Practical Awareness‐Based Strategies for Eating (PASE): A Pilot and Feasibility Randomized Trial

ABSTRACT

Background

Few prior interventions for obesity have focused on reward‐related eating. Researchers developed a mobile health mindfulness‐based intervention, Eat Right Now (ERN), for improving reward related eating; ERN has not yet been feasibility tested as a weight loss intervention.

Objective

To obtain 6‐month pilot and feasibility data in patients using the ERN intervention alone or with asynchronous coaching.

Methods

A 6‐month, two‐arm pilot and feasibility randomized trial was conducted involving 20 adults with a BMI ≥ 25 kg/m² recruited from Kaiser Permanente Washington. Participants were randomized to ERN alone or ERN plus individualized coaching. Weight loss was assessed via a Wi‐Fi‐enabled digital scale.

Results

Among the 20 randomized participants, 17 (85%) started the intervention and remained enrolled in six months. Among these, 82% (14/17) had weight data captured by a Wi‐Fi scale and 70.5% (12/17) completed follow‐up questionnaires at the 6‐month time point. The coached ERN participants had 4.3% weight loss (95% confidence interval (CI): 2.3%, 6.3%) and uncoached participants had 3.6% weight loss (95% CI: 1.2%, 5.9%). Participants in both approaches reported reductions in reward‐related eating.

Conclusions

This pilot randomized trial of ERN demonstrates the feasibility and acceptability of the intervention for the majority of both the coached and uncoached groups. Further efforts are needed to ensure high retention and data capture in future studies.

1. Introduction

Despite decades of research on the behavioral treatment of obesity, its prevalence continues to increase worldwide [1]. It remains unclear what specific behavior change approaches are most effective for promoting sustained weight loss. Historically, standard behavioral treatments for obesity have relied heavily on cognitive‐behavioral willpower‐based self‐regulation strategies (e.g., self‐monitoring, stimulus control, contingency management, and substitution strategies) yielding short‐term weight loss that begins to recede after just 6 months [2]. These cognitive behavioral strategies often fail to address the more automatic and reactive, reward‐related eating behaviors, which appear to have a major impact on the ability to use planned, reasoned behavioral pathways to sustain weight loss [3, 4, 5]. Failure to address reward‐related eating behaviors may set the stage for self‐regulation failures, leading to a cycle of disinhibited reward‐related eating, negative affective states, and long‐term weight regain [6, 7, 8]. Recent advances in the field of obesity pharmacotherapy and bariatric surgery have pointed to the importance of food reward [9, 10, 11], but to date, few behavioral interventions have sought to target these behaviors. Thus, a current major gap in the field of obesity treatment is the lack of scalable and effective behavioral interventions that address automatic pathways, including reward‐related eating, that govern behaviors.

Reward‐related eating behaviors are those that are driven by the automatic rewards or relieving aspects of consuming highly palatable food [12, 13, 14, 15]. These behaviors are often used unconsciously to alleviate negative affective states, such as stress, sadness, and boredom, and they manifest as cravings, pre‐occupation with food, lack of satiation, overeating, and loss of control eating [14, 16]. These behaviors are selected, learned, and reinforced by rewarding responses from the environment, a concept known as reinforcement learning [17]. Once the brain forms a connection between a behavior and a reward, it creates a powerful emotional memory that increases the probability of enacting reward‐yielding behavior in the future [18]. This is the process of food reward valuation, where the subjective hedonic processing in the orbitofrontal cortex of the brain assigns the expected reward value of food and eating behaviors [19]. Highly palatable foods (e.g., sweets) are well‐documented to stimulate dopamine release along the same pathway as substances of abuse [20, 21, 22, 23, 24, 25]. This in turn drives motivation to consume food [26, 27].

Repeatedly overconsuming sugar‐laden food can then condition individuals to expect pleasurable responses not only upon consuming highly palatable food, but also when observing environmental cues that they associate with the food. This can activate learned associations that trigger automatic, cue‐induced, reward‐related eating—in the absence of hunger [28, 29, 30, 31, 32]. Reward‐related eating behavior is congruent with the psychosomatic theory of “emotional eating” (eating in response to emotional triggers rather than feelings of physical hunger) as well as “disinhibited eating” (overeating in the presence of palatable foods or other stimuli, i.e., stress) [33, 34, 35, 36, 37]. The more one engages in these behaviors, for example, by experiencing a negative affective state (trigger), eating tasty foods (behavior), and receiving temporary relief (feeling better, being distracted from negative affect, etc.), the more reinforced and habitual these behaviors become. Each time one engages in these behaviors, they further obscure the ability to differentiate between homeostatic (true) hunger and reward‐related eating.

Mindfulness training may recalibrate food rewards by increasing awareness of habitual behaviors, consciously evaluating the reward value of eating patterns, and enabling strategies to respond to triggers without automatic reactions. This approach works by targeting reinforcement learning mechanisms, where habitual behaviors are driven by their perceived reward value. By fostering curiosity and present‐moment awareness, mindfulness disrupts automatic pathways, allowing individuals to observe the outcomes of their eating behaviors—including the negative aspects of overindulgence [13]. This facilitates recalibration of reward values in the orbitofrontal cortex and ventromedial prefrontal cortex, key regions implicated in reward processing [38]. Mindfulness also engages neurocognitive systems related to attentional control, emotion regulation, and self‐referential processing, enhancing the ability to identify triggers, such as stress or emotional distress, that lead to maladaptive eating [39]. Over time, this process shifts values toward healthier eating behaviors, making them more intrinsically rewarding and reducing reliance on willpower‐based self‐control, which often fails under stress [15]. Mindfulness training can help individuals identify mental states that trigger eating in the absence of hunger, the types of food they eat, and the way they eat. This awareness can help individuals to assess the relative reward of continuing to engage in their current habit or choosing a potentially more rewarding behavior. Mindfulness training has been shown to reduce maladaptive eating behaviors (e.g., emotional eating, external eating, binge eating, reactivity to food cravings, restrained eating, and mindless eating) across multiple studies [13, 40, 41, 42]. However, to our knowledge, there are no scalable mindfulness interventions developed to address these behaviors.

