a Remotely Delivered Behavioral Intervention for ...

OBES SURG DOI 10.1007/s11695-016-2337-3

ORIGINAL CONTRIBUTIONS

Project HELP: a Remotely Delivered Behavioral Intervention for Weight Regain after Bariatric Surgery Lauren E. Bradley 1 & Evan M. Forman 2 & Stephanie G. Kerrigan 2 & Stephanie P. Goldstein 2 & Meghan L. Butryn 2 & J. Graham Thomas 3 & James D. Herbert 2 & David B. Sarwer 4

# Springer Science+Business Media New York 2016

Abstract Background Weight regain following bariatric surgery is common and potentially compromises the health benefits initially attained after surgery. Poor compliance to dietary and physical activity prescriptions is believed to be largely responsible for weight regain. Patients may benefit from developing specialized psychological skills necessary to engage in positive health behaviors over the long term. Unfortunately, patients often face challenges to physically returning to the bariatric surgery program for support in developing and maintaining these behaviors. Remotely delivered interventions, in contrast, can be conveniently delivered to the patient and have been found efficacious for a number of health problems, including obesity. To date, they have received little attention with bariatric surgery patients. The study aimed to evaluate a newly developed, remote acceptance-based behavioral intervention for postoperative weight regain. Methods Patients at least 1.5 years out from surgery who experienced postoperative weight regain were recruited to re-

ceive the 10-week intervention. Participants were assessed at baseline, mid-treatment, post-treatment, and at 3-month follow-up. Results Support for the intervention’s feasibility and acceptability was achieved, with 70 % retention among those who started the program and a high mean rating (4.7 out of 5.0) of program satisfaction among study completers. On average, weight regain was reversed with a mean weight loss of 5.1 ± 5.5 % throughout the intervention. This weight loss was maintained at 3-month follow-up. Significant improvements in eating-related and acceptance-based variables also were observed. Conclusions This pilot study provides initial support for the feasibility, acceptability, and preliminary efficacy of a remotely delivered acceptance-based behavioral intervention for postoperative weight regain. Keywords Postoperative weight regain . Behavioral intervention . eHealth . The research described in this paper

* Lauren E. Bradley [email protected]

James D. Herbert [email protected] David B. Sarwer [email protected]

Evan M. Forman [email protected] Stephanie G. Kerrigan [email protected]

1

Department of Behavioral Sciences, Rush University Medical Center, 1645 W. Jackson Blvd. Suite 400, Chicago, IL 60625, USA

Stephanie P. Goldstein [email protected]

2

Department of Psychology, Drexel University, Philadelphia, PA, USA

Meghan L. Butryn [email protected]

3

Weight Control and Diabetes Research Center, Warren Alpert Medical School of Brown University, Providence, RI, USA

J. Graham Thomas [email protected]

4

Center for Obesity Research and Education, Temple University, Philadelphia, PA, USA

OBES SURG

was supported in part by grants to the first author from the American Psychological Association and the Society for a Science of Clinical Psychology.

Introduction/Purpose Although bariatric surgery is the most successful weight loss treatment, 20–30 % of patients fail to reach targeted weight loss (i.e., 25 to 35 % of total body weight) or begin to regain large amounts of weight beginning 6 to 24 months postoperatively [1–4]. Longer-term data suggest that 30 to 70 % of patients fail to maintain a 20 % weight loss 10 years after surgery, depending on type of surgery received [1]. These suboptimal weight outcomes have critical implications, as they have been linked with poorer postoperative health outcomes [2, 3, 5, 6]. Poor long-term weight outcomes are largely attributed to behavioral factors, including non-adherence to the recommended postoperative diet and the development or re-emergence of maladaptive eating behaviors (e.g., grazing, loss of control eating, emotional eating) [7–10]. Immediately after surgery, compliance to postoperative dietary recommendations is facilitated by the physical and metabolic effects of surgery. Specifically, the volume of food that can be consumed is limited by the restrictive nature of surgery and changes in gut hormones are believed to mediate changes in hunger and food preferences [11, 12]. However, it is possible that over time, some effects deteriorate, making adherence increasingly difficult for patients. Specifically, following the initial weight loss phase, patients report increased hunger, food cravings, and ability to physically consume greater amounts of food [13–16]. Long-term changes in eating behavior are therefore necessary for sustained success post-surgery. Given the need to enhance compliance with recommended postoperative eating behaviors in order to improve weight loss and related health outcomes, a number of studies have evaluated the efficacy of postoperative behavioral weight control interventions. A meta-analysis of controlled trials of these interventions found minimal differences in percent excess weight loss between treatment and control conditions (i.e., 1.6 %) 6–12 months following the beginning of the intervention [17], suggesting that the provision of behavioral skills alone has modest effects on weight. Standard interventions such as these provide participants with behavioral strategies (e.g., self-monitoring, stimulus control) to target weight regain. However, given the challenges reported by patients who are regaining weight (e.g., return of food cravings, increased hunger, ability to consume greater amounts of food), there is a need for interventions that also provide psychological skills to help patients appropriately engage in weight control behaviors. Acceptance-based treatments (ABTs) are a type of behavioral intervention that provides patients with skills to enhance

