Addressing Rural Health Disparities by Optimizing “High-Touch” Intervention Components in Digital Obesity Treatment: the iREACH Rural Study

Abstract

Background:

Rural residents are more impacted by obesity and related comorbidities than their urban counterparts. Digital weight management interventions may produce meaningful weight loss among rural residents.

Objectives:

The iREACH Rural Study aims to identify “high-touch” component(s) that contribute to meaningful weight loss (≥1.5 kg) at 6-months, over and above what the 24-week core online program produces. Three treatment components are assessed: group video sessions (yes/no); self-monitoring feedback (counselor-crafted/pre-scripted, modular); and individual coaching calls (yes/no).

Design:

The iREACH Rural Study is a factorial experiment (n=616).

Methods:

Participants receive up to 3 “high-touch” components (weekly synchronous facilitated group video sessions, weekly counselor-crafted self-monitoring feedback, and individual coaching calls) to determine which contribute meaningfully to 6-month weight loss. Participants complete assessments at baseline, 2 months, 6 months, and 12 months. Weight loss at 6 months (primary outcome) and 12 months (secondary outcome) is measured by Bluetooth-enabled scales. The study seeks to identify the weight loss approach for underserved rural residents which optimizes weight change outcomes and also examines costs associated with delivering different treatment constellations.

Summary:

The iREACH Rural Study is the first of its kind to isolate digital weight loss intervention components to determine which meaningfully contribute to long-term weight loss among rural residing individuals. The results may be used to refine digital weight loss programs by enhancing their effectiveness to allow broad dissemination.

1. Introduction

Over 70% of adults in the United States have overweight or obesity,¹ which increases risk of chronic conditions including cardiovascular disease, diabetes, and certain forms of cancer.^2–4 Weight losses of as little as 5% attenuate obesity-associated comorbidities,^5,6 and are achieved with lifestyle interventions like the federally-supported Diabetes Prevention Program (DPP).^7,8 Despite the success of the DPP, program sites are not equitably distributed: to date, 14% of rural counties have a DPP program site compared to 48% of urban counties.⁹ As rural residents are often more geographically dispersed, digital weight loss programs offer promise and have resulted in average weight losses ≥5% at 6 months.^10–12 Importantly, “high-touch”, personnel-intensive elements associated with in-person delivery, such as individual coaching sessions, appear to be critical to remote program success^13,14 but increase costs.^15,16 Therefore, there is need to establish which “high-touch” components are most impactful on weight loss in remote interventions and thus maximize program effectiveness and minimize costs. The purpose of this manuscript is to describe the development and methodology of a digitally-delivered weight loss study for rural adults (iREACH Rural). The study aims to optimize weight loss by using a factorial experiment to efficiently identify “high-touch” component(s) which make more than a minimal contribution (≥1.5 kg) to weight loss above what the core digital program produces.

2. Methods

2.1. Study overview

The study examines three distinct personnel-intensive digital treatment components: 1) weekly synchronous facilitated group video sessions; 2) weekly counselor-crafted self-monitoring feedback; and 3) individual coaching calls (Figure 1). These three candidate treatment elements all require significant personnel time and cost,^16,17 and it is unclear if they are necessary for optimal weight loss. All other elements of the core iREACH program are comparable across all participants, allowing the presence of the “high-touch” digital components to be the only thing differing between conditions. The study is approved by the Institutional Review Board (IRB) at the University of South Carolina and acknowledged by the University of Virginia’s IRB.

Figure. 1.

Conceptual Model.

2.2. Factorial experiment design

We use a 2 × 2 × 2 factorial experimental design within the Multiphase Optimization Strategy (MOST) framework¹⁸ (Table 1) to examine the effects of three “high-touch” intervention components. This framework was selected as it allows for direct comparisons of intervention components (in contrast to a randomized controlled trial, which often examine intervention “packages”). We will enroll participants in seven waves over 36 months, with approximately 88 randomized participants anticipated in each wave.

Table 1.

