Introduction

Mindfulness meditation skills, such as those taught in mindfulness-based eating awareness training (MB-EAT) may address factors related to the problematic eating patterns and eating disorders that contribute to obesity (Kristeller and Hallett 1999). Mindfulness has been defined as “bringing one’s complete attention to the present experience on a moment-to-moment basis” (Marlatt and Kristeller 1999, p. 68) and is practiced with an attitude of acceptance and without judgment or interpretation of thoughts, feelings, and sensations (Kabat-Zinn 1994). When applied to eating, mindfulness involves attending to the sensory aspects of food such as the taste, aroma, and appearance, as well as the thoughts, emotions, and bodily sensations that accompany eating.

Obese individuals have been found to have higher rates of binge eating disorder (BED) and night eating syndrome (NES) in comparison with those who are not obese (Tanofsky-Kraff and Yanovski 2004; Urquhart and Mihalynuk 2011). These disorders or subclinical presentations of these disorders can contribute to weight gain. Specific factors related to weight gain include eating to soothe painful emotions, stress, limited interoceptive awareness, and prioritizing external versus internal cues when making decisions about eating (Mitchell et al. 2016). Difficulty coping with painful emotions can be an underlying cause of overeating and binge eating (Heatherton and Baumeister 1991). Stress has been linked to disordered eating and obesity (Dallman 2010; Sinha and Jastreboff 2013), especially in emotional eaters and obese individuals (Adam and Epel 2007; Torres and Nowson 2007). Stress and aversive emotions can increase preference for foods that are calorie-dense and high in far, sugar, and/or salt (Dallman 2010; Sinha and Jastreboff 2013). Limited interoceptive awareness may be another factor associated with problematic eating and obesity. Individuals who are unable to accurately differentiate internal cues of hunger from other physical processes may unknowingly eat when they are not physically hungry (Craighead and Allen 1995; Ganley 1989; Mitchell et al. 2016). Cues in the environment can contribute to mindless eating and overeating. External eating (the tendency to eat in response to sensory cues for eating) has been associated with obesity (Blair et al. 1990; Delahanty et al. 2002). Typical diets may exacerbate this problem, as they consist of externally dictated rules for what, when, and how much to eat, potentially contributing to the loss of one’s ability to recognize, accept, or respond to internal cues of hunger, taste, satiety, and fullness (Ganley 1989).

The body of research evidence for mindful eating interventions is small in terms of the number of published studies, but results are favorable. MB-EAT follows a 10-session protocol that has been used in several published studies (e.g., Kristeller and Hallett 1999; Kristeller et al. 2013; Miller et al. 2014). As described by Kristeller and Epel (2014), the purported mechanisms of change for MB-EAT are improved self-regulation: first, if individuals can improve awareness of internal physical signals such as hunger and fullness, they will make more skillful decisions regarding when and how much to eat, and second, mindfulness skills may boost the pre-frontal cortical processes that help overcome the tendency to undervalue postponed rewards, like weight maintenance, or improved health that may be a consequence of not acting on impulses to eat high calorie, nutrient-poor foods. Studies of mindful eating interventions other than MB-EAT that have been published appear to incorporate common elements, including mindful eating practices, mindfulness meditation, information about nutrition, and mindful yoga or walking. Recent reviews of mindful eating interventions range from 1 to 24 sessions with 7–150 participants and include both patient and non-patient participants who are overweight or obese. These reviews show significant improvements in binge eating and emotional eating (Godsey 2013; Katterman et al. 2014; O’Reilly et al. 2014).

Kristeller and Hallett (1999) used MB-EAT with 18 overweight/obese women (average BMI = 40) with BED. They found a significant improvement in participants’ perceived control of eating and awareness of hunger and satiety cues. Participants who practiced mindfulness between sessions more frequently and for longer periods of time experienced better outcomes. In MB-EAT, participants are encouraged to practice daily. In a follow-up study, Kristeller et al. (2013) conducted a randomized controlled trial (RCT) of MB-EAT for 150 individuals with an average BMI of 40.3, 66% of whom met DSM-IV-TR (American Psychiatric Association 2000) criteria for BED. The comparison groups were a psychoeducational/cognitive behavioral intervention or a wait list control. The MB-EAT and psychoeducation participants showed comparable improvement after 1 and 4 months post-intervention in bingeing and depression. At 4 months post, 5% of those who had met criteria for BED in the MB-EAT group continued to meet criteria for BED, compared with 24% of those in the psychoeducation group. The amount of mindfulness practice between sessions predicted weight loss and other outcomes. Larger effect sizes were found for the MB-EAT group as compared to the psychoeducation group on measures of disinhibition and hunger, indicating greater self-regulation and behavioral control in the MB-EAT group.