one’s willingness to experience a range of internal experiences (e.g., thoughts, emotions, urges) in order to behave in ways consistent with what is most important to them [18–20]. These interventions are largely based on Acceptance and Commitment Therapy (ACT) [21]. ABTs for weight control specifically target the challenges faced by postoperative patients by helping individuals enact goal-directed behaviors (e.g., healthy eating, physical activity) that are in line with one’s values in spite of aversive (or non-preferred) internal experiences, such as those that lead to dietary inadherence [22, 23]. These skills (i.e., distress tolerance, presentmoment awareness of internal states, clarity of one’s personal values and linking values to in-the-moment decision-making, psychological distancing, or Bdefusion^ (as described in [20])) are synthesized with behavioral elements known to be crucial for weight control. It is hypothesized that ABTs will improve post-surgery patients’ adherence to demanding dietary guidelines which will, in turn, enable weight loss maintenance in the years following bariatric surgery. Previous research supports the use of ABTs for weight loss in nonsurgical populations [23–25], and several studies have applied acceptance-based strategies to bariatric surgery patients; however, data are limited [26–28]. Weineland et al. [26] reported improvements in disordered eating, body dissatisfaction, and quality of life following an ABT compared to treatment as usual; however, changes in weight were not measured. Two mindfulness-based interventions have shown to enhance postoperative weight outcomes; however, sample sizes were small (i.e., n = 1 and n = 7 [27, 28]). Our previous research also provides support for ABT for this population. We conducted an open trial of a newly developed, 10-week, ABT in-person group intervention for bariatric surgery patients who displayed a 10 % weight regain since their lowest weight after surgery (n = 11; [29]). The intervention was shown to be feasible and acceptable, with 72 % retention (100 % retention in those who attended more than 1 session) and high mean rating (4.25 out of 5.00) of program satisfaction. In addition, a mean weight loss of 3.6 ± 3.0 % throughout the 10-week intervention was observed. There were also significant improvements in eating-related variables, including decreased responsivity to internal cues. Even with these encouraging results, delivering the intervention postoperatively remains a significant challenge. Patients struggle to return to their bariatric programs for regular follow-up, limiting the effective delivery of treatment. Over two thirds of patients (72 %) miss appointments during the first two postoperative years [30], and only 40 % of patients return for their first four annual follow-up visits [31]. Studies evaluating behavioral postoperative interventions similarly report high attrition. For example, in one study, only 20 % of participants who expressed interest in a postoperative intervention enrolled in the study and approximately 40 % of consented participants withdrew from the study prior to group