To address the lack of interventions targeting reward‐related eating behaviors, researchers at Brown University developed a theory‐driven, mobile‐health (mHealth)‐delivered mindfulness intervention, Eat Right Now (ERN), that emphasizes food reward valuation, mindful eating, and other mindfulness skills in addition to standard behavioral treatment concepts (e.g., goal setting, self‐monitoring, and stimulus control) [15, 43]. Published pilot data from the ERN program show reductions in craving‐related eating as well as a reduction in reward value of craving‐related eating behaviors [12, 44]. These results suggest a mechanistic pathway by which mindfulness training may help patients maintain long‐term weight loss by using awareness training to recalibrate food rewards. However, the extent to which ERN promotes weight loss is unknown. Additionally, prior research has suggested that asynchronous coaching may augment app‐based weight loss programs [45]. Thus, researchers sought to test the feasibility of using ERN for weight loss in a pilot randomized trial comparing individualized asynchronous coaching of the ERN program with that of the ERN program alone.

2. Materials and Methods

The goal of the Practical Awareness‐Based Strategies (PASE) study was to obtain preliminary data on the feasibility of the mindfulness‐based, reward‐related eating intervention (Eat Right Now) for weight loss using a pilot randomized trial involving 20 KPWA patients with BMI ≥ 25 kg/m². All aspects of this study were reviewed and approved by the KPWA Institutional Review Board.

2.1. Participants and Randomization

KPWA members who were between 18 and 65 years of age with a BMI ≥ 25 kg/m² were recruited using social media advertisements about the PASE study, which were posted on the KPWA Health Research Institute's Facebook, Twitter and LinkedIn accounts with language inviting individuals interested in participating in a 6‐month study of two mindfulness‐based interventions for weight loss. Social media recruitment activities occurred for 1 week in April 2021. All interested participants were directed in the social media recruitment materials to a web‐based screening survey using REDCap (www.redcap.org) to verify eligibility. Additional inclusion criteria required having a smart phone or tablet with Wi‐Fi access and an interest in mindful eating. Exclusion criteria included being currently pregnant, currently with a weight ≥ 396 pounds (the limit of the digital scale), currently taking insulin, already enrolled in a weight management program, or having had bariatric surgery in the past five years. If eligible, individuals were called by study staff to confirm eligibility and obtain oral consent to receive an email that included an online consent form to review and sign and a baseline survey to complete. Researchers aimed to enroll 20 participants in this pilot study.

After participants completed the consent process, they were randomized to one of two approaches, using REDCap's randomization module. Approach 1 was the Eat Right Now program and weekly study materials reinforcing the contents from ERN (cohort referred to as “uncoached”). Approach 2 was the same as Approach 1, with the addition of asynchronous coaching (cohort referred to as “coached”). Both groups were provided a Wi‐Fi scale and were instructed to weigh themselves the first thing in the morning at least once per week throughout the study, wearing the same amount of clothing.

All enrolled participants had an initial call with a coach to orient the participant to the program, tell their randomization assignment, and to establish a weight loss goal of 5%–7% of total body weight.

2.2. Eat Right Now Intervention Overview

Eat Right Now (ERN) emphasizes food reward valuation, mindful eating, and other mindfulness skills in addition to standard behavioral treatment concepts (e.g., goal setting, self‐monitoring, and stimulus control). ERN delivers content using brief (5–10 min), daily video and audio lessons through a smartphone or tablet app (iOS and Android) and an online platform [12]. Each lesson includes a video or audio lecture with animation and straightforward guidance as to how to practice the mindfulness principles being taught (typically a meditative or mindful eating technique). Modules for future days are locked to prevent participants from skipping ahead; however, participants have unlimited access to review previous modules. New modules are not unlocked until the participant has completed the current module and a new calendar day has started. Participants can take days off from the app and resume where they left off, review previous modules, and/or delay starting a new module to practice their current skills.

In addition to the daily modules, the ERN intervention also includes (1) app‐delivered check‐ins/reminders to encourage engagement, (2) user‐initiated in‐app guided exercises, (3) weekly Zoom‐based, virtual large group classroom learning, led by Dr. Brewer and other trained ERN facilitators, and (4) a large online community, where ERN users can connect to peers to provide and receive support, post questions for Dr. Brewer (and other facilitators), as well as access additional content emphasizing ERN concepts. The weekly large group classroom sessions focused on practical application of the ERN Food Reward Awareness Training (described below), where ERN facilitators guided several participants through the 3 steps using examples from participants' own lives. While there was time for a few participants to ask questions during classroom sessions, the asynchronous coaching provided individualized feedback and an opportunity to check‐in on questions in a more personal setting. The online community provided an opportunity for participants to post questions in a more general setting, which could yield responses from other community members as well as trained facilitators, but that setting did not provide an opportunity for participants to ask questions confidentially. For the PASE study, both groups received a weekly email summarizing their goals for that week, which generally included (1) continuing to work through the ERN program at their own pace, (2) gradually increasing their physical activity toward a goal of 150 min of moderate‐to‐vigorous activity each week, and (3) a brief set of questions or assignments related to content in the ERN program. Coached participants also received additional weekly contact from their personal coach (as described further below).

2.3. Eat Right Now Curriculum

The ERN curriculum tested in this pilot study covered two major content areas: (1) Food Reward Awareness Training, and (2) Mindful Eating and other Mindfulness Practices. The ERN content is organized into weekly “themes” that include 6 days of new content on a topic followed by a one day “Week in Review” that reinforces content from that week. The Core ERN curriculum included four weeks of video content, and 14 additional “theme” weeks of audio only content were available.

