutrition, exercise science, public health, or related health professions. Interventionists may include registered dietitians, exercise specialists, and other personnel with experience in behavioral weight management. All interventionists will receive standardized training in study procedures, intervention delivery, motivational communication strategies, and treatment fidelity monitoring before participant contact. To promote consistency across treatment conditions, interventionists will deliver both the core weight loss maintenance program and the intervention components to which participants are assigned using a structured facilitator guide that provides standardized content, procedures, and delivery recommendations. Ongoing supervision, protocol review, and fidelity monitoring procedures will be used throughout the trial.

2.5.3. Randomization

Participants meeting eligibility criteria for the extended care program will receive the core program and be randomized via computer-generated assignment to receive 0-4 of the ‘minimally disruptive’ intervention components, reflecting 16 intervention combinations. A permuted block randomization with varying block sizes blinding the study investigators to forthcoming treatment assignments will be applied with equal probability of allocation across the 16 conditions. Because of the behavioral nature of the intervention components, interventionists and participants cannot be blinded to component assignment. However, research staff conducting outcome assessments and primary analyses will remain blinded to treatment allocation.

2.5.4. Extended Care Intervention

Core Program.

All participants will receive core program content targeting dietary, PA, and behavioral topics to support WLM through sustainment of self-regulation skills. Participants will continue behavioral goals and self-monitoring practices introduced during the initial weight loss program, with modifications as needed to support WLM. Core intervention content will be covered during 24 brief 1:1 contacts (≤30 minutes) delivered remotely (via Healthie) that decrease in frequency over time (i.e., 12 weekly for 3 months, 6 biweekly for 3 months, and 6 monthly for 6 months). Individual contacts will be scheduled collaboratively by interventionists and participants and will generally occur at the same day and time each week throughout the program, with adjustments made as needed to accommodate participant availability and scheduling preferences. Individual sessions will be conducted primarily via video call through Healthie, although telephone contact or secure messaging may be used when participants or interventionists are unable to meet by video. Individual contacts focus on progress review, problem solving, reinforcement of behavioral strategies, and support for assigned intervention components. The core maintenance intervention is delivered individually; however, two intervention components described below (home-based RT and ACT workshops) also include group-based workshops that provide skills training, experiential activities, and intervention-specific content.

Extended Care Components.

Candidate intervention components were selected based on three criteria: (1) empirical evidence supporting weight management or weight loss maintenance (Table 3), (2) alignment with theoretical domains of behavior maintenance included in our conceptual model,¹⁸ and (3) potential to be delivered in a lower-burden manner consistent with principles of minimally disruptive medicine.

Table 3.

Weight loss maintenance (WLM) strategies to enhance capacity with less burden

WLM strategy	Description	Empirical Support	Primary Maintenance Construct Targeted
Reduced food variety (RFV)	▪ Limiting intake of high-energy, low-nutrient dense foods by consuming a limited variety of these foods (e.g., repetition of the same entrée and snack food several times per week)	Less variety in eating occasions reduces energy intake, and limited variety in energy-dense foods produces greater initial weight loss and is a common characteristic of successful weight loss maintainers.30–36	Habits
Home-based resistance training (RT)	▪ A complementary strategy to support PA during WLM and includes the provision of short-term RT skills and self-regulatory training, home equipment and low dose RT prescriptions	RT is common among PA and WLM maintainers; promotes comparable weight loss and WLM to aerobic exercise when paired with caloric restriction; and is associated with improved eating, PA behaviors, and psychosocial health.37–46	Resources
Buddy training and support	▪ Leveraging individuals’ existing social networks by providing training on effective techniques (e.g., reflective listening) for support and encouragement	Identifying and training buddies from one’s social network supports WLM and may reduce dropout.11,47	Contextual influences
Acceptance commitment therapy (ACT) workshops	▪ Short-term ACT training condensed for briefer delivery to strengthen acceptance, values, and mindfulness skills	Limited contact ACT workshops are effective for initial weight loss and WLM.48	Maintenance motives

RFV: reduced food variety; RT: resistance training; ACT: acceptance and commitment therapy; PA: physical activity; WLM: weight loss maintenance

RFV.