Other researchers investigated mindful eating interventions that did not adhere to the MB-EAT protocol or that involved mindful eating interventions developed for specific settings or studies. Daubenmier et al. (2011) assessed a mindful eating intervention based on MB-EAT with 47 overweight/obese women who were randomly assigned to the 4-month intervention or a waitlist control, to explore the effects on stress eating. Measures included self-report questionnaires as well as cortisol levels. Participants improved in mindfulness, anxiety, and externally based eating but did not differ on average cortisol awakening response, weight, or abdominal fat over time. Mindful eating interventions have also been implemented in community settings (Bush et al. 2014). A 10-week group intervention integrating mindfulness and intuitive eating skills was delivered to 124 female university employees, excluding those with diagnoses of anorexia or bulimia. The goal of the intervention was to reduce body dissatisfaction and decrease problematic eating behaviors. Participants in the intervention as compared to wait list controls reported higher levels of body appreciation and lower levels of problematic eating. In addition, mindfulness scores served as a partial mediator of change in outcomes. Finally, Kidd et al. (2013) administered an 8-week mindful eating group intervention with the purpose of investigating changes in self-efficacy for weight loss, depression, body fat, and blood pressure in 12 obese women living in an urban area. A focus group was conducted following the intervention to understand the participants’ experiences with mindful eating. The variable that improved statistically on self-report measures was self-efficacy in regard weight loss. Thematic analyses of focus group content confirmed increased self-efficacy over weight loss, and the participants described improvements in mood, food choices, and eating behavior.

In response to increasing obesity rates, the numbers of bariatric surgeries performed have increased dramatically, and laparoscopic techniques have decreased the risk of surgical complications (Nguyen et al. 2011). Bariatric surgery has consistently been found to be the most effective treatment for sustained weight loss and reduction of related medical conditions for obese individuals (Chang et al. 2014). Obese individuals who pursue bariatric surgery report high rates of problematic eating, including loss of control over eating and night eating (Royal et al. 2015), as well as binge eating and chronic overeating (Kalarchian et al. 1998). Greater mindfulness was related to fewer binges and less emotional eating, as well as less habitual overeating and grazing, in a study of 820 patients seeking bariatric surgery (Levin et al. 2014). Even when individuals presenting for bariatric surgery do not meet criteria for a clinical eating disorder, these disordered eating patterns may prevent optimal adherence to post-surgical eating guidelines. Given that weight loss with bariatric surgery is associated with the improvement or resolution of medical comorbidities as well as improvements in patient-reported quality of life, weight regain is of great concern (Hachem and Brennan 2016).

The first year after surgery for most patients is characterized by rapid weight loss and patient satisfaction, but once patients transition from the active weight loss phase to weight maintenance, pre-surgery eating problems may recur. In one study following gastric bypass patients over an average of 28 months, 79% regained some weight after reaching their lowest weight and 15% experienced a weight increase of 15% or more from their lowest weight (Odom et al. 2010). Eating-related factors associated with weight regain following bariatric surgery have been reported in numerous research studies since the 1990s. Binge eating and loss of control over eating have repeatedly been found to be related to weight gain (Mitchell et al. 2001; Odom et al. 2010). A recent review reported that in 14 of the 15 studies examined, binge eating and loss of control reported post-operatively was associated with less weight loss or more weight regain (Meany et al. 2014). Other factors associated with weight gain are eating in response to painful emotions, regularly eating past the point of fullness and eating continuously throughout the day (Grothe et al. 2006; Mitchell et al. 2016; Zimmerman et al. 2007). In a study that followed bariatric surgery patients over 10 years, disinhibition over eating (i.e., eating in response to various internal and external cues instead of physical hunger) was associated with weight gain (Kottinen et al. 2015). Patients at the highest risk for weight loss failure may need additional support as early as 6 months to 1 year following surgery (Courcoulas et al. 2013).

Preventing weight regain and supporting weight maintenance is of utmost importance for bariatric surgery patients. Although weight loss is the most commonly measured variable in outcome studies of bariatric surgery, the associated improvements in physical health, psychosocial functioning, and quality of life can be profound (Magallares and Schomerus 2015). In recent reviews of psychosocial interventions for post-bariatric surgery patients, behavioral lifestyle interventions were associated with greater weight loss, though these findings must be interpreted with caution, because the interventions did not follow standardized treatment manuals, were delivered by various types of health professionals, and small sample sizes were studied (Rudolph and Hilbert 2013; Stewart and Avenell 2016). Mindfulness-based cognitive therapy (MBCT), developed by Segal et al. (2002), is an intervention shown to be effective in preventing depression relapse in individuals with recurrent depression (Piet and Hougaard 2011). An intervention that could similarly prevent relapse in terms of weight regain associated with problematic eating in individuals who undergo bariatric surgery would thus be of great value.