The Core ERN lessons and theme weeks teach mindful eating and other core mindfulness practices, with content based on prior published mindful eating and mindfulness interventions while emphasizing awareness of the reward valuation of certain habits [13, 15, 44, 46, 47, 48]. In the mindful eating lessons, individuals learn to pay attention to three key aspects of eating, which are part of the reward valuation process: the why, what, and how. Individuals learn to identify mental states that trigger eating in the absence of hunger (the why), the types of food they eat when these triggers are present (the what), and the way they eat—which is typically mindless and excessive (the how). Other guided mindful eating exercises promote awareness of physical versus hedonic/emotional hunger, sensations of satiety, and taste satisfaction. This helps users identify maladaptive, reward‐related eating habits and then to assess the relative reward of continuing to engage in their current habit or of choosing a potentially more rewarding behavior. Users also receive training in traditional mindfulness practices, including guided breathing, body scans, walking meditation, loving‐kindness, noting practice, and open awareness, which foster and strengthen curiosity, concentration (attentional control), interoceptive (body) awareness, non‐judgmental acceptance, and self‐compassion. Cultivating each of these mindfulness characteristics provides support to users in practicing the three‐step reward valuation process.

2.4. Eat Right Now Food Reward Awareness Training

One major innovation of the ERN program is its emphasis on food reward awareness training, which is a stepwise process of training participants to develop greater awareness of the rewards, results, or outcomes (i.e., bodily sensations, emotions, thoughts) that arise from various reward‐related eating behaviors. The approach has 3 core steps [12, 44]: Step 1 focuses on increasing awareness of one's habitual behavior, as it is essential that one becomes aware of the triggers (i.e., emotional experiences, places, people, situations; e.g., being hungry and pressed for time) and behavioral responses (e.g., eating fast food) that comprise eating habits. Awareness of the why, what, and how of eating is an integral part of this step. Step 2 focuses on making the reward value of eating patterns conscious and linking behavior with the outcomes (rewards/results) that follow from enacting a given habit (e.g., “What do I get from eating this type/amount of food?”). Importantly, the second step emphasizes noticing all the rewards/results (especially bodily sensations and emotions; i.e., not just focusing on positive or negative thoughts that might arise). This step promotes the recalibration of the experience of reward, which is a critical component of reinforcement learning: future behavior is driven by perceived reward value [18, 49]. Step 3 is learning to exist with behavioral triggers rather than (always) acting upon them, with the goal of developing additional strategies for responding to one's specific triggers that are centered around self‐care. Here users practice choosing alternative behaviors—for example, riding out cravings and eating mindfully—and noticing the positive results of healthy eating and not overeating etc. This 3‐step food reward awareness training is emphasized throughout the program's daily lessons, in‐app exercises, coaching, group classroom teaching, and online community content.

To assist ERN users in developing food reward awareness, a novel in‐app exercise called the “Craving Tool” was developed for craving‐related eating behaviors [44]. When ERN users identify that a craving is occurring, they are encouraged to open the Craving Tool and are guided through a three‐step process to update the reward value of the craving. Step 1: imagine eating the type or amount of food that is craved and notice the rewards/results. Step 2: rate how much you are still craving the food relative to when you started. This serves as a record of the current reward value for the behavior. Step 3: choose to either ride out the craving (using a mindfulness exercise) or to follow a guided mindful eating exercise. After Step 3, users are prompted to reassess their rewards/results (this can be prompted to repeat in 5–20 min, as it can take time to feel the effects of overeating). For example, someone might crave cake, go through Step 1, and still record a strong craving in Step 2 (indicating a high stored reward value in their brain). In Step 3, if they overeat and pay attention to the results (e.g., uncomfortably full, sluggish, etc.), this reduces the reward value of that behavior, which then influences future behavior. Or they might eat the cake mindfully and stop eating before overindulging. This increases their reward value for eating mindfully, and they may be less likely to overeat in the future.

2.5. Approach 2: Eat Right Now Plus Asynchronous Coaching Intervention

The Approach 2 “coached” intervention included all the above ERN content, plus a once‐weekly check‐in email (that included 4–5 questions about the participant's experience with the prior week's content) from their assigned coach, and participants were also able to chat asynchronously as needed with the coach using the Eat Right Now app throughout the program. In brief, coached participants were asked to read and respond to a weekly coaching check‐in email that had a standard format: (1) a question about that weeks' coaching content, (2) a question about that week's physical activity goal, (3) a question about any challenges anticipated in the coming week, and (4) a question about progress in the Eat Right Now modules. The coach responded to any questions that the participant raised, provided guidance on addressing challenges, and offered encouragement around any successes. Coaches completed an 8‐week coaches' training for the Eat Right Now program (led by JB) that involved a weekly 1‐h observation of the ERN virtual classroom learning session followed by a 90‐min group‐based discussion of core Eat Right Now concepts and small group role‐playing activities. The two trained PASE coaches (DA and RG) conducted all the coaching check‐ins and responded to participants' chats, questions and concerns within 24–48 h. Coaches asked questions about the weekly reading materials and followed up with participants on the current module they were currently working through on the ERN app.

Approach 1 participants had an estimated contact/intervention time commitment of 13.5 h over the 6‐month study period. Approach 2 participants had an estimated time commitment of 20 h over the study period.

2.6. Physical Activity Recommendations

Each group was advised to include physical activity each week as follows: Weeks 1–5, 45 min per week; Weeks 6–7, 60 min per week; Weeks 8–9, 80 min per week; Weeks 10–11, 100 min per week; Weeks 12–15, 120 min per week; Weeks 16–17, 150 min per week; Weeks 18–19, 200 min per week; Week 20, 225 min; Weeks 21 and up, 250 min per week.