Factorial Design

Experimental Condition (Cell)	Facilitated Weekly Group Sessions	Type of Weekly Feedback	Individual Coaching Calls	Anticipated N per Cell
1	Group	Counselor	Calls	77
2	Group	Pre-Scripted	No Calls	77
3	Group	Pre-Scripted	Calls	77
4	Group	Counselor	No Calls	77
5	No Group	Counselor	Calls	77
6	No Group	Counselor	No Calls	77
7	No Group	Pre-Scripted	Calls	77
8	No Group	Pre-Scripted	No Calls	77

2.3. Participant recruitment, selection, and randomization

Adult participants (i.e., 18 years or older) are recruited nationally from rural areas (defined as having a zip code corresponding to a Rural-Urban Commuting Area code of 4–10, defined as rural by the Federal Office of Rural Health Policy, or living in an area that qualifies for the Centers for Medicare and Medicaid Services Rural Health Clinic Program based on 2010/2020 Census Bureau data).¹⁹ Our recruitment approach uses multiple methods to publicize participation opportunities to likely eligible individuals, including: 1) ResearchMatch; 2) boosted Facebook posts; 3) ClinicalTrials.gov; 4) postings in relevant newsletters and listservs (e.g., sororities, fraternities, churches, organizations consisting predominately of men like hunting clubs); 5) press releases to state news organizations, and 6) referrals of friends/family members/co-workers from former program participants.

Interested individuals from across the United States are directed to an online portal which provides a narrated presentation describing the study and answers common questions about study participation. Individuals who are interested in participating are directed to an online screening application which obtains basic information to determine likely eligibility. Individuals who appear likely to be eligible after completing the screener are contacted for a Zoom screening visit, in which additional eligibility criteria are confirmed, a detailed description of study participation is provided, and any questions are answered. The consent form is also reviewed during the screening visit; those who wish to pursue enrollment are sent a link to the consent within REDCap.^20,21 Once the consent form has been signed, potential participants receive a secure, personal link to the baseline REDCap questionnaires. After these questionnaires are completed, individuals are mailed a Bluetooth-enabled Fitbit Aria electronic scale and Fitbit activity tracker, and they are asked to monitor their weight, physical activity and dietary intake for 7 days as a behavioral run-in. If potential participants complete all of these steps and provide an additional weight on the specified randomization weigh day (baseline weight), they are randomized.

We aim to enroll 616 rural-dwelling adults, with specific recruitment targets of 40% men and 22% who self-identify with an National Institutes of Health-designated ethnic or racial minority group, reflecting the proportion of self-identified racially minoritized individuals who reside in rural areas.²² Individuals are eligible to participate if they have a Body Mass Index (BMI) between 25–55 kg/m2, are free of medical problems that might contraindicate participation in a behavioral weight loss program containing an exercise component, not pregnant or lactating, not currently on medication to promote weight loss, not enrolled in another weight loss program, and have no history of bariatric surgery. Participants must have a computer or tablet with Internet access (at home or work) and a smartphone to access all program elements.

Participants are randomized individually to one of the eight treatment conditions using the randomization scheme built into the study’s REDCap database by a biostatistician with no direct participant contact. We use sex as a stratification factor. Randomization assignment is coded in the study management system.

2.4. Intervention components

Core intervention.

All participants receive the core iREACH behavioral lifestyle program, as described previously.^10–12 Briefly, the 24-session group-based program delivers weekly online interactive behavioral modules, with a goal of fostering 5–10% weight losses and changing dietary and exercise habits with goal-setting, self-monitoring, and problem solving strategies. The intervention is based on social cognitive theory²³ and uses a self-regulation approach to produce and maintain weight loss.²⁴ The program focuses on skill development and offers asynchronous social support through a private curated and moderated discussion board.