The published research on mindful eating group intervention studies with post-bariatric surgery participants is very limited (Chacko et al. 2016; Leahey et al. 2008). While the problematic eating behaviors exhibited by obese individuals who undergo surgery are similar to those who do not undergo surgery, and the same interventions may thus be effective, bariatric surgery patients are unique in that surgical alterations lead to changes in eating that are not shared by non-surgical patients. For example, bariatric surgery patients commonly experience abdominal discomfort after eating foods like bread and pasta, and experience dumping syndrome, characterized by nausea, fatigue, and dizziness, after consuming foods high in sugar or fat. During psychosocial interventions targeting eating, it is expected that these experiences will be raised by participants. Thus, there is a strong clinical and research rationale for investigating mindful eating interventions with a homogenous bariatric surgery patient group, rather than in a mixed group of surgical and non-surgical participants.

Chacko et al. (2016) randomly assigned nine participants to a mindfulness intervention and nine to a standard intervention. The mindfulness intervention integrated aspects of mindfulness-based stress reduction (Kabat-Zinn 1982) and MB-EAT with traditional behavioral strategies for obesity including goal setting, problem-solving, self-monitoring, and social support and was delivered over ten consecutive weekly classes of 90 min each. The aim of this intervention was to improve coping skills so as to support long-term weight maintenance. The standard intervention was a 1-h individual counseling session with a registered dietitian. No significant group differences in weight were found between groups at 12 weeks and 6 months post-baseline. A decrease in emotional eating after the mindfulness intervention that was statistically significant at 6 months was reported. Interestingly, participants in the mindfulness group reported experiencing more stress after the intervention compared to before the intervention, but were less reactive to the stress.

Leahey et al. (2008) conducted a group that incorporated cognitive behavior therapy (CBT) strategies and mindful eating practice over 10 weekly 75-min sessions in a hospital setting. The intervention incorporated CBT strategies such as keeping an eating journal that included eating-related thoughts and feelings, and removing triggering foods from the home. Mindful eating practices were facilitated in each session to improve awareness of reactions to food and eating. The group was composed of seven patients who had either undergone Roux-en-y gastric bypass (RYGB) or gastric banding and who reported subjective binges with loss of control and eating to manage emotions. Post-intervention, patients reported improvements in eating, emotion regulation, and depression, changes that theoretically should help patients reduce problematic eating and thus prevent weight regain.

The research reviewed so far supports the effectiveness of MB-EAT and other mindful eating interventions for overeating and related problematic eating patterns associated with weight gain, and evidence-based interventions for preventing weight regain after bariatric surgery are needed. Therefore, the objective of the current study was to investigate the feasibility and outcomes of MB-EAT for post-bariatric surgery patients. We hypothesized that participants would maintain their BMI and show improvements in eating psychopathology, depression, anxiety, mindfulness, and emotion regulation immediately following the intervention and at 4 months follow-up.

Method

Participants

Our goal was to enroll 30 participants recruited from a patient group (n = 2396) that underwent the RYGB or sleeve gastrectomy (SG) between 2009 and 2013 at the Toronto Western Hospital Bariatric Surgery Program (TWH-BSP). The Toronto Western Hospital is a large academic hospital within the University Health Network (UHN), a multi-site network of research and teaching hospitals in Toronto, Canada. Approximately 90% of bariatric surgeries performed at the TWH-BSP are the RYGB, and the SG is only offered if surgically indicated. Patients are followed by the TWH-BSP for 5 years post-surgery. Within the first year post-surgery, patients are scheduled to meet with members of the multi-disciplinary healthcare team at 1, 3, 6, and 12 months and yearly thereafter up to 5 years. Patients were included in this study if they were 1 year or more post-bariatric surgery, aged 18 or older, and experiencing self-reported emotional or other overeating related difficulties adhering to post-surgery eating guidelines. Exclusion criteria were active suicidal ideation, serious mental illness (i.e., psychotic disorder, bipolar disorder), substance use disorder, severe depression (defined as score of 20 or more on the patient health questionnaire) (PHQ-9) (Kroenke et al. 2001), or anxiety (defined as a score of 15 or greater on the generalized anxiety disorder questionnaire) (GAD-7) (Spitzer et al. 2006). Interested patients were screened by a research assistant over the phone. Potential participants were informed of the study via posters in the clinic, announcements at the twice monthly patient support group, and by individual clinicians at routine follow-up appointments.

Recruitment occurred over a period of 9 months. Of the 46 individuals who self-referred to the study based on recruitment materials and methods, 28 were recruited. Please see Fig. 1, CONSORT diagram for details of the participant flow. Participants were considered non-completers if they attended fewer than five sessions. Completers (76%, n = 22) on average attended 7.05 (SD = .10) sessions, while non-completers (21%, n = 6) attended on average two sessions (SD = .63). Of the completers, 9.1% attended five sessions, 18.2% attended six sessions, 31.8% attended seven sessions, and 40.9% attended all eight sessions. Of the non-completers, 16.7% attended only one session, 66.7% attended two sessions, and 16.7% attended three sessions. Three separate groups were run: June–August 2014 (n = 8), September–November 2014 (n = 7), and January–March 2015 (n = 7). Study completers and non-completers did not differ significantly on any pre-intervention variables, p ≤ 0.05.