OBES SURG

assignment [32]. In another study, 60 % of patients attended half or less than half of the intervention sessions [33]. Several studies have found a correlation between travel distance and postoperative follow-up [34–36], indicating that in-person appointments may be difficult for many patients to attend. Remote interventions (e.g., delivered via phone and/or Internet) appear to be well-suited to address the challenges of engaging bariatric surgery patients postoperatively. Internet-based interventions targeting weight loss are particularly desirable, due to cost-effectiveness, increased access, and reduced patient burden [37]. Internet-based interventions have been shown to result in meaningful weight losses in nonsurgical populations [38, 39]. Previous research has also supported the application of remotely delivered ABTs, including those for type 2 diabetes [40], chronic pain [41], smoking cessation [42], anxiety disorders [43–45], depression [46], and tinnitus [47]. In addition, Weineland and colleagues’ (2012) ABT intervention for bariatric surgery patients (described above [26]) combined face-to-face sessions with Internet modules; 12 of 16 participants were retained from study enrollment to follow-up [48]. The current study builds on this prior work and was designed to investigate the feasibility, acceptability, and preliminary efficacy of a novel remotely delivered ABT intervention for postoperative weight regain. Based on our preliminary data [29], we hypothesized that at post-treatment, patients would demonstrate significant weight loss and improvements in self-reported eating behavior, as well as change in the postulated mechanisms of action (i.e., acceptance of internal experiences, defusion, or psychological distancing).

Materials and Methods Participants Participants were recruited by contacting respondents to a survey [13] of attitudes of patients on postoperative interventions from the University of Pennsylvania Bariatric Surgery Program and by posting advertisements through community flyers and Craigslist. Inclusion criteria were 18–70 years old, weight loss surgery (i.e., RYGB, gastric sleeve, gastric banding) at least 1.5 years out from surgery, ≥10 % weight regain of maximum weight loss or 5 % of their minimum weight post-surgery, and weight regain lasting for at least 3 months prior to enrollment. Potential participants were excluded if they were enrolled in a structured weight loss program, were pregnant/planned to become pregnant within 6 months of enrollment, had a medical condition that had the potential to affect weight (e.g., Cushing’s disease) or would limit one’s ability to make dietary or physical activity changes, exhibited psychiatric symptoms that would interfere with the ability to benefit from the intervention (i.e., inability

to provide consent or to implement behavior changes), or reported acute suicidality. Medications known to affect body weight (gain or loss), such as chronic systemic steroids or psychiatric medications including lithium, tricyclic antidepressants, and anti-psychotic agents [49], were required to be stable for at least 3 months. Intervention The intervention (Project HELP: Healthy Eating and Lifestyle Post-surgery) was delivered via online modules. Specifically, 10 weekly sessions were developed using an e-learning software suite (i.e., Articulate) and hosted on Coursesites (a popular e-learning platform). Module content was developed by translating material from the treatment protocol used for the inperson group program version of the intervention previously described [29]. These modules included the presentation of material using images, text, audio, and video to convey session content comprehensively. Other components included interactive exercises, examples of other Bpatients^ utilizing ABT skills in the moment, quizzes that aimed to support participants’ understanding of the material, and directed assignments to be completed throughout the week (i.e., BSkill Builders^). Participants were assigned to view each module over the course of the week (at any time they chose, with the ability to control the flow of material, including replaying and rewinding module materials). They were also asked to record their food intake daily using MyFitnessPal and to record their weights and average daily calories in an online spreadsheet each week that selfpopulated a graph to visually represent progress. This intervention focused on ABT strategies with an emphasis on willingness to experience less pleasurable (e.g., choosing low-calorie foods vs. more pleasurable calorically dense foods) and aversive internal experiences (e.g., hunger, food cravings, negative emotions). Strategies to increase willingness were taught, including defusion (i.e., gaining distance from internal experiences such that one gains the ability to act independently of them). Mindful decision-making, as it relates to eating and exercise, was also emphasized. Clarification of, and commitment to, core values was another key component, as living life in accordance with one’s values (e.g., health) makes willingness to engage in difficult weight control behaviors worthwhile. Standard behavioral techniques for weight loss (i.e., self-monitoring, stimulus control, portion control, psychoeducation) were also included in each module. These skills were framed as the core behaviors necessary for weight control, while the acceptance-based skills were presented as essential tools to enable the patients to continue to engage in these behaviors over the long term despite the difficulties in doing so (see [50] for more details on the application of these skills). Following the introduction of each ABT skill, a behavioral application section was included. In this section, participants received specific examples of utilizing the ABT skill to engage