Intervention content is tailored to unique experiences and needs of rural populations using cultural adaptation principles.²⁵ For example, physical activity options recognize that gym-based opportunities and availability of walking trails could be limited in rural areas and other more feasible exercise possibilities are suggested. Infrequent shopping due to long distances to grocery stores is acknowledged, and suggestions are provided for how best to stock up to keep healthy foods available in the home, as well as suggestions about shopping on a limited budget. Whenever possible, we provide module content that reflects likely lived experiences of rural residents; for example, we mention research suggesting that social support for dietary changes and physical activity may be lower among rural residents²⁶ and then suggest ways of cultivating social support. Adaptations focus predominantly on the examples offered (and associated recommendations provided) rather than altering core behavioral strategies provided.

All participants receive calorie goals tailored to baseline weight (ranging from 1200–1800 kcal per day) and physical activity goals (graded to reach 250 Active Zone Minutes per week). All participants are asked to self-monitor their dietary intake daily using the Fitbit app on their smartphone. Participants are also given a Fitbit device to track physical activity daily. Daily weighing is prescribed, and an Aria Bluetooth-enabled electronic scale (e-scale) is provided. Data are transmitted to the password-protected study website via a Fitbit API and are accessible only to the participant and study staff. All participants are provided with weekly feedback emails focused on their self-monitoring and other weight control behaviors to emphasize successful enactment of self-regulatory behaviors and provides constructive suggestions.

In addition to the core digital program, participants may be randomized to receive additional “high-touch” treatment components: 1) synchronous facilitated group video sessions (yes/no); 2) counselor-crafted self-monitoring feedback or pre-scripted, modular feedback; 3) individual coaching calls (yes/no). The three components are described in detail below.

Synchronous facilitated group video sessions.

Facilitated group video sessions were selected as a component because group meetings have been associated with superior weight losses among both in-person²⁷ and digital programs,²⁸ likely reflecting the importance of social support for weight loss success.^29,30 Furthermore, online group sessions may help alleviate noted deficiencies in social support for weight management among rural populations specifically.^26,31,32Participants randomized to receive weekly synchronous (i.e., in real time) video group sessions (50% of sample; blue cells in Table 1) attend weekly “face-to-face” virtual groups of 15–25 participants at a consistent time. These 60-minute Zoom³³ groups are facilitated by a trained professional who reinforces the information and behavioral strategies introduced in weekly modules. The session leader elicits participants’ experiences in their efforts to establish new diet and exercise habits, guides problem solving, and promotes social support within the group. The session protocol has a structured format: starting with a “check-in,” group members review their successes and challenges in meeting goals over the previous week and collectively problem-solve identified barriers. Then, the group discusses the self-regulation topic introduced in the weekly module, with an emphasis on experiential engagement around the topic and a goal of promoting behavioral skill development. The session ends with a review of next week’s behavioral goals. Participants randomized to a “no group” condition do not participate in any formal group sessions but have asynchronous social support through a private curated and moderated discussion board, which provides weekly prompts that seek to promote engagement around the information and strategies introduced in the weekly module.

Self-monitoring feedback.

Self-monitoring feedback was selected for examination because although individuals who self-monitor most frequently tend to lose the most weight,³⁴ efficacy of counselor-crafted feedback in encouraging self-monitoring and producing better weight loss compared to pre-scripted modular feedback in a remotely-delivered program is less well known.^28,35,36 For those randomized to receive counselor-crafted feedback (50% of the sample; orange cells in Table 1), a trained interventionist reviews digital dietary, physical activity, and weight monitoring for the week, completion of online modules, and composes an individualized feedback message. Emailed feedback provides positive reinforcement for successful goal achievement, identifies possible areas for improvement, and suggests possible strategies for identified challenges.³⁷ Consistent with previous research,²⁸ individuals randomized to pre-scripted, modular feedback receive weekly emailed feedback that is constructed from a bank of pre-scripted messages that align with success, partial success, or absence of self-monitoring within the following domains: dietary monitoring, physical activity monitoring, and self-weighing. Pre-scripted feedback is “modular” in that one of three options is selected electronically using an established algorithm (see Table 2). The feedback message also notes whether the weekly online module was viewed.

Table 2.