Fig. 1
figure 1

CONSORT diagram of participant flow

Full size image

Participants’ mean age was 55.41 years (SD = 9.44). The sample was 100% female and 21 of the 22 participants identified as Caucasian (95%). In terms of marital status, 11 (50%) were married or cohabiting, eight (36.4%) had never been married, and three (13.6%) were separated or divorced. Participants were generally highly educated: 14 (68.2%) had completed a college diploma or university degree, and two (9%) completed some college or university courses. Two (9.1%) were high school graduates, and three (13.6%) had not completed high school. Participants’ mean BMI was 32.82 kg/m2 (SD = 5.31), in the range of Class I obesity, with 21 having undergone a RYGB and one a SG. The average time since surgery was 36.36 months (SD = 20.96, min = 15, max = 114). Comparable pre, post, and follow-up survey data exists for between 13 and 17 participants depending on the measure. The reason for missing data was lack of participant completion of measures despite repeated attempts to contact by the research assistant.

Procedure

A single-group pre-post treatment design with a follow-up of 4 months to establish minimal stability of symptoms was used. Given the lack of previous data on MB-EAT for bariatric surgery patients, this exploratory approach was a first step in providing information about basic treatment feasibility. A similar design was used to explore effects of MB-EAT for treating binge eating disorder (Kristeller and Hallett 1999) and in other studies of mindful eating for obese individuals (Dalen et al. 2010) and bariatric surgery patients (Leahey et al. 2008). The study was approved by the UHN Research Ethics Board. Informed, signed consent was obtained from participants, and they were provided with a copy of the consent form.

The 8-session MB-EAT protocol for pre-surgery bariatric patients was used with minor modifications to accommodate the typical dietary restrictions encountered by post-surgery patients (Kristeller et al. 2010). For example, due to abdominal discomfort experienced by many post-surgery patients when consuming sugar and fat, sugar-free/fat-free food items were made available for mindful eating exercises that involve sweet foods. Each session was 2 h long. The overarching aim of MB-EAT is to help group members cultivate the ability to integrate their own food preferences, physical, and emotional needs (“inner wisdom”) with knowledge about healthy eating (“outer wisdom”), when making decisions about eating. Each MB-EAT session includes experiential and didactic components. Facilitators provide instruction on a variety of mindfulness practices lasting from 5 to 20 min including breath meditation, body scan, self-compassion, and self-forgiveness meditations as well as mindful eating practices with small portions of different types of food. Walking meditation practice and gentle chair yoga are introduced and facilitated as mindfulness practices. Session 6 involves a potluck so that members can practice mindful eating in a social setting. Time is set aside each session for participants to share homework, ask questions, and discuss their experiences. Nutrition education and a behavioral chain analysis are incorporated. A chain analysis is a therapeutic technique used in cognitive behavior therapy and dialectical behavior therapy that is a detailed analysis of the thoughts, emotions, behaviors, and environmental factors associated with a problematic behavior. This analysis can be helpful in identifying solutions for preventing the behavior in similar future situations (Linehan 2015).

Participants were assigned homework between sessions that involved guided mindfulness practice audio recordings, readings, and worksheets. The mindfulness recordings were created by the first and second author. Mindfulness practiced ranged in duration from 10 to 20 min and consisted of formal sitting meditation including breath awareness, body scan, and self-forgiveness that were to be practiced 6 days per week. Shorter practice recordings were assigned in early sessions and became longer over the course of the group as participants gained experience with mindfulness. Mindfulness practice homework for the following week was described at the end of each session so that participants were clear about expectations. Audio recordings to be used for the coming week were emailed to participants after each session. Homework also included instructions to practice mindful eating with meals and snacks. Readings were one-page handouts describing various aspects of mindful eating, while worksheets were provided on such topics as identifying events that triggered emotional eating. For a more detailed description of MB-EAT, please see Kristeller and Wolever (2011).

All MB-EAT groups were co-facilitated by the first two authors, both of whom completed the MB-EAT facilitator training with J. Kristeller and A. Lieberstein. SW is a clinical psychologist with 7 years of clinical experience with bariatric surgery patients and 12 years of clinical experience with mindfulness interventions, Dialectical Behavior Therapy, and eating disorders. CD is a psychometrist (a clinician with a master’s degree in psychology who is supervised by a psychologist) with 5 years of clinical experience with bariatric surgery patients. She has completed training in mindfulness-based cognitive therapy (Segal et al. 2002) training and MBSR (Kabat-Zinn 1982) and has facilitated groups according to both protocols. Both facilitators have long-term personal meditation practices in Buddhist traditions, engage in ongoing study of Buddhist meditation, psychology, and philosophy, and participate in regular meditation retreats.