2.7. Study Measures

Data were collected using Wi‐Fi scales, self‐report questionnaires, from the KPWA electronic medical record, and from ERN app use metrics (described below). Self‐report questionnaires were administered at baseline and at 6 months of follow‐up. Wi‐Fi scales collected weight data throughout the study period.

2.7.1. Anthropometric Measures

The primary outcome of interest was change in body weight from baseline to 6 months of follow‐up. Body weight was measured using a Withings Body Wi‐Fi enabled digital scale that was provided to participants via mail. Study staff called participants 7 days after the scale was mailed to train participants on using the scale and ERN app. Participants were instructed to weigh themselves at baseline 3 days in a row at the same time of day to establish a baseline weight. The participants weighed themselves at least weekly at the same time of day.

2.7.2. Self‐Report Questionnaires

Participants completed an online self‐report questionnaire at baseline and 6 months post enrollment that assessed secondary outcomes related to reward‐related eating and exploratory measures related to eating disorders and lifestyle patterns.

2.7.3. Food Craving Questionnaire—Trait—Reduced (FCQ‐T‐R)

The 15‐item FCQ‐T‐R assesses (1) pre‐occupation with food, or obsessive thoughts about food and eating, (2) loss of control overeating or difficulty regulating eating behavior, (3) positive outcome expectancy or believing that eating is positively reinforcing, and (4) emotional craving, or the tendency to crave food when experiencing high levels of emotion [50]. Higher FCQ‐T‐R scores have been associated with more frequent thinking and eating of high calorie snacks, and weight gain over time via increases in disinhibited eating. The 15 Items are answered on a 6‐point scale from 1 (never) to 6 (always). The total score is computed as the sum of all items, with higher scores indicating greater trait reward‐related eating behaviors.

2.7.4. Reward Based Eating Drive Scale (RED)

The 9‐item RED scale addresses reward‐driven eating [31]. Sample items include “When I start eating, I just can't seem to stop” (lack of control), “I don't get full easy” (lack of satiety), and “Food is always on my mind” (preoccupation with food). Participants answer items on a scale from 1 (strongly disagree) to 5 (strongly agree). Greater RED scores have been associated with greater weight and weight gain over time (Epel et al., 2014) as well as stronger daily craving experiences (Mason et al., 2015), and reductions in RED scores have been associated with weight loss (Mason et al., 2016). The total score is computed as the sum of all items, with higher scores reflecting a higher reward‐based eating drive.

2.7.5. Acceptance and Action Questionnaire for Weight (AAQ‐W)

The AAQ‐W is a validated 22‐item measure assessing experiential avoidance and psychological inflexibility related to body weight, food and eating [51]. Respondents indicate the degree of fit of each statement from 1 (never) to 7 (always) true. Higher scores indicate greater weight‐related experiential avoidance.

2.7.6. SCOFF Questionnaire

The validated SCOFF 5‐item questionnaire SCOFF (named for a keyword from each of the 5 questions Sick, Control, One stone, Fat, Food) is used as a screening tool for detecting eating disorders, including anorexia nervosa, bulimia, and binge eating disorder [52].

2.7.7. Exercise Vital Sign (EVS) for Physical Activity

This validated 2‐item measure was developed as part of the Exercise is Medicine initiative and is adapted from the Behavioral Risk Factor Surveillance System (BRFSS) physical activity questions [53].

2.7.8. Weight and Lifestyle Inventory (WALI)

The WALI is designed to obtain information about weight and dieting history. Researchers adapted the WALI to identify areas to work with participants during the initial coaching call and for participants in Approach 2 [54].

2.7.9. Six Factor Questionnaire (6‐FQ)

The 6‐FQ is a validated 27‐item instrument that measures a respondent's lifestyle patterns, which include (1) convenient diner, (2) fast pacer, (3) easily enticed eater, (4) exercise struggler, (5) self‐critic, and (6) all‐or‐nothing doer [55]. This questionnaire was administered to understand participants' self‐identified lifestyle patterns and if these changed over the course of the intervention.

2.7.10. Engagement and Satisfaction

Additional questions at the 6 months follow up questionnaire asked participants about their satisfaction with components of the study, including ERN, the Wi‐Fi digital scale and with the study overall. Approach 2 participants were also asked questions on satisfaction with the coaching aspect of the study.

2.8. Analysis

As we were conducting a small, feasibility pilot study, we did not perform any statistical hypothesis tests, and all analyses are descriptive in nature. The primary outcome of this feasibility pilot was to demonstrate retention of enrolled participants. We therefore described the percentage retained and the percentage achieving complete data capture for weight loss and for self‐reported questionnaire‐based outcomes.

In exploratory analyses, we also examined the percent total weight loss (%TWL) at 6 months, calculated as follows: weight in kilograms at 6 months/weight in kilograms at baseline × 100. The weight analyses used the average of the first three pre‐treatment weight measures to establish each person's baseline weight.

Average weight loss was estimated via random effect regression on weight over time with a random intercept, controlling for baseline weight [56]. A restricted cubic spline with two knots was used to model weight change over time, and interaction terms with treatment group were included to allow different estimated weight trajectories for each group. The percentage weight loss at 180 days was then calculated via marginal standardization, using the predicted weight at 180 days divided by baseline weight as the value to be estimated.

To estimate the percentage of people achieving 5% weight loss at 180 days, data was restricted t to one record per person, creating variables denoting the weight closest to each 30‐day period, then performing multiple imputation via chained equations using those weight values, race, ethnicity, and intervention group, creating 100 imputed datasets [57]. In each dataset, a participant was coded as meeting the threshold if their weight at 180 days divided by their baseline weight was less than or equal to 0.95 (5% weight loss). Logistic regression was then run on the multiply imputed data and used marginal standardization to obtain estimated values of the proportion of individuals achieving the weight loss thresholds.