Modular Feedback Example

Focused Feedback on Week 8
Weight Monitoring	Dietary Intake	Physical Activity	Module	Opening	Closing
a = 6–7 days b = 1–5 days c = 0 days	Monitoring on a day = recorded > 3 eating occasions Stayed in calorie goal = add up calories of days monitoring/#days monitored & if ≤calorie goal = success a= monitored ≥ 5 days AND stayed within calorie goal ≥ 5 days b= monitored ≥ 5 days but exceeded calorie goal on AND stayed within calorie goal c= monitored 1– 4 days d= no dietary monitoring	A = met Active Zone Minute goal of 250 min B = has some active minutes but not at goal (their active zone minute between 0 but ≤ 150 min) C= no active minutes D= no Fitbit activity data	A = opened module 8 B = did not open module 8
a) Strong finish to Month 2 of the program! Weighing yourself every day this week is the way to do it! Congrats! b) Problem solving is this week's theme so try out your problem-solving skills with a focus on weighing yourself every morning. It’s a gift you can give yourself - - those who weigh daily are more likely to succeed at weight loss! c) No weight values in your record this week? We definitely need you to hop on the scale for your Month 2 Weigh Day! Please get in touch if you are experiencing any technical difficulties with weighing yourself. We will work with you to fix them. Problem Solving is this week's theme and we can work it out together!	a) Great self-monitoring this week. Excellent!! b) Good self-monitoring this week. Use those problem-solving skills from this week's module to modify your calorie intake to stay within your goals. c) You've got some partial self-monitoring this week. Use those problem-solving skills from this week's module to increase the completeness of your recording. d) No dietary recording this week? Use those problem-solving skills from this week's module to start recording. Start with dinners and evening snacks - or other eating times that tend to be high calorie.	a) You did it! You aced the week 8 goal of 250 Active Zone Minutes this week. Well done!! b) Good work with your physical activity this week - you're moving towards the Active Zone Minute goal of 250 for week 8. Use those problem-solving skills from this week's module to pump things up. c) Your exercise could use some attention. Try applying those problem-solving skills from the week 8 module to get in your Active Zone Minutes. Start small if you've had a lapse. No need to jump back in at the current week's goal if you need to work up to it. Take it at your own pace if you need to. d) No physical activity data for week 8?!. Use those problem-solving skills from this week's module to help get yourself wearing your Fitbit and transmitting the data to us. If you are having technology difficulties, let us know!	a) Nice work taking a look at the Problem-Solving module this week. Research shows that problem-solving skills predict long term weight loss! So, hone those skills!! b) Seems like you have not yet looked at the Problem-Solving module. Research shows that problem-solving skills are strongly related to success in weight loss. Makes sense, right?! There are a lot of challenges out there to solve. So, don’t miss out on this critical module.	Greetings! We have now had our first follow up data collection. Thanks to all who completed it right on time (you know who you are!) And if you are among those who have not yet finished sending us your official weight, exercise and dietary information, or completed your online questionnaires, it’s not too late. Do it now! There is a nice little incentive for those who complete everything. So, get it done and treat yourself to a little prize for all your hard work. You deserve it!	Have a great week! [facilitator name]

Individual coaching calls.

Finally, we chose to evaluate the impact of individual coaching calls because these sessions may enhance digital treatment outcomes.³⁸ Individuals randomized to receive the coaching condition (50% of sample; green cells in Table 1) are offered three 30-minute video/phone calls (based on their preference) from an experienced interventionist. The first call occurs prior to starting the core program and emphasizes amplifying motivation and promoting early treatment engagement by exploring personal reasons for weight loss and lifestyle change, and reinforcing and expanding change talk using a motivational interviewing style.³⁹ Subsequent calls use a motivational interviewing-informed approach to problem solve barriers to self-monitoring, particularly dietary self-monitoring because of the strong role it plays in promoting weight loss.³⁴ Drawing on self-regulatory theory⁴⁰ and the problem solving therapy framework,²⁴ participants are guided to identify points at which their self-monitoring “breaks down” and determine which steps in the problem solving progression (i.e., orientation, definition, brainstorming, selection/implementation and evaluation) might best be applied to enhance dietary self-monitoring. We employ an adaptation of “toolkits” outlined in the problem solving therapy handbook,⁴¹ which have been tailored to self-monitoring by the research team. The toolkit helps participants collaboratively develop an action plan that moves them closer to engaging in consistent dietary self-monitoring, while supporting participant autonomy in developing the plan. The second call occurs in week 3 or 4 and the third call in week 10 or 11. Coaches are guided by a semi-structured interview format in REDCap which they complete to provide context for future calls. Participants who are randomized to a “no coaching” condition do not participate in coaching calls and begin the intervention with an introductory email from their group facilitator and the first module in week 1.