Measures

Demographics (age, gender, ethnicity, education, work status, and marital status) were collected, and BMI was calculated immediately prior to the first session. BMI was calculated again at the last session and at 4-month follow-up. Four-month weights were self-reported rather than measured at the clinic.

Eating psychopathology was assessed with the binge eating scale (BES) (Gormally 1982) and the emotional eating scale (EES) (Arnow et al. 1995). The BES is a 16-item self-report measure designed specifically for use with individuals with obesity that assesses binge eating behaviors (e.g., amount of food consumed) and associated cognitions and emotions (e.g., guilt, shame). Scores on the BES are categorized so that scores of less than 17 are considered “non-bingeing”; “moderate bingeing” refers to scores between 18 and 26, and “severe bingeing” refers to scores 27 and greater (Gormally 1982). The EES is a 25-item self-report measure that assesses the tendency to cope with painful emotions by eating. Respondents are presented with 25 emotions and are asked to rate the strength of their urge to eat on a scale from 0 (no desire to eat) to 4 (an overwhelming urge to eat) when experiencing each emotion. The EES results in three subscales that summarize emotions indicating eating in response to anger, anxiety, and depression. The BES and EES have been used previously in bariatric surgery patient populations (Shakory et al. 2015; Sockalingam et al. 2016). Clinical cutoffs for the EES were determined as per the work of Schneider et al. (2012), based on Arnow et al.’s (1995) article which examined the EES in individuals reporting binge eating. Participants were classified as high in EES anger if their score was > 25.4, high in EES anxiety if they scored > 15.9, and high in EES depression if they scored > 12.5.

Depression was assessed via the PHQ-9 (Kroenke et al. 2001), a 9-item self-report measure of depression severity. Respondents are asked to rate the frequency with which they have experienced depressive symptoms over the last 2 weeks on a scale ranging from 0 (not at all) to 3 (nearly every day). Scores on the PHQ-9 can range from 0 to 27 with higher scores indicating more depression. Scores of 10 and above signify clinically significant depression (Manea et al. 2012). Anxiety was assessed with the GAD-7 (Spitzer et al. 2006), a 7-item self-reported measure of anxiety severity. It was developed to assess generalized anxiety disorder but has proved to be a good screening instrument for other anxiety disorders including panic disorder, social phobia, and post-traumatic stress disorder. Respondents are asked to rate the frequency with which they experienced anxiety symptoms over the last 2 weeks on a scale ranging from 0 (not at all) to 3 (nearly every day). Scores on the GAD-7 range from 0 to 21, and mild, moderate, and severe levels of anxiety symptoms correspond to cut-off scores on 5, 10, and 15, respectively. Both the PHQ-9 and GAD-7 have been used to measure symptoms in bariatric surgery patients (Cassin et al. 2013; Sockalingam et al. 2016).

Mindfulness was measured with The Five Facets of Mindfulness Questionnaire (FFMQ) (Baer et al. 2006). The FFMQ is a 39-item self-report measure with items scored from 1 (never or very rarely true) to 5 (very often or always true). A total score as well as five subscales can be calculated: observing, describing, acting with awareness, nonjudging, and nonreactivity. We chose a global measure of mindfulness based on other published studies of mindful eating research, the majority of which used global mindfulness measures and specific eating measures. All measures of mindfulness have limitations (e.g., Bergomi et al. 2013; Sauer et al. 2013), and we chose the FFMQ, because it is widely used across mindfulness research.

General emotion regulation was assessed with the difficulties in emotion regulation scale (DERS) (Gratz and Roermer 2004). The DERS is a 36-item questionnaire that measures general difficulties regulating emotions rather than in response to specific events. Items are rated on a 5-point scale with higher scores indicating better emotion regulation. A total score and six subscales can be computed: nonacceptance of emotional reactions; difficulties engaging in goal directed behavior; impulse control difficulties; lack of emotional awareness; and limited access to emotion regulation strategies. Emotion regulation has previously been studied in bariatric surgery patients (Taube-Schiff et al. 2015).

Participants were administered brief measures at the end of each session. A one-item measure to evaluate participants’ overall rating of the session (“How helpful was the session you just completed?”) was developed for this study and used a scale of one (less than helpful) to seven (completely helpful). Participants recorded the number of days per week and minutes per day they practiced mindfulness in the previous week, except for the first session. They were not asked to differentiate between mindful eating practice and other types of mindfulness practices. Previous research on mindfulness interventions has demonstrated the relationship between time spent in mindfulness practice and outcomes (Carmody and Baer 2008; Kristeller and Hallett 1999; Kristeller et al. 2013).