Additional exploratory descriptive analyses were conducted on the self‐reported measures related to reward‐related eating, eating disorders, and lifestyle patterns, calculating means and standard deviations for these continuous variables at baseline and at 6 months follow‐up.

Analyses were conducted in Stata 15.1 (StataCorp. 2017. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC.) Figures were created in RStudio (R Core Team 2019).

3. Results

Thirty‐two people expressed initial interest in the study and among these 20 were enrolled and randomized to the two intervention groups (Figure 1). Among these 20 participants, two participants were lost to follow‐up before starting the intervention (both were assigned to the coached group; they did not return phone calls to set up the initial coaching call and did not begin using the app hence were not yet informed of their randomization status), and one participant in the coached group disenrolled from the study after 61 days citing lack of time (Figure 1). These three individuals were all females, were slightly younger than the overall population (average age of 38.4 years), and they had an average baseline average BMI of 35.4 kg/m². No adverse events were reported during the study.

FIGURE 1.

Flow diagram of recruitment and retention of study participants. *Group 1 = Eat Right Now only; Group 2 = Eat Right Now plus coaching.

3.1. Participant Characteristics at Baseline

Among the 17 participants who enrolled and completed 6‐month follow‐up, they had a mean age of 48.65 ± 9.5 and a mean BMI of 34.2 ± 9.6 kg/m² at baseline (Table 1). The sample was predominantly female (82.4%) and was highly educated with 94% having completed college or post‐collegiate education. Participants in the uncoached group were slightly older (mean age 50.6 vs. 45.8) and more educated (100% college or post‐collegiate education vs. 60%) than the coached group. In the six months before the intervention, two participants (one in each group) lost 10 or more pounds, while the rest were stable or increasing in weight.

TABLE 1.

Characteristics of the 17 study participants who initiated the intervention and remained enrolled through 6 months.

	Overall	Approach 1: ERN only	Approach 2: ERN + coaching
Total (n)	17	10	7
Demographics
Female, n (%)	14 (82.4%)	8 (80%)	6 (85.7%)
Age at baseline, years (mean [SD])	48.65 (9.45)	50.60 (6.83)	45.87 (12.36)
Education, n (%)
< High school	0 (0.0%)	0 (0.0%)	0 (0%)
High school	2 (11.8%)	0 (0.0%)	2 (20%)
Some college	2 (11.8%)	0 (0.0%)	2 (20%)
College	7 (41.1%)	5 (50.0%)	2 (20%)
> College	9 (52.9%)	5 (50.0%)	4 (40%)
Weight and height
BMI, kg/m2 (mean, [SD])	34.18 (9.62)	35.72 (10.88)	31.97 (7.71)
Overweight (25.0–29.9), n (%)	7 (41.2%)	3 (30.0%)	4 (57.1%)
Class I obesity (30.0–34.9), n (%)	4 (23.5%)	3 (30.0%)	1 (14.3%)
Class II obesity (35.0–39.9), n (%)	2 (11.8%)	2 (20.0%)	0 (0.0%)
Class III obesity (40.0+), n (%)	4 (23.5%)	2 (20.0%)	2 (28.6%)
Weight in lbs (mean, [SD])	213.00 (59.71)	219.60 (59.77)	203.57 (63.02)
Height in inches (mean, [SD])	66.22 (3.79)	66.08 (4.21)	66.43 (3.41)
During the past 6 months my weight has…
Decreased by more than 10 lbs.	2 (11.8%)	1 (10.0%)	1 (14.3%)
Decreased by 5–10 lbs.	0 (0.0%)	0 (0.0%)	0 (0.0%)
Been relatively stable	10 (58.9%)	6 (60.0%)	4 (57.1%)
Increased by 5–10 lbs.	1 (5.9%)	0 (0%)	1 (14.3%)
Increased by more than 10 lbs.	4 (23.5%)	3 (30.0%)	1 (14.3%)
Medical history
Diabetes (type I or II)	2 (11.8%)	2 (20.0%)	0 (0.0%)
Arthritis	4 (23.5%)	3 (30.0%)	1 (14.3%)
Hypertension	5 (29.4%)	3 (30.0%)	2 (28.6%)
Physical activity
Enjoyment of physical activity
Greatly	6 (35.3%)	4 (40.0%)	2 (28.6%)
Moderately	7 (41.2%)	4 (40.0%)	3 (42.9%)
Slightly	3 (17.6%)	2 (20.0%)	1 (14.3%)
Not at all	1 (5.9%)	0 (0.0%)	1 (14.3%)
Daily lifestyle activity (mean, [SD]) a	4.47 (2.10)	4.00 (2.16)	5.14 (1.95)
Minutes of activity per week (mean, [SD])	98 (96.77)	106 (121.58)	90 (69.6)
Average days of PA per week
1–2 days	9 (52.9%)	5 (50.0%)	4 (57.1%)
3–4 days	2 (11.8%)	1 (10.0%)	1 (14.3%)
5–7 days	5 (29.4%)	3 (30.0%)	2 (28.6%)
Missing	1 (5.9%)	1 (10.0%)	0 (0.0%)
Average minutes of exercise
0 min	0 (0.0%)	0 (0.0%)	0 (0.0%)
10 min	1 (5.9%)	1 (10.0%)	0 (0.0%)
20 min	4 (23.5%)	3 (30.0%)	1 (14.3%)
30 min	3 (17.6%)	1 (10.0%)	2 (28.6%)
40 min	4 (23.5%)	2 (20.0%)	2 (28.6%)
60 min	4 (23.5%)	2 (20.0%)	2 (28.6%)
Missing	1 (5.9%)	1 (10.0%)	0 (0.0%)
Baseline scores
FCQ‐T‐R score (mean, [SD]) b	50.88 (13.61)	44.80 (11.46)	59.57 (12.15)
RED score (mean, [SD]) c	19.23 (9.81)	21.60 (10.23)	15.86 (8.76)
AAQ‐W score (mean, [SD]) d	82.82 (17.21)	79.00 (16.32)	88.29 (18.20)
SCOFF Score (mean, [SD]) e	1.18 (0.81)	1.00 (0.82)	1.43 (0.79)
Six factor total score (mean, [SD]) f	30.41 (10.66)	28.10 (9.67)	33.71 (11.88)
Convenient diner (%)	0.0 (0%)	0 (0.0%)	0 (0.0%)
Fast pacer (%)	5 (29.4%)	3 (30.0%)	2 (28.6%)
Easily enticed eater (%)	6 (35.3%)	4 (40.0%)	2 (28.6%)
Exercise struggler (%)	2 (11.8%)	1 (10.0%)	1 (14.3%)
Self‐critic (%)	3 (17.6%)	1 (10.0%)	2 (28.6%)
All or nothing doer (%)	1 (5.9%)	0 (0.0%)	1 (14.3%)