RFV will enhance capacity via habit formation (i.e., increased automaticity and reduced effort in deciding among food options) to consume meals and snacks that promote decreased energy intake.^32,49,50 At the start of each month, participants will identify two foods (i.e., one dinner entrée and one snack food) to regularly consume during the month. Entrées will be classified as those foods providing the largest amount of energy at a meal (e.g., pizza, pasta, meat). Snack foods will be classified as foods that are commonly eaten, but not as an entrée (e.g., baked goods, popcorn, desserts, chips). The goal for this condition is to consume their chosen entree (≥2 times per week) and snack food (≥4 times per week), while limiting variety of other foods in these groups. Participants will develop meal plans specifying when chosen foods will be consumed (e.g., “Taco Tuesdays and Fridays”). Participants will be encouraged to repeat meal plans over the course of the intervention and to consume their chosen foods whenever they are in unplanned eating situations.

Home-based RT.

Home-based RT will enhance WLM capacity via provision and development of resources that promote increased PA and psychosocial characteristics (e.g., self-efficacy, self-regulatory skills, mood). Participants are encouraged to achieve recommended levels of RT (≥2 days per week) to support weight loss and WLM.^51,52 Participants will attend six 1-hour workshops delivered during the first 3 months of the WLM phase before transitioning to an unsupervised home-based RT program for the remainder of the intervention. Workshops are designed to address common barriers to RT while providing skills training and short-term supervision.^53,54 Between workshops, participants are encouraged to practice RT on 2 days each week. Participants receive RT equipment, supporting materials, and low dose RT prescriptions to support behavioral enactment. Each RT session will target major muscle groups using a full-body training approach. Low-dose RT prescription, defined as the completion of at least one set per exercise with additional sets performed when feasible, are implemented to enhance feasibility and adherence while remaining consistent with established guidelines.^52,55 The RT program will be progressed using a pre-specified algorithm.

Buddy training and support.

Buddy training and support will enhance WLM capacity by promoting contextual influences that support WLM behaviors.^47,56 Participants will be asked to choose an adult friend or family member who can support their weight loss efforts throughout the trial. Buddies will receive structured training on the provision of effective social support, including completion of ≥4 of 6 pre-recorded online training webinars. Webinars will cover energy balance, effective social support strategies, motivation, problem solving, healthy eating, and support during setbacks. After training, buddies will be asked to check in with their participant 1-2 times per week during the 12-month intervention via a modality of their mutual choosing (i.e., in-person, phone, text, email, social media).

ACT workshops.

ACT workshops will enhance capacity by strengthening maintenance motives (e.g., congruence between behaviors and identity/values) to sustain WLM behaviors. Workshops will include six 1-hour group meetings held remotely via Zoom every 2 weeks during the first 3 months of the WLM phase by interventionists with training in ACT. The workshop will teach mindfulness, willingness, and values clarification skills, with materials adapted from treatment protocols that have been successfully used by our team. Modules focus on mindfulness, acceptance of difficult internal experiences, and clarification of personal values to support sustained engagement in weight control behaviors.

2.5.5. Measures

A schedule of study measures is provided in Table 4. Demographics, medical history, and medical clearance for exercise (if indicated) will be self-reported. Anthropometric measures will include height (baseline only) and weight. The primary outcome is WLM at Month 12. This will be operationalized primarily as percent change in body weight from randomization to Month 12. Treatment adherence during the weight loss induction and extended care phases will be assessed by the treatment sessions attendance. Adherence to the four extended care components will be assessed by brief component-specific weekly surveys distributed through Healthie. Adherence measures are defined as follows: (1) number of times the chosen entrees and snacks were consumed per week (RFV), (2) number of skills training workshops attended and the frequency of RT per week (home-based RT), (3) number of buddy support contacts reported per week (buddy training and support), and (4) number of group-based workshops attended and reports of ACT skills utilization per week (ACT workshops). All participants complete the same outcome assessments and measures of the hypothesized mediators regardless of randomized treatment assignment. Component-specific adherence measures are collected only from participants assigned to the corresponding intervention component because they assess engagement with intervention content unique to that component. Treatment burden for each intervention component will be assessed longitudinally using participant-reported ratings of time and effort associated with intervention participation using an adapted version of an established treatment burden measure.⁵⁷ Treatment mediators consistent with our conceptual model and specific to each treatment component will also be assessed, including habit strength (Self-Report Habit Index^50,58), psychological resources (Exercise Self-Efficacy Scale^59–61), social support (Social Support Scales for diet and exercise⁶²), and motivation and values (Treatment Self-Regulation Questionnaire⁶³). Additional constructs relevant to the core WLM program and intervention components will also be assessed, including self-regulation skills (Self-Regulation Questionnaire – short version^64,65); mindful awareness of eating and activity experiences (Philadelphia Mindfulness subscale⁶⁶); acceptance, or willingness to engage in weight loss behaviors (Food Craving Acceptance and Action Questionnaire⁶⁷); and physical resources (exercise environment questionnaire⁶⁸).