Data Analyses

All analyses were performed using R, version 3.2.2 (R Development Core Team 2008). Descriptive statistics including means, standard deviations, and frequency counts were calculated to describe participant characteristics. Data analyses were conducted with study completers, because we were interested in the effects of the protocol as delivered. We chose not to conduct an intent-to-treat (ITT) analysis, as these are best suited to randomized controlled trials as a means of providing comparisons between groups for all participants regardless of treatment completion (Ten Have et al. 2008). The ITT approach estimates the effects of allocating an intervention in practice, while we were specifically interested in evaluating the effects of MB-EAT on the subgroup of participants who received the intervention as per protocol. This approach is appropriate when the primary goal is exploratory and to evaluate new treatments (Hollis and Campbell 1999).

Differences between pre, post, and 4-month follow-up outcome measures were examined with paired samples t tests. Due to the small sample size, significance levels of p < .10 are noted. Cohen’s d was calculated as a measure of the effect size for paired samples t tests by dividing the mean difference by its standard deviation, with a d value between 0 and 0.3 being considered a small effect size, between 0.3 and 0.6 a moderate effect size, and over 0.6 a large effect size. Clinical significance was assessed by examining the percentage of patients who met clinical severity cut-offs for each outcome measure, where this information was available.

Results

We analyzed outcomes with paired sample t tests from pre- to post-group (see Table 1) and from pre to 4 months post-group to examine durability of changes (see Table 2). Changes in eating psychopathology, depression, anxiety, emotion regulation, emotional eating, and mindfulness are reported.

Table 1 Pre- and post-MB-EAT scores with paired samples t tests for completers
Full size table
Table 2 Pre and 4 months post-MB-EAT scores with paired samples t tests for completers
Full size table

First, from pre- to post-MB-EAT, a trend toward improvement was found for binge eating, t(17), p = 0.08. Pre-MB-EAT, 47.6% of participants were in the “non-bingeing” category, 23.8% were in the “moderate bingeing” category, and 28.6% were in the “severe bingeing” categories. At post-MB-EAT, 64.7% were in the “non-bingeing” category, 23.5% in the “moderate bingeing” category, and 11.8% in the “severe bingeing” category. Statistically significant improvements were found for depressive symptoms as measured by the PHQ-9, t(17) 2.29, p = 0.04, Cohen’s d = 0.39, with 28.6% of participants scoring above the clinical cut-off for depression pre-MB-EAT, and 17.6% above the cut-off post-MB-EAT. In terms of anxiety as measured by the GAD-7, significant changes were not found, t(17) 1.68, p = 0.11, Cohen’s d = 0.42. Pre-MB-EAT, 61.9% of participants did not meet threshold for clinically significant anxiety, 23.8% were in the “mild anxiety” category, and 14.3% in the “moderate anxiety” category. Post-MB-EAT, 82.4% did not meet threshold for clinically significant anxiety, and 17.6% met the cut-off for “mild” anxiety. Emotional eating as measured by the EES yielded mixed findings depending on the subscale. For emotional eating in response to anxiety, t(17) 1.94, p = 0.07, Cohen’s d = 0.44; emotional eating in response to depression, t(17) 1.85, p = 0.09, Cohen’s d = 0.38; and emotional eating in response to anger, t(17) 1.37, p = 0.19, Cohen’s d = 0.25. Pre-MB-EAT, 42.9% of participants scored high in EES anger, 76.2% were high in EES anxiety, and 61.9% were high in EES depression. Post-MB-EAT, 35.3% of participants scored high in EES anger, 47.1% were high in EES anxiety, and 41.2% were high in EES depression. Significant changes were not found for difficulties in emotion regulation as measured by the DERS, t(16) 1.28, p = 0.22, Cohen’s d = 0.15. Clinical cut-offs have not been established for the DERS.

From pre-MB-EAT to the 4 month follow-up period, significant changes were not found for binge eating, t(17), p = 0.16. At the follow-up, 43.8% of participants were in the “non-bingeing” category, while 25% were in the “moderate bingeing,” and 31.3% were in the “severe bingeing” category. Significant improvements were not maintained in depressive symptoms as measured by the PHQ-9, t(17) 0.43, p = .76, Cohen’s d = 0.06, with 17.6% of participants scoring above the clinical cut-off for depression. Anxiety as measured by the GAD-7 continued to show no significant change, t(17), p = 0.49, Cohen’s d = 0.12. At follow-up, 55.6% of participants did not meet threshold for clinically significant anxiety, 33.3% were in the “mild anxiety” category, and 11.1% in the “moderate anxiety” category. Emotional eating as measured by the EES again yielded mixed findings depending on the subscale. For emotional eating in response to anxiety, t(17) 1.98, p = 0.07, Cohen’s d = 0.28; emotional eating in response to depression, t(17) 1.07, p = 0.30, Cohen’s d = 0.26); and emotional eating in response to anger, t(17) 1.34, p = 0.16, Cohen’s d = 0.25. In regard to clinical significance, 31.3% of participants scored high in EES anger, 68.8% were high in EES anxiety, and 43.8% were high in EES depression. Significant changes were found for difficulties in emotion regulation as measured by the DERS, t(2.26) p = 0.04, Cohen’s d = 0.12).