^a On a scale from 1 to 10, where 1 is very sedentary and 10 is very active.

^b FCQ‐T‐R Score ranges from 15 to 90 and a higher score equals worse cravings.

^c The RED score ranges from 9 to 45 and a higher score equals a worse reward‐based eating drive.

^d The AAQ‐W score ranges from 22 to 154 and a higher score equals worse weight‐related experience and acceptance.

^e SCOFF score ≥ 2 indicates a likely case of anorexia nervosa or bulimia.

^f The Six Factor Score includes: Convenient Diner: importance of having regular meal rhythm, calorie control and awareness, more healthy foods and planning. Fast Pacer: importance of time management, prioritization of self‐care and effect of stress. Easily Enticed Eater: importance of food temptation and regulation; environmental cueing of food intake and hedonic reward of particular food. Exercise Struggler: importance of physical activity and inactivity in weight management; experience various barriers that deter from increasing physical activity. Self‐Critic: body dissatisfaction and body image disparagement. All‐or‐nothing Doer: dichotomous thinking (perfectionism, high personal standards, self‐criticism, difficulty with adaption to new behaviors, high emotional distress) or lack of moderation.

Participants in the uncoached group more often reported diabetes (20% vs. 0%) and arthritis (30% vs. 14%), but had a similar prevalence of hypertension (30% vs. 28.6%). Participants in the uncoached group also more often reported that they enjoyed physical activity greatly (40% vs. 28.6%). The mean minutes of physical activity per week at baseline were 106 for uncoached and 92 for coached participants.

On measures of reward‐related eating, coached participants reported higher FCQ‐T‐r scores (59.6 ± 12.2 vs. 44.8 ± 11.5) but had similar scores on the RED (15.8 ± 8.8 vs. 21.6 ± 10.2). The coaches also had higher scores on the AAQ‐W (88.3 ± 18.2 vs. 79.0 ± 16.3), SCOFF (1.4 vs. 1.0) and the Six Factors (33.7 ± 11.8 vs. 28.1 ± 9.7).

3.2. Primary Outcomes of Interest: Retention, Engagement, and Satisfaction

The characteristics of the 17 participants that started the intervention and remained enrolled through six months were similar to the overall recruited population (Table S1).

At 6 months follow‐up, 82% (14/17) had weight data captured by Wi‐Fi scale, as three participants did not weigh themselves at the 6‐month timepoint. The mean number of weights captured was 86, with a range of 36–330. Six‐month questionnaires were completed by 12/17 (70.5%).

Among the 17 participants that started the intervention and remained enrolled, they completed an average of 22.9 of the 28 core ERN modules, and 11 out of 17 participants (64.7%) completed all 28 core ERN modules during the 6‐month follow‐up. In terms of ongoing engagement with the app, 12 out of 17 (70.5%) participants were still logging into the app in the final month of the 6‐month intervention period. The mean number of app logins was 212 (range 55–475). Participants were engaged with the app for a long time, with a mean duration of engagement of 144 days (range 47–180 days).

When asked to rate their satisfaction with the program using a 4‐point Likert scale ranging from not satisfied (1) to very satisfied (4), the mean score was 3.5, with 75% of participants being satisfied or very satisfied. Three were a little satisfied, 2 were satisfied and 7 were very satisfied.

3.3. Exploratory Outcome: Weight Loss

Figure 2 shows the individual trajectories of %TWL over the six months of follow‐up for each of the 17 participants that started the intervention and remained enrolled. The trajectories show wide heterogeneity in treatment response, where some participants achieved ∼10% weight loss, others clustered around 4%–5% weight loss, and some lost little to no weight.

FIGURE 2.

Percent change in weight over time for study participants.

The estimated mean %TWL from the random‐effects model at 6 months was 4.3% (95% confidence interval: 2.3%, 6.3%) for coached participants and 3.6% for uncoached participants (95% CI: 1.2%, 5.9%). The estimated proportions achieving clinically meaningful weight loss (≥ 5%) from the logistic regression model were 30.4% (95% confidence interval: 0.0%, 63.8%) of coached and 27.9% (95% Confidence interval: 0.2%, 55.6%) of uncoached participants.

3.4. Exploratory Outcomes: Survey Measures

Participants in both approaches had reductions in reward‐related eating after 6 months (Table 2). On the FCQ‐T‐r, scores were reduced by 34.1% from baseline (mean 51.8 ± 12.2 at baseline, 34.2 ± 12.9 at 6 months). On the RED scale, scores were 38.7% lower (mean 14.3 ± 6.1 at baseline, 8.7 ± 8.8 at 6 months) for participants who completed both baseline and 6‐month surveys.