Table 4.

Schedule of study measures

Measures	Month - 4	Month 0	Month 6	Month 12
Demographics	X
Medical history	X
Medical clearance (if indicated)	X	X
Anthropometry	X	X	X	X
Treatment adherence
RFV: Weekly consumption of chosen entrees and snacks		X	X	X
Home-based RT: Workshop attendance & weekly RT engagement		X	X	X
Buddy training: Frequency of buddy contacts		X	X	X
ACT workshops: Workshop attendance & ACT skills utilization		X	X	X
Treatment burden		X	X	X
Treatment mediators
Habit Strength	X	X	X	X
Exercise Self-efficacy	X	X	X	X
Social support	X	X	X	X
Motivation and values	X	X	X	X
Other Measures
Self-regulation skills69,70	X	X	X	X
Mindfulness71	X	X	X	X
Acceptance72	X	X	X	X
Home Exercise Environment73	X	X	X	X

PA: physical activity; RFV: reduced food variety; RT: resistance training; ACT: acceptance and commitment therapy

3. Results

3.1. Aim 1

Aim 1: Identify active intervention components that promote WLM and assemble an optimized intervention package that maximizes WLM. We hypothesize that the optimized package will significantly contribute to WLM between randomization and month 12.

3.1.2. Statistical analysis plan for Aim 1

All statistical analyses will be conducted in an intent-to-treat manner. The primary outcome for all efficacy analyses is percent change in body weight from randomization to Month 12.

In primary analyses, we will use a linear mixed model approach with effect coding.^47,56 First, we will estimate the main effects of the four individual intervention components (i.e., factors) by means of a full factorial experiment. A main effect will be defined as the difference between the “on” (included) versus “off” (excluded) level of that factor, averaged across all other factors. The latter will be achieved by effect coding the study measures as 1 (factor set to “on”), and −1 (“off”). For instance, the main effect for RFV will be the difference between its levels averaged across each of the two levels of home-based RT, buddy training and support, and ACT workshops. For each of the components, we will determine whether there is a difference in change across time using Month 0 (randomization) as the reference cell (i.e., Month 12 vs. Month 0). Statistically these effects will be modeled as component by time interactions, with the Month 12 outcome as the primary endpoint. Next, we will assess interactions (i.e., two-, three-, and four-way interactions) between components (e.g., two-way interactions of RFV x home-based RT x time). Significant interactions will be interpreted alongside identified active components.

Decision-making process.

In brief, the goal is to identify intervention combinations for an optimized treatment package.⁶⁹ Accordingly, potential best combinations of components with individual components being ‘robust’ to produce the desired effect will be identified based on the hierarchical ordering principle. Specifically, the change in estimates for each main effect (component x time interaction) will first be assessed and those that produce a significant effect for weight change when turned “on” will be identified as ‘active’ (i.e., screened-in set) components. Next, interactions that include active individual components will be assessed (e.g., synergistic, antagonistic) to determine whether components remain in the ‘screened-in’ set (see Supplemental Figure 3 for an example of the decision-making process for a two-way interaction). Once the screened-in components are identified, they will be ranked in order of maximal WLM to form the final optimized intervention. Because all the intervention components were designed to be lower-burden and less intensive than many traditional approaches, our default decision-making is informed by the maximum efficacy (i.e., magnitude of WLM) of the intervention package.

3.2. Aims 2 and 3

Aim 2: Examine whether intervention-related improvements in habit strength, psychological resources, social support, and motivation explain (mediate) the effects of intervention components on WLM.

Aim 3: Explore whether participant characteristics, including age, race, and magnitude of initial weight loss, influence (moderate) the effectiveness of intervention components.