Participants rated the intervention as highly helpful, m = 6.59/7 (SD = .73) across all eight sessions, with ratings increasing each session. The most highly rated session was session 6, which involved a potluck so that participants could practice mindful eating in a social setting.

We investigated whether self-reported time spent practicing mindfulness each week was associated with outcomes from pre to post-MB-EAT and from pre to 4 months post-MB-EAT. Participants reported practicing mindfulness on average 3.51 days per week (SD = 2.04, range = 7), for an average of 16.60 min per session (SD = 12.41, range = 50). From pre- to post-MB-EAT, time spent practicing mindfulness was significantly associated with a decrease in emotional eating in response to anger (p = 0.05) and reached a near-significant relationship with improvement in binge eating (p = 0.06) and depression (p = 0.06). At 4 months follow-up, time spent practicing during the course of MB-EAT was associated with trends to improvement in emotional eating in response to anxiety (p = 0.07) and mindfulness (p = 0.08). We did not collect data on participants’ mindfulness practice during the 4-month follow-up period itself.

Discussion

The results of this pilot feasibility study support the use of MB-EAT as an intervention to prevent weight recidivism and improve psychosocial functioning post-bariatric surgery. MB-EAT appears to be an effective means for addressing the types of problematic eating behaviors and related psychological factors that can interfere with successful weight loss maintenance. This study was designed to examine the effects of a mindfulness-based eating intervention with post-bariatric surgery patients that used an established protocol with 4 months post-intervention follow-up. From pre- to post-MB-EAT, participants maintained their BMIs, reported statistically significant improvements in depression and improvements in binge eating and emotional eating in response to anxiety and anger. Effect sizes were moderate for these outcomes. Time spent practicing mindfulness between sessions varied among participants, but the more participants practiced, the greater the improvements in binge eating, emotional eating in response to anger, and depression. Furthermore, participants rated each session favorably, and the ratings increased over time, supporting the feasibility and acceptability of MB-EAT from the participant perspective. Although 46 patients contacted the research assistant to express interest in participating, only 28 enrolled in MB-EAT. Of the 18 patients who were excluded or chose not to participate, 15 did not respond to the research assistant’s repeated attempts to reach them. The reasons for lack of enrollment were not collected. It may be that the demands of voluntarily attending an 8-week program and completing homework were overly burdensome for some patients.

With a mean age of 54.29, participants in this study were older compared with other published studies of bariatric surgery patients that report a mean age between 40 and 45 years, and all participants were female, compared to more typical gender ratios of 70–80% (Chang et al. 2014). We did not restrict our study to women, and the reasons for the lack of male participants are unclear. Research on the effects of gender on outcomes in mindfulness-based interventions is limited, but one study showed that women may benefit more in terms of greater improvements in anxiety and depression compared to men (Rojiani et al. 2017). The authors speculated that this may be due to gender differences in emotion regulation techniques and the association of mindfulness with decreasing rumination, a cognitive process more prevalent in women.

As our participants were entirely female, our findings may not be applicable to male bariatric surgery patients. Most of our participants were Caucasian and thus belonged to the dominant ethnic group of the region. Other characteristics that we did not measure such as religion and culture may have played a role in outcomes. Previous studies of mindfulness-based interventions suggest that these characteristics should be taken into account when designing or adapting protocols to specific groups (Amaro et al. 2014; Sibinga et al. 2011).

The findings from pre-MB-EAT to 4 months post-MB-EAT suggest that some of these positive outcomes may be enduring. Statistically significant improvements in emotion regulation were reported, and the improvements in emotional eating in response to anxiety immediately post-MB-EAT were maintained 4 months post-intervention. We found that time spent practicing during MB-EAT was associated with reduced emotional eating in response to anxiety and with increased mindfulness at follow-up. Finally, no changes in BMI were found from pre to 4 months post-MB-EAT. Including the 8-week-long intervention, this is a total period of nearly 6 months.