OBES SURG

in particular weight control behaviors. Specifically, in most modules, participants were given a list of common challenges relating to the topic covered and they were able to choose which challenge was most relevant to them. They would subsequently see a specific example of a character utilizing ABT skills in response to that behavioral challenge. Participants were then prompted to plan for how they would personally use these ABT skills in a similar situation (see Fig. 1 for an example). In order to ensure that participants were understanding and applying the skills delivered, a brief (i.e., 20 min) telephone call with a member of the study team (i.e., program coach) was conducted every 2 weeks. Program coaches were advanced graduate students with at least 1 year of experience delivering ABTs for weight control. The goal of these calls

Fig. 1 Screenshots from modules

was to discuss and clarify the content of the session, discuss how the participant utilized skills demonstrated in the modules, problem-solve difficulties in utilizing the skills, and review homework. Feedback regarding weight losses and food records was also provided. This component is particularly important, as prior research conducted on Internet-based weight loss treatments has shown that such feedback results in significantly greater weight loss compared to when it is not provided [51]. A discussion board feature was also available to participants in the program. Topics were posted on the discussion board by research staff designed to facilitate social support among participants (which has been shown to be particularly important for bariatric surgery patients [52]). As the primary aim of this study was to develop a feasible and acceptable treatment, we implemented a structured plan to

OBES SURG

collect feedback from participants. After completing each module, participants rated its helpfulness, ease of use, and level of engagement on a 3-point scale (1 = BNot at all,^ 2 = BSomewhat,^ 3 = BVery^). Participants were also prompted to report technological challenges and provide free-form feedback at the end of each module. In addition, all participants completed a treatment acceptability questionnaire following the completion of the intervention to gather additional feedback.

Treatment Acceptability and Feasibility

Procedures

Eating-Related and Physical Activity Variables

Interested participants underwent a phone screen by study personnel in order to determine eligibility. Eligible participants were invited to complete the 10 weekly online modules (as well as phone check-ins with a program coach). Prior to beginning the program, participants were required to complete a brief tutorial, during which their assigned program coach would walk them through the technical components of the online program. Assessments time points included: (1) baseline (within 2 weeks prior to starting the intervention), (2) mid-treatment (after completing the fifth online module), (3) post-treatment (at the completion of the final module), and (4) follow-up (3 months following completion of the final module). Participants were compensated $15 for completing the mid-treatment assessment and $25 for the post-treatment and 3-month follow-up assessments. All assessments were conducted remotely via online questionnaires.

Caloric intake was measured with an online system for dietary self-monitoring (i.e., MyFitnessPal). Participants were instructed to complete daily food records starting at baseline through the end of the intervention. Calorie data from records at baseline and post-treatment (calculated via MyFitnessPal) were analyzed to calculate average daily caloric consumption during the first and last week of the program. Use of electronic tools has been shown to improve compliance with selfmonitoring compared to using paper records [54]. In addition to serving as an outcome variable, this system also served as the self-monitoring component of the intervention. A selfmonitoring record was considered complete for 1 day if a participant entered at least three separate meals or if he/she recorded at least 50 % of his/her daily calorie goal (as used by [55]). Loss of control eating was examined using the Eating Disorder Examination Questionnaire (EDE-Q; [56]). Because many bariatric surgery patients cannot physically consume an objectively large amount of food in one sitting, both subjective and objective binge episodes as assessed by the EDE-Q were considered loss of control episodes [57]. Good concurrent validity with the EDE interview has been established for the EDE-Q [58]. Disinhibition, restraint, and reactivity to internal and external cues were measured with The Eating Inventory [59], which has been shown to be reliable and valid [60]. Disinhibition scores have been shown to decrease with weight loss treatment as well as after bariatric surgery [7, 61], while cognitive restraint has been shown to increase post-surgery [7]. The Emotional Eating Scale (EES) was used to assess emotional eating, which is a self-report measure that assesses the relationship between overeating and negative emotions [62]. The EES is classified into three subscales (anger/frustration, anxiety, depression). This measure has good construct validity and adequate test-retest reliability [62] and has been previously used with post-bariatric surgery patients [27]. Grazing behavior over the past 2 weeks was assessed by self-report based on the definition provided by Colles et al. [9], i.e., Bthe consumption of smaller amounts of food