2.5. Measures

Participant data are collected at baseline, 2 months, 6 months, and 12 months (unless otherwise noted), either through technology-based physical/behavioral measurements or by computer-administered REDCap questionnaires. Our conceptual model (Figure 1) outlines our primary and secondary outcomes, as well as hypothesized mediators and moderators.

2.5.1. Demographic measures

Participants complete a sociodemographic questionnaire (e.g., race, ethnicity, sex, gender, age) at screening. Home address provided at screening is used to classify participants with respect to rurality.

2.5.2. Primary outcome measure

Change in body weight at 6 months is the primary study outcome. Weight outcomes are obtained by e-scale provided to each participant and transmitted to the study via an API from Fitbit. Participants weigh themselves once on the designated data collection day first thing in the morning, after voiding and before dressing.⁴² Measurement concordance between e-scales and calibrated clinic scales has been previously reported and is strong, with correlations of 0.99, regardless of gender, BMI, race or age.⁴³

2.5.3. Treatment engagement mediators

Treatment engagement, operationalized as self-monitoring adherence and online module completion, is captured to assess mediation of 6-month weight loss (Figure 1).

Self-monitoring and online module completion.

We record self-monitoring for weighing, dietary intake, and physical activity during treatment from data transmitted by Fitbit. Dietary self-monitoring data are considered as present on each day on which a minimum of 800 calories were recorded.⁴⁴ A valid day of self-monitoring physical activity is defined as a day with > 10 hours of wear time, assessed as minutes of non-zero heart rate measurement captured by the Fitbit device.^45–48 Weight is considered monitored on days with a weight value transmitted from the study-provided e-scale.⁴⁴ Self-monitoring rates are calculated as number of days with self-monitoring data/number of days. Weekly module completion is also tracked to characterize engagement. We track attendance at weekly group video sessions and individual coaching calls completed for those who are randomized to receive those components. We also examine sustained engagement with self-monitoring tools for 6-months following the end of the formal intervention period (i.e., 12-month assessment).

Physical activity.

Self-reported minutes of moderate-to-vigorous physical activity (MVPA) is measured using the International Physical Activity Questionnaire (short form).⁴⁹ In addition, we examine Active Zone minutes captured by the Fitbit monitor over a 7-day wear time at baseline, 6 and 12 months, to objectively-quantify changes in physical activity associated with the experimental components.

2.5.4. Social cognitive mechanisms

Five social cognitive variables are measured to assess mediation of treatment engagement and subsequent 6-month weight loss (Figure 1): self-regulation, supportive accountability, problem solving, motivation, and social support for weight control behaviors.

Supportive accountability.

Supportive accountability may increase treatment adherence when a trustworthy and caring person with specific expertise is perceived to be monitoring one’s actions, and individuals feel an obligation to deliver on a commitment or explain reasons for failing to achieve an outcome. Two indices of supportive accountability (i.e., 8-item Support Accountability Inventory,⁵⁰ 10-item Accountability subscale of the Supportive Accountability Measure)⁵¹ have been examined in the context of weight management; we obtain both at 2 months, 6 months, and 12 months given the lack of clarity as to which is the timepoint most strongly associated with digital weight outcomes.

Problem solving.