Statistically significant changes were not found in mean binge eating as measured by the BES. This may be due to the relatively low mean score for this measure at baseline (i.e., 17.65). The BES is a commonly used scale in the mindful eating literature and comparisons between studies illustrate notable pre-intervention differences in BES scores that are likely due to characteristics of the population being studied. For example, Dalen et al. (2010) reported a mean score of 16.2 (SD = 5.4) pre-intervention with a statistically significant decrease to 9.5 (SD = 5.1) after their 6-week-long group intervention for 10 obese adult patients. Barnes and Kristeller (2016) reported a mean BES score of 9.9 (SD = 7.20) pre-MB-EAT and 10.9 (SD = 7.3) post-MB-EAT for 18 adolescents, a non-significant difference. Finally, Kristeller et al. (2013) reported a BES score of 28.98 (SD = 7.78) pre-MB-EAT and 15.24 (SD = 8.06) post-MB-EAT for 53 adults. These notable differences in scores between studies make it difficult to draw conclusions or generalize between populations. In the current study, examination of changes in clinical cut-off scores revealed that improvements in binge eating immediately post-MB-EAT were not maintained at follow-up. Post-MB-EAT, 64.7% of participants were in the “non-bingeing” category, but this decreased to 43.8% at follow-up. While only 11.8% of participants were in the “severe bingeing” category post-MB-EAT, this increased to 31.3% at follow-up. A longer intervention or the incorporation of other therapeutic techniques may be needed maintain changes to eating behavior. For future research with bariatric surgery patients, measures that assess control over eating such as the Three-Factor Eating Questionnaire (Stunkard and Messick 1985) or the Dutch Eating Behaviors Questionnaire (van Strien et al. 1986) may also be useful.

Surprisingly, participants in this study did not report increases in mindfulness as measured by the FFMQ (Baer et al. 2006). This finding contrasts with most mindful eating intervention studies that do show improvements in mindfulness (e.g., Bush et al. 2014; Dalen et al. 2010; Daubenmier et al. 2011) but are consistent with the results of another study, which likewise did not show improvements in mindfulness (Kidd et al. 2013). In a review of mindful eating interventions specifically for weight loss, Olson and Emery (2015) noted that research investigating the relationship between changes in mindfulness and weight loss remains inconclusive in that improvements in mindfulness do not always occur and do not always appear to be related to improvements in eating behavior or weight loss. Based on the current literature, improvements in mindfulness using available measures do not support change in mindfulness as a mechanism of change in outcome.

The results from this small pilot study are consistent with the overall intervention outcome literature on mindful eating for individuals who are overweight/obese and report overeating and emotional eating (e.g., Kristeller et al. 2006; Kristeller et al. 2013, Wolever and Best 2009). Our results echo those from another published pilot study on an integrative CBT mindful eating group for post-bariatric surgery patients that similarly found improvements in depression, emotional eating, and emotion regulation (Leahey et al. 2008). The improvements in emotion regulation and emotional eating reported by the participants in this study support the theoretical and applied research findings examining the links between emotion regulation, eating behavior, and weight (Heatherton and Baumeister 1991; Taube-Schiff et al. 2015; Whiteside et al. 2007). Further research with a larger sample is needed to replicate these findings, to understand mechanisms of change and to increase durability of changes in binge eating.

Limitations

The small sample size in the current study, lack of a control group, and reliance on self-report measures means we can draw only tentative conclusions. Although the design was appropriate given the early stages of development of mindful eating interventions for bariatric surgery patients, these limitations resulted in lack of statistical power, contributing to difficulty determining the degree of change related to MB-EAT versus bariatric surgery or other confounding factors. Since not all of our participants completed questionnaires at every time point, incomplete data may have contributed to this limitation. We anticipate that a larger sample size would result in more statistically significant improvements in the measures used in the current study, given trends toward statistical significance in several outcomes. Important variables were not controlled, such as psychiatric comorbidity, although recent research suggests that psychiatric comorbidity is not a significant predictor of bariatric surgery outcomes (Thomson et al. 2016). We did not assess expectancy effects to rule out the contribution of placebo effects to measured changes in outcomes. The measures we used may not have been sufficiently precise to capture the constructs under investigation, or participants’ expectancy of improvement may have been related to our findings. We did not collect information about participants’ previous exposure to or experience with contemplative practices, and this may have influenced intervention adherence and outcomes. We cannot rule out therapist effects, given that the same two therapists delivered the treatment. Although the facilitators received formal training in MB-EAT and followed the protocol, a treatment fidelity check was not done. The participants were enrolled in a hospital-based bariatric surgery program and were scheduled for yearly follow-up appointments with an interdisciplinary team. It is possible that these appointments occurred for some participants during the study period, and this may have contributed to the continued improvement we observed in some outcomes. We did not collect information about mindfulness practice at follow-up or informal mindfulness practice. Participation was voluntary, and participants who enrolled may have had characteristics that are not shared by the majority of bariatric surgery patients, limiting generalizability of our findings. Mindfulness-based interventions, like many psychotherapeutic interventions, are complex, integrative, and consist of various elements. It is difficult to identify the specific mechanisms of change. For example, group support, an element that is common to all group psychotherapies, may have contributed to the observed improvements in addition to specific mindfulness interventions and practice. Future research should include qualitative methods and other relevant designs to investigate these mechanisms.