TABLE 2.

Baseline and 6‐month follow‐up survey scores.

	Baseline			6 Month follow up
	Overall	Approach 1: ERN only	Approach 2: ERN + coaching	Overall	Approach 1: ERN only	Approach 2: ERN + coaching
Total (n) with survey data	12	7	5	12	7	5
FCQ‐T‐R score (mean, [SD]) a	51.83 (12.22)	45.86 (10.12)	60.20 (10.35)	34.17 (12.92)	31.14 (12.47)	38.40 (13.69)
RED score (mean, [SD]) b	14.25 (6.11)	16.57 (6.83)	11.00 (3.16)	8.73 (8.81)	11.25 (10.44)	7.29 (8.26)
AAQ‐W score (mean, [SD]) c	85.00 (14.37)	82.86 (12.46)	88.00 (17.78)	69.36 (10.44)	64.29 (9.39)	78.25 (4.65)
SCOFF score (mean, [SD]) d	1.17 (0.83)	1.00 (0.82)	1.40 (0.89)	0.58 (1.00)	0.29 (0.49)	1.00 (1.41)
Six factor total score (mean, [SD]) e	31.17 (11.42)	29.00 (9.78)	34.20 (13.99)	21.08 (9.30)	18.86 (10.49)	24.20 (7.19)
Convenient diner (%)	0.0 (0%)	0 (0.0%)	0 (0.0%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
Fast pacer (%)	3 (25.0%)	2 (28.6%)	1 (20.0%)	2 (16.7%)	1 (14.3%)	1 (20.0%)
Easily enticed eater (%)	5 (41.7%)	3 (42.9%)	2 (40.0%)	2 (16.7%)	1 (14.3%)	1 (20.0%)
Exercise struggler (%)	1 (8.3%)	1 (14.3%)	0 (0.0%)	1 (8.3%)	1 (14.3%)	0 (0.0%)
Self‐critic (%)	3 (25.0%)	1 (14.3%)	2 (40.0%)	1 (8.3%)	0 (0.0%)	1 (20.0%)
All or nothing doer (%)	1 (8.3%)	0 (0.0%)	1 (20.0%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
Factors endorsed (mean, [SD])	1.08 (1.16)	1.0 (0.82)	1.20 (1.64)	0.5 (0.67)	0.43 (0.79)	0.6 (0.55)
No factors endorsed (%)	4 (33.3%)	2 (28.6%)	2 (40.0%)	7 (58.3%)	5 (71.4%)	2 (40%)

^a FCQ‐T‐R Score ranges from 15 to 90 and a higher score equals worse cravings.

^b The RED score ranges from 9 to 45 and a higher score equals a worse reward‐based eating drive.

^c The AAQ‐W score ranges from 22 to 154 and a higher score equals worse weight‐related experience and acceptance.

^d SCOFF score ≥ 2 indicates a likely case of anorexia nervosa or bulimia.

^e The Six Factor Score includes: Convenient Diner: importance of having regular meal rhythm, calorie control and awareness, more healthy foods and planning. Fast Pacer: importance of time management, prioritization of self‐care and effect of stress. Easily Enticed Eater: importance of food temptation and regulation; environmental cueing of food intake; and hedonic reward of a particular food. Exercise Struggler: importance of physical activity and inactivity in weight management; experience various barriers that deter from increasing physical activity. Self‐Critic: body dissatisfaction and body image disparagement. All‐or‐nothing Doer: dichotomous thinking (perfectionism, high personal standards, self‐criticism, difficulty with adaption to new behaviors, high emotional distress) or lack of moderation.

In the other survey measures, the AAQ‐W scores improved by 16.2% from a mean of 85.0 ± 14.4 at baseline to a mean of 69.4 ± 10.4 at 6 months. The mean SCOFF score improved by 50% from a baseline mean of 1.2 ± 0.8 to 0.6 ± 1.0 at 6 months. Finally, the Six Factor total score improved by 32.3% from a mean of 31.2 ± 11.4 at baseline to 21.1 ± 9.3 at 6 months. The mean number of factors endorsed decreased from 1.08 ± 1.2 to 0.5 ± 0.7, and the factors that showed the greatest improvement were the Easily Enticed Eater and the Self‐Critic.

4. Discussion

This 6‐month pilot and feasibility randomized trial of ERN, an mHealth delivered, mindfulness‐based weight‐loss intervention that targets reward related eating, demonstrates that the majority of eligible participants remain enrolled in the intervention and complete follow‐up assessments. Among 17 participants who started the intervention and remained enrolled at 6 months, over 80% had weight data available by Wi‐Fi scale and 70.5% completed the follow‐up questionnaire.

The study also provided preliminary data for weight loss in both the coached and uncoached intervention arms. Coached ERN participants achieved slightly greater average weight loss (4.3% vs. 3.6%) than uncoached participants. 30.4% of coached and 27.9% uncoached participants achieved 5% or greater weight loss, which is widely considered to be clinically meaningful. However, given the pilot nature of this study, these preliminary estimates of weight loss outcomes deserve confirmation in larger studies with greater follow‐up.