The 25-item Problem-Solving Test⁵² is administered to capture problem solving skill at baseline and 6 months. The measure includes 2 scales of Effective Problem Solving (Positive Problem Orientation and Planful Problem Solving), with higher scores on these scales indicating greater positive problem-solving skill levels. The measure also includes 3 scales of Ineffective Problem Solving (Negative Problem Orientation, Impulsive/Careless and Avoidance), with higher scores indicating greater utilization of that ineffective approach to problem solving.

Motivation.

We measure motivational factors using the 12-item Treatment Self-Regulation Questionnaire.⁵³ This is a self-determination theory-based measure which assesses autonomous reasons for engaging in weight control efforts (personal reasons for change or motivations that reflect an internalized rationale for change) and controlled motivation (reasons for change that are externally imposed or arise from others). The measure has previously been used to explore obesity treatment response.^24,40 Higher scores indicate greater levels of autonomous and controlled motivation, respectively.

Social Support.

We measure two aspects of perceived social support. First, we measure group cohesion at 2 months and 6 months using the 20-item Group Cohesion Scale.⁵⁴ Social support for weight control behaviors in the past month is measured using the 36-item Social Support for Healthy Behaviors Scale with higher scores indicating greater support for weight control behaviors from family and friends.³⁰

2.5.5. Cost and treatment satisfaction

Cost.

Non-research-related costs (evaluated at competitive market rates) associated with delivering each treatment component are tracked from baseline to 6 months and will be assessed in constant dollars using the Consumer Price Index. These include the hourly wages (including fringe benefits) of the minimum required occupations for training/implementation, and equipment costs for e-scales and activity trackers.

Treatment satisfaction.

Participants are asked about the helpfulness of each core treatment component, how satisfied they are with the program overall, and how likely they would be to recommend the program to family/friends. Higher scores indicate greater satisfaction or perceived helpfulness. These data are collected at 2 months and 6 months.

2.6. Sample size justification

Our primary outcome on which the trial is powered is weight change from baseline to 6 months. Our interest is in identifying the treatment elements that make more than a minimal contribution to average weight loss, or conversely to identify those components that have minimal or no effect on weight loss. We define a more-than-minimal contribution as an effect size of ≥.25 and thus have powered the trial to detect effects of .25 or greater, which translates to a ≥1.5 kg difference in weight loss (using 6-month data from our previous study, with a pooled standard deviation of 6.0).¹¹ A sample size of 518 is needed to detect a main effect of ≥1.5 kg at 6-month with 80% power and 0.05 type I error. With attrition at 6-months estimated at 15%, we will recruit 616 participants (allowing equal cell sizes across the eight treatment combination cells).

2.7. Statistical analysis

Primary outcome analysis.

An intention-to-treat approach will be used to estimate whether treatment components result in meaningful and statistically significant weight loss at 6 months. Specifically, we will address missing 6-month weight data using selection models or multiple imputation.⁵⁵ To evaluate which “high-touch” intervention components contribute meaningfully to 6-month weight loss, a General Linear Mixed Model (GLMM), adjusted for each of the three components and all two-way interactions between the components, will be used. Interactions will allow us to examine how the presence of one component potentially interacts with the effect of another component. Additionally, several demographic and relevant biological variables will be included as a priori chosen control variables: sex,⁵⁶ race/ethnicity,⁵⁷ age,⁵⁸ and baseline weight, given their previously established relationship with weight loss success. The GLMM also provides a structured approach to model potential mechanisms behind missing data and to test for the presence of selection bias.⁵⁹

Analysis of optimal treatment package.

To identify an optimized treatment package, we will perform a mixed effects model analysis using effect coding, which will keep any covariances among main effects and interactions to a minimum. We will then employ a decision-making process in alignment with the MOST framework¹⁸ by examining main effects, and then considering interactions in subsequent steps, as follows: Step 1: Tentative identification of components that show a significant main effect on weight. If any of the 3 components have a significant main effect, they will tentatively be selected for inclusion in the treatment package. Step 2: Examine interactions for indications that we should consider revising the components included in the package because when the two components are both included, they boost or diminish weight losses relative to what is achieved with each single component alone. Step 3: Identify the treatment package that maximizes weight loss; if more than 1 package maximizes weight loss, we will consider both weight loss and cost in making the final determination of the optimal treatment package.