Measures Anthropometric Data Weight was self-measured by participants using a digital scale that they were required to purchase (if they did not already own). Prior to weighing, a scale check was then performed to assess the reliability of participants’ scales and participants were given written instructions to enhance the accuracy of self-measured weights (e.g., removing shoes or extra layers of clothing, placing the scale on a hard, level surface, weighing first thing in the morning). Participants weighed themselves three times during each assessment and the average of these weights was recorded. If recorded weights differed by >1 lb., participants were contacted to check accuracy of recording. Self-reported weights in bariatric surgery patients have been shown to be reasonably accurate, i.e., on average within 1 kg [53]. Within our sample, self-recorded weights were very reliable (i.e., SD of mean recorded weights within each subject was 0.1 lbs). Participants also self-reported their height, which was used to calculate BMI.

Treatment acceptability was measured using a questionnaire, adapted by the investigators from previous measures used for this purpose, using a 5-point Likert scale (1 = BNot at all,^ 3 = BSomewhat,^ 5 = BVery^) to evaluate how helpful participants found the treatment, their satisfaction with it, and how likely they would be to recommend it to a friend. Feasibility of the modules was assessed via a brief questionnaire following the viewing of each module, as described above.

OBES SURG

continuously over an extended period of time, eating more than the subjects considers best for them,^ (p. 616). A questionnaire was created by the researchers due to the lack of validated measures for this eating behavior [63]. In addition to confirming these criteria in the form of yes or no questions, participants were asked to what extent they grazed over the past 2 weeks (from BNever^ to BAlways^). The Food Cravings Questionnaire-Trait (FCQ-T; [64]) was used to assess food cravings, which has been shown to have excellent internal consistency (including in a bariatric surgeryseeking population, [65]) and good test-retest reliability [64]. The Paffenbarger Physical Activity Recall [66], a 15-item, interview-based measure for assessing physical activity, was used. By converting these activities into metabolic equivalents based on body mass, total expenditure from physical activity was calculated.

Acceptance-Based Process Variables The Acceptance subscale of the Philadelphia Mindfulness Scale (PHLMS) was used to assess acceptance of internal experiences [67]. Internal consistency and concurrent validity with established measures has been demonstrated [67]. Acceptance of food-related internal experiences was measured with the Food-Related Acceptance and Action Questionnaire (FAAQ), which has demonstrated good reliability and validity [68]. Defusion was measured with the Drexel Defusion Scale (DDS), a self-report measure assessing the extent of the ability to defuse from different internal experiences, which has been shown to have good internal consistency [69]. Acceptance of physical activity-related internal experiences was evaluated with the Physical Activity Acceptance Questionnaire (PAAQ). This 10-item measure assesses self-reported acceptance of psychological and physical discomfort associated with engaging in physical activity. Good internal reliability and concurrent validity has been demonstrated [70]. All measures were completed at each assessment point, except that height was reported at baseline only and treatment acceptability at post-treatment only.

Data Analysis All variables are reported as mean ± standard deviation or frequency and percentages. T tests were used to assess changes in weight and secondary variables (eating behaviors and acceptance-related constructs) pre- to post-treatment. Completer analyses and intent-to-treat analyses (using last weight carried forward for participants who provided a baseline assessment weight) were conducted for weight data.