3.2.1. Statistical analysis plan for Aims 2 and 3

We will analyze mediation and moderator models separately and then analyze their joint effects.⁷⁰ First, we will explore the role of constructs from our conceptual model as mediating variables between the intervention and primary outcome. Pre-specified mediators include measures of habit strength, resources, social support, and motivation. We will use the product of coefficients test to compute the mediated effects.^71,72 Specific moderators will include age, race, and magnitude of initial weight loss achieved during the induction phase. Magnitude of initial weight loss will be examined as a prespecified moderator of intervention effects on weight loss maintenance and may also be included as a covariate in sensitivity analyses. Moderation effects will be tested with multiple regression analysis, where all predictor variables and their interaction terms will be centered prior to model estimation to improve interpretation of regression coefficients. Our model will be similar to the one in our primary analysis, except we will augment the model with an additional two-way interaction term between moderator (e.g., race, age, initial weight loss) and the intervention component (moderator x component) and a three-way interaction term (moderator x component x time). If three-way interactions are observed, post-hoc tests will be carried out to determine the differential influence of the moderator to one treatment versus another, controlling for family-wise error rate. We will explore the use of modeling approaches such as shrinkage-based predictive risk models and classification and regression trees to evaluate heterogeneity in treatment effects as additional secondary analyses.

3.3. Sample size and power

Power calculations for Aim 1 were conducted using the SAS macro %FactorialPowerPlan that is specific for MOST designs,⁷³ assuming an alpha level of 0.05 and power of 0.80. We powered to estimate a small-to-moderate effect (Cohen’s d=0.30) and pre-post correlation of 0.6.^74,75 This corresponds to a component main effect of 2% difference in weight change between conditions with the component turned “on” versus conditions with the component turned “off”. This is consistent with effect sizes and power calculations used in similar WLM trials and represents a conservative estimate since a weight change no greater than 3% has been recommended to define weight maintenance.^76–78 Further, since these power calculations are based on effect coding rather than dummy coding, we have the same power to detect interactions as we do main effects. Accounting for 15% attrition (based on our previous weight management trials that achieved retention >85%) resulted in a randomized sample size of 260 participants (226 completers). To allocate equal numbers of participants across the 16 conditions, we increased the final sample to N=272. For Aim 2, a sample size of 226 achieves 80% joint power to detect the regression coefficient of 0.3 for the model with a given mediator as the outcome and the study components as the exposure; and regression coefficient of 0.2 for the model with weight loss difference as the outcome and the study components as the exposure given the mediator is zero. Based on our previous trials we estimated that 50% of participants would achieve sufficient weight loss to be eligible for randomization and, as such, we plan to recruit and enroll 544 participants at baseline.

4. Conclusion

Behavioral treatments for obesity are effective but difficult to sustain long-term, making weight regain common. This optimization trial provides an efficient and rigorous evaluation of strategies to promote WLM, intended to result in an optimized intervention package that improves WLM by enhancing individuals’ capacity for behavior maintenance while limiting treatment burden. To our knowledge, this is the first use of a MOST-inspired optimization trial to develop an intervention package specifically focused on WLM. Further, the study explores unconventional but evidence-based treatment strategies for WLM and applies a ‘minimally disruptive medicine’ paradigm (i.e., care designed to advance patient priorities while minimizing treatment burden)^16,17 to behavioral maintenance in an integrated conceptual framework. The trial will also examine treatment moderators that may provide valuable information on future treatment refinements and tailoring.

This project will provide 12 months of WLM data. While longer-term follow-up is desirable, the most rapid and significant weight regain occurs during the first year following treatment,⁷ suggesting this is a critical period to study WLM. The decision-making process for this optimization trial prioritizes identification of the package that maximizes WLM without formal specification of key constraints (e.g., cost, time). This is because all components were developed and selected with reduced burden in mind. Depending on context, alternate decision criteria can be considered. The same interventionists will deliver all component combinations to the participants, minimizing potential “interventionist effects” and ensuring continuity of care for participants across time and conditions. Adherence to several intervention components is assessed primarily through participant self-report because many targeted behaviors (e.g., social support interactions and food selection patterns) are difficult to capture objectively. Although objective monitoring devices were considered, they were not included because of resource constraints and the study’s primary focus on optimization of intervention components based on weight loss maintenance outcomes.

Upon completion of the optimization trial, activities for an evaluation phase will constitute a future proposal in which the optimized package will be evaluated to confirm its effectiveness via a conventional RCT (i.e., optimized treatment package vs. standard care). While it is hypothesized that the optimized package will result in significantly more weight loss than standard care, nonsignificant findings would also be informative, indicating a need to return to the MOST preparation phase, in which the conceptual model can be re-evaluated, additional or revised intervention components can be identified, and additional pilot work can be conducted. Future evaluation studies should assess implementation costs, reimbursement potential, and cost-effectiveness of the optimized intervention package. The overarching goal of the study is to improve long-term weight outcomes through WLM intervention strategies that are feasible and potentially sustainable for individuals. Ultimately, this work may inform best practices for WLM interventions and improve long-term weight management outcomes.