These findings are consistent with prior studies involving mHealth delivered weight loss interventions, such as Weight Watchers (WW) and Noom, which both include a coaching component. Noom has published several research studies [58, 59, 60, 61, 62, 63]. The most rigorous and relevant study enrolled 202 adults with BMI ≥ 25 kg/m² and prediabetes and randomized them to Noom or regular medical care including a paper‐based behavioral intervention [63]. After six and 12 months of intervention, the Noom group had lost 3.7% (95% CI: −2.5% to −4.9%) and 2.5% (95% CI: −1.3% to −3.7%) of total body weight compared to −0.2% (+1.1 to −1.4%) and +0.3 (+1.7 to −1.1%) for controls. In a study of 279 WW adults with BMI 27–40 kg/m², participants were randomized to 12 months of WW alone, 12 months of WW with an activity tracker, or an online newsletter (control) [64]. At 3 months, weight loss was greater in the WW group (2.7 kg, 95% CI: 2.0–3.5 kg) than in the control group (1.3 kg, 95%: 0.5–2.0 kg). At 12 months, there was no significant difference in weight loss between all 3 groups (p > 0.52). 25.5% of WW achieved > 5% weight loss at 12 months compared to 12.9% of controls (p = 0.04). The approach appears to have preliminary efficacy and could be compared with these programs in the future. These findings are consistent with other behavioral weight management programs delivered in a variety of settings [65], including primary care settings [66].

Although our response rate to the follow‐up questionnaire was limited (70.5%), the exploratory findings from this pilot trial are consistent with the prior pilot studies, which suggest that ERN leads to improvements in reward‐related eating behaviors. The first pilot study of ERN recruited adult women with overweight or obesity who were experiencing food cravings (n = 104, M age = 46.1 ± 14.6; M BMI = 31.2 ± 4.3 kg/m²) [12]. At 1 month after completing the final module of the intervention, participants had significant pre‐to post‐intervention reductions in craving related eating as measured by FCQ‐T‐R (25% reduction, p < 0.001) and RED (29% reduction, p < 0.001) and a text‐based EMA that assessed the presence or absence of craving‐related eating within the past hour found a 40% reduction in craving‐related eating (t = 5.68, p < 0.001) [12]. These findings were confirmed in a second, 8‐week, single‐arm pilot study, involving 46 adult women (mean age: 43 years; mean BMI: 28.6 kg/m²) who endorsed experiencing food cravings [44].

A mindfulness‐based program such as this one that addresses reward‐related eating behaviors may also be of benefit to individuals who undergo bariatric surgery. As noted in the review by Sarwer et al., impulsivity is a central factor underlying maladaptive eating behaviors in adults with severe obesity [67]. Coaching on reward‐related eating may help limit impulsivity and mitigate other psychological and behavioral factors such as binge‐eating that can contribute to long‐term weight regain after bariatric operations. Prior research has also highlighted the importance of hedonic hunger, which is one's pre‐occupation with or desire to consume palatable foods for the purposes of pleasure and in the absence of hunger [68].

Limitations of this study include its small sample size, short duration of follow‐up, and low response rate to the 6‐month survey. This low response rate could introduce bias if the participants who provided weight and questionnaire data in follow‐up were different from the those who disenrolled or were lost to follow‐up. This team is currently developing a 22‐week ERN program to supplement the 4‐week core intervention with modules on nutrition and physical activity, modeled after the diabetes prevention program, for weight loss; future research should assess the longer‐term impact of this intervention on health outcomes. Because of the sample size and response rate, the current study has insufficient statistical power to detect any clinically significant between‐group differences in weight loss and reward‐ and craving‐related eating behaviors. We recruited individuals who expressed an interest in mindful eating, which may limit the generalizability of our study beyond adults who have similar interests. Furthermore, the educational level of our study participants was high, which may also limit our generalizability to adults with lower educational attainment or socio‐economic status, since education level may correlate with having the time and resources to engage in mindfulness training, interest in mindfulness, and ability to acquire psychological skills through an mHealth intervention [69, 70, 71, 72, 73].

Overall, this study and the prior pilot studies demonstrate that the ERN intervention: (1) is safe, highly acceptable, and engaging among adults with BMI ≥ 25 kg/m²; (2) engages its behavioral target through a specific mechanism: reward‐related eating via reward valuation; (3) is associated with improvements in validated measures of reward‐related eating through at least 6 months follow‐up with large effect sizes (> 1.1) among those with and without endorsements of food cravings at baseline; [12] (4) is associated with moderate weight loss through 6 months, particularly among those with sustained use of the program; and (5) in a prior pilot study [12], correlates greater reductions in reward‐related eating with greater weight loss.

This study leaves several unanswered questions that deserve further research. First, is the effect of the ERN on reward‐related eating durable beyond six months? Second, is the improvement in reward‐related eating with ERN greater than what can be achieved with other industry‐leading standard behavioral weight loss interventions, such as Noom or WW? Finally, can a year‐long ERN intervention curriculum be more effective than the original 28‐day “core” ERN intervention that was tested in this pilot study? Another question is how best to provide asynchronous support. The results from this study suggest that asynchronous coaching could be helpful for some but not necessary for others, and a stepped care approach (i.e., providing extra support to those struggling) could be more effective and less resource intensive.

Author Contributions

D.A., R.G., D.R., J.N., and J.B. study design, obtaining funding, data collection, data interpretation, literature search, writing of manuscript. E.J., data analysis, data interpretation, editing of manuscript. K.M. study design, data collection, interpretation of data, editing of manuscript.

Ethics Statement

The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Kaiser Permanente Washington Institutional Review Board.

Consent

Written informed consent was obtained from all participants in the study.

Conflicts of Interest

Dr. Brewer was a paid adviser to Sharecare, the digital healthcare company that owns the Eat Right Now program, during the conduct of the study. He is not currently a paid advisor to Sharecare. No other researchers on this project have any financial ties or conflicting interests. Dr. Brewer worked closely with the Brown University Conflict of Interest committee to ensure subject safety and data integrity and developed a plan that was reviewed and updated yearly to ensure adequate protections. DA, RG, DR, and JN have received a separate contract from Sharecare, Inc. to assist them with the development of an Eat Right Now diabetes prevention program. Funds for this contract were paid to Kaiser Permanente Washington.

Supporting information

Table S1