Results Participants A total of 114 potential participants who indicated interest in the study (via a recruitment survey, Craigslist ads, flyers, and referrals from other research studies) were contacted. Of those 114, 60 were reached by phone or email and provided with further information about the study. Of those reached, 8 did not meet eligibility criteria, 20 were lost to contact, and 9 were no longer interested in treatment. The remaining 23 potential participants met eligibility criteria and agreed to join the program. Twenty of those who met criteria were enrolled in the program (i.e., completed the baseline assessment and tutorial), and 16 of those 20 interacted with the first module (i.e., treatment Butilizers^) and were included in analyses. Eleven of these participants completed the intervention and post-treatment assessment (i.e., treatment Bcompleters^). See the CONSORT diagram (Fig. 2) for details on screening and enrollment. Participant demographics are listed in Table 1. Of note, the sample was predominantly White and female. Self-reported surgery data (Table 1) indicated that the majority of participants received gastric bypass surgery (75 %), and mean time since surgery was 5.1 ± 1.0 years. Treatment Feasibility and Acceptability Program retention was defined as completing at least 8 of 10 modules. Out of the 20 participants who enrolled in the study (i.e., completed the tutorial and baseline assessment), 12 participants completed 8 modules (60.0 %). However, of the 16 participants who utilized the modules, 75 % met retention criteria defined above. The average time to program completion was 12.2 ± 4.0 weeks. Although the intervention was designed to be completed within 10 weeks, participants were unable to move onto the subsequent weekly module until they completed the previous week’s module. The first two completers took an average of 20.2 ± 0.1 weeks to complete the program (i.e., twice the intended amount of time). Following this observation, personalized schedules, checklists, and structured reminders were implemented and the remaining participants took an average of 10.4 ± 0.6 weeks to complete the intervention. The average time spent interacting with each module was 26.2 ± 10.2 min, and average time of each phone coach call was 16.5 ± 3.6 min. On average, participants answered quiz questions correctly 85.5 % of the time, indicating adequate understanding and retention of material. Program completers completed all five bi-weekly phone sessions. For program utilizers, 77.5 % of phone coach calls were completed. During the majority of the program, the discussion board was unused. Qualitative feedback indicated that the discussion board was not user-friendly, which we were unable to modify given the use of an already-developed

OBES SURG Fig. 2 Consort diagram

platform. Additional instructions were provided to assist participants and targeted posts were implemented; however, these steps resulted in only one participant using this feature. Among completers (those who completed the acceptability questionnaire, n = 11) acceptability ratings were high, including for overall helpfulness of ABT strategies (4.5/5 ± 0.8), overall satisfaction with the program (4.7 ± 0.6), and confidence in recommending the program to others (4.7 ± 0.6). The average ratings of individual modules also indicate high acceptability, with participants rating them as Bvery^ helpful 71.1 % of the time, Bvery^ easy to use 90.6 % of the time, and Bvery^ engaging 78.7 % of the time. The highest rated acceptance-based strategies (rated at least 4.5 out of 5) were acceptance, willingness, and mindful decision-making

Table 1 Participant demographics and surgery information

strategies. The helpfulness of the phone coach calls were not formally assessed; however, discussions with participants indicated high acceptability and usefulness. Participants were able to view modules 75.2 % of the time without reporting technological issues. The most commonly cited issue was user activity not being automatically recorded by the program, requiring participants to re-enter quiz responses and feedback (i.e., 21.6 %). When modules were viewed, skill builders were completed 79.8 % of the time, weekly self-monitored weights 93.8 % of the time, and weekly self-monitored average calories 85.3 % of the time. Participants only entered complete food records 67.4 % of days in the program. However, treatment completers recorded their food intake 74.8 % of the time.

Total sample

Completers

Utilized only

Enrolled only

(n = 20)

(n = 11)

(n = 5)

(n = 4)

Age (year; M ± SD) Women (%) White (%) African American (%) Married or living with partner (%) Employed full-time (%)

54.3 ± 12.1 85.0 % 80.0 % 20.0 % 55.0 % 60.0 %

50.7 ± 13.7 72.7 % 81.8 % 18.2 % 54.5 % 54.5 %

59.2 ± 5.3 100 % 80.0 % 20.0 % 80.0 % 60.0 %

58.0 ± 12.9 100 % 75.0 % 25.0 % 25.0 % 75.0 %

Gastric bypass (%) Gastric sleeve (%) Gastric banding (%) Time since surgery (year; M ± SD)

75.0 % 15.0 % 10.0 % 5.1 ± 1.0

72.7 % 27.3 % 0% 5.1 ± 1.1

80.0 % 0% 20.0 % 5.5 ± 0.7

75.0 % 0% 25.0 % 4.6 ± 1.1

Completers participants who completed all modules and assessments; Utilized only interacted with at least the first module, but did not complete the program; Enrolled only completed baseline assessment and tutorial only