Mediation and moderation analysis.

The main outcome GLMM will be modified to explore potential mediating effects of variables (e.g., accountability, social support, problem solving, motivation, and self-regulatory processes, treatment engagement [self-monitoring, web utilization, physical activity, dietary intake]) on weight outcomes. It will also be modified to include interaction terms between components and potential moderators (e.g., sex, race/ethnicity, baseline weight, age). Exploratory findings will provide preliminary results to inform a future trial to examine the optimal treatment package within salient sociodemographic subgroups and will allow us to better understand the effects of the components.

Cost analysis.

We will conduct analyses from the provider’s perspective to assess the incremental costs and potential cost-savings associated with each component. We will also estimate incremental costs associated with each combination of components, costs per participant, and marginal cost per kgs weight loss at 6-months. Incremental cost-effectiveness ratios (ICERs) will be calculated for all possible optimized treatment packages compared with their next best nondominated alternative.^60,61 Economic evaluation analysis will take the factorial design into consideration.⁶² Specifically, all possible combinations of the components will be treated as mutually exclusive and will be evaluated simultaneously instead of pairwise. Uncertainty around the ICERs will be graphically illustrated using bootstrapping methods.^62,63 We will summarize the joint uncertainty using cost-effectiveness acceptability curves,⁶⁴ and will be used to show whether any combinations are cost-effective compared with the core program with any additional component, for a range of providers’ willingness-to-pay thresholds.

Exploratory 12-month outcomes.

We will extend the GLMM model to examine weight trajectories at 12-months. For each of the three components, we will explore post-treatment weight outcomes. This allows a preliminary examination of components that may optimize weight maintenance among rural residing individuals.

3. Summary

Rural residents are at an increased risk of obesity and obesity-related comorbidities relative to their urban counterparts,^65,66 but they are underserved by traditional lifestyle modification weight loss programs.⁹ Remotely delivered weight-loss interventions are a promising and potentially cost-effective means of reaching rural residents who are geographically dispersed. However, remotely-delivered programs are not currently reimbursed by the Centers for Medicare and Medicaid Services, due to insufficient evidence of effectiveness.⁶⁷ Admittedly, few studies of remote weight loss programs have focused on rural residents.^68–71 Of those available, most are small, of short duration, uncontrolled, and/or limited to a single sex/gender. The current study will redress this by enrolling a large national rural sample, providing an important opportunity to understand digital weight control at a large geographic scale and informing questions of broad dissemination.

Remote interventions have produced clinically meaningful weight loss among participants when “high-touch”, personalized intervention components are included in the intervention.^13,14 The addition of “high-touch” intervention components to remote weight loss programs increases their cost but may increase the effectiveness. This study isolates the impact of three labor-intensive and thus expensive “high-touch” components, allowing for the refinement of a digital weight loss intervention to essential and complementary elements. This distillation of remote weight loss programs to their most successful and cost-effective elements may in turn inform policy decisions about reimbursements for these services, therefore expanding their reach in the long-term. Thus, although the specific components tested are not novel, this systematic approach to identify the optimal digital treatment design is novel and fills an important gap in the literature.

In addition, it has been well-documented that individuals who participate in weight loss programs, delivered both in-person and remotely, are susceptible to long-term weight regain.⁷² There is little information about which personnel-intensive treatment components might curtail weight rebound after digital weight control programs end. Our study follows participants for 6 months post-treatment, allowing a glimpse into the maintenance of behavior change and weight following the end of a digital program. Thus, this study will offer innovative insights for obesity management that can direct future research. Taken as a whole, this study will set the stage for confirming the most promising digital intervention to reduce obesity rates among rural residents and provide essential evidence for informing policy decisions regarding dissemination without geographic borders.