OBES SURG

Weight Outcomes Among treatment completers, 10 out of 11 (i.e., 90.9 %) participants demonstrated weight stabilization (within 0.2 kg) or weight loss. Including treatment utilizers who provided at least one additional weight after beginning treatment, 13 out of 14 (i.e., 92.9 %) participants demonstrated weight stabilization or weight loss. Treatment completers demonstrated significant weight loss from pre- to post-treatment (5.1 ± 5.5 %; 5.9 ± 6.5 kg, t(10) = 3.02, p = .01). Intent-to-treat analyses also revealed significant weight loss pre- to post-intervention (3.9 ± 5.0 %; 4.4 ± 5.8 kg, t(15) = 3.05, p = .01). Percent weight losses for each treatment completer are displayed in Fig. 3, and weight trajectories since 2 years prior to starting the intervention are shown in Fig. 4. All treatment completers were assessed at 3-month followup. Weight losses were maintained, with an average additional weight loss of 0.6 ± 2.7 % from post-treatment to follow-up. Total average weight loss from pre-treatment to 3-month follow-up was 5.7 ± 6.1 %. Process Variable Outcomes Changes in acceptance-based process variables from pre-to post-treatment were generally large, in the expected direction and supported hypotheses (Table 2). However, acceptance of general internal experiences did not evidence improvement. Significant and generally medium to large improvements in eating-related variables were also observed (Table 2). In addition to the changes in scores on these validated measures, changes were also observed on items from our grazing questionnaire. In particular, the percentage of participants who endorsed problematic grazing decreased from 36.4 to 9.1 % from pre- to post-treatment. In addition, items from the EDE-Q revealed that average frequency of loss of control eating episodes decreased from 4.3 times to 0.9 times within the previous 4 weeks (M = 3.36, SD = 6.04, t(10) = 1.85, p = .09). In addition, the percentage of participants who endorsed loss of control eating episodes decreased from 63.6 to 27.3 % pre- to post-treatment.

Fig. 4 Mean weight change since surgery from 2 years pre-treatment to post-treatment. Black square self-reported weights (via a screen), black triangle self-measured weights

Based on food record data, treatment completers who provided calorie data from pre- and post-treatment (n = 8) reduced their average daily calorie intake from 1364.5 ± 342.7 to 1227.1 ± 69.6 (Mchange = 137.4, SD = 245.9, t(7) = 1.58, p = .16); however, this difference did not reach statistical significance. Self-reported caloric expenditure (based on the Paffenbarger physical activity questionnaire) did not improve; in fact, average calories expended per week evidenced a small and statistically insignificant decrease from 4048.1 to 3834.8 kcal/week (M c h a n g e = 213.3, SD = 2513.9, t(9) = 0.27, p = .80). Due to our small sample size, we were not able to conduct formal mediation analyses. However, exploratory analyses were conducted to identify potential mediators. Correlations between residualized changes in process measures from preto mid-treatment and residualized change in weight from preto post-treatment were conducted. Pre- to post-treatment weights were used for these analyses rather than mid- to post-treatments because the majority of weight loss seen in weight control interventions often occurs towards the beginning of the intervention. Residualized changes in several of our hypothesized mediators were strongly correlated with residualized changes in weight including defusion (r = −0.58, p = 0.06), disinhibition (r = 0.55, p = 0.08), reactivity to internal cues (r = 0.71, p = 0.02), eating in response to depression (r = 0.63, p = 0.04), food cravings (r = 0.54, p = 0.09), and food-related acceptance (r = −0.50, p = 0.12).

Discussion

Fig. 3 Percent weight change from pre- to post-treatment by participant

The current study provides preliminary support for the use of a remotely delivered acceptance-based behavioral intervention for weight regain after bariatric surgery. Specifically, our pilot study supported the feasibility, acceptability, and preliminary efficacy of this novel 10-module online program. Overall, the program proved to be feasible and acceptable, with the vast majority of participants rating the program as

OBES SURG Table 2 Pairwise comparisons of process measures from pre- to post-treatment

Baseline

Post-treatment

Paired t test (df = 10)

EI disinhibition EI cognitive restraint EI internal EI external

M 8.3 12.4 4.4 2.3

SD 4.2 3.2 2.6 1.8

M 5.6 17.6 2.3 1.7

SD 2.8 2.3 1.7 1.3

t 2.9 −4.08 3.0 1.60

p 0.02*