Introduction

Emotional and/or binge eating could be a problematic strategy for managing unpleasant feelings among patients pursuing services for weight management, and is disproportionately common among individuals living with obesity as compared to the general population [1]. This reliance on food as a source of comfort can have a deleterious impact on weight management outcomes. Affect regulation theory suggest that individuals who lack more adaptive coping strategies use food to distract from uncomfortable emotions—a behaviour which is subsequently reinforced and/or maintained by the accompanying (though temporary) reduction in negative affect [2]. Given the relationship between emotional and/or binge-type eating, affective dysregulation, and obesity [3, 4], it stands to reason that interventions aimed at addressing the underlying factors which contribute to problematic eating behaviours are necessary to optimize weight management outcomes.

Dialectical behaviour therapy (DBT) was developed by Linehan [5] as a treatment for borderline personality disorder (BPD). Affective dysregulation is a core feature of BPD, which involves an over-reliance on impulsive, maladaptive behavior in attempt to cope with emotional distress [6]. There is preliminary evidence for the use of DBT in reducing binge-type and/or emotional eating [7, 8]. Researchers in the field of obesity and bariatrics have been exploring the use of adapted DBT skills training groups in addressing binge-type and emotional eating, with promising outcomes in terms of eating behaviour, emotion regulation, and weight management [3, 4, 8,9,10].

Traditionally, cognitive-behavioral treatment (CBT), interpersonal psychotherapy (IPT), and pharmacotherapy have been the treatments of choice for binge eating disorder (BED). However, these treatments have not directly addressed affective dysregulation. Further, although mindfulness-based stress reduction (MBSR) and mindfulness-based cognitive therapy (MBCT) involve skills building techniques aimed at the development of conscious awareness to the here-and-now, as well as increased attention to emotional triggers and automatic thought patterns, these interventions also lack the emotion regulation component that is central to DBT.

The current study investigated the impact of a modified 12-week DBT skills training group for outpatients of the weight management clinic (WMC) at The Ottawa Hospital, a clinic that provides treatment for patients who are overweight. The goal of this pilot project was to determine whether a DBT skills-training group, adapted for use within a weight management population, would prove useful in improving affect regulation skills among its participants. More specifically, we hypothesized that participation in this group would help patients develop adaptive coping strategies to assist in better managing unpleasant emotions. In turn, it was anticipated that improvements in participants’ affect regulation skills would translate to improvements in measures of general distress, emotional eating, mindfulness, and interpersonal relationships which together, could positively impact weight management.

Method

The interventions studied were three consecutive 12-week psychotherapy groups held at the WMC (June–August 2015, September–December 2015, December 2015–February 2016). The first group was facilitated by the third author (DM) who completed DBT training and has been practicing DBT since 2004. The clinic psychologist (DS), and a clinic social worker observed the first group as part of their training and co-facilitated subsequent groups. The groups were semi-closed such that new patients could join every 4 weeks at the beginning of a new module.

The 12-week intervention comprised three modules, each spanning four once weekly sessions. Each module began with one full session of core mindfulness. The remaining three sessions in each module then introduced and expanded upon the topics of emotion regulation, interpersonal effectiveness, and distress tolerance. Skills taught in the interpersonal effectiveness module focused on helping to ask for things or say no to unwanted requests, while at the same time maintaining or improving relationships. The emotion regulation skills module included skills to help reduce the intensity of painful emotions, skills to help “surf” emotions and skills to help change emotions. Distress tolerance skills included crisis survival and acceptance strategies. The modules involved psychoeducation, skills development, and homework. A guidebook (unpublished) based on Linehan’s skills training manual was developed for this adaptation. Each session was 1 h and 45 min in duration. The group room was open 10 min prior to group start to facilitate “settling in”; 10–15 min were dedicated to a mindfulness exercise; 40–45 min were allotted for homework review; and approximately 45 min for the introduction of new material. A 15 min grace period following group allowed for crisis intervention. In addition to homework, participants had the option of completing a diary card which recorded use of core mindfulness skills, and explored goals and values. Group discussion and sharing of experiences was encouraged during homework review. Group members were either self-referred or referred by their WMC health care provider from the surgical, medical, or behavioural stream.

Patients experiencing emotion regulation or interpersonal difficulties, emotional eating, who had current or historical diagnoses of BPD (and/or sub-clinical traits) were included. Medical stability was ensured by requiring a minimum of 6 weeks post-surgery. Patients with unstable medical conditions, with acute psychosis, mania, or substance abuse were excluded and referred elsewhere. Patients attended a 1-h pre-group screening to provide education about the group, discuss goals, and confirm treatment fit. This pilot study included three groups, each with six participants. Group 2 included an additional patient who declined research participation. Group 3 ended two weeks early due to unexpected provider leave (i.e., 10 rather than 12-weeks in duration). The study was approved by the Ottawa Health Science Network Research Ethics Board and all participants completed a signed consent form.

A longitudinal pre-post non-randomized design with a 3-month follow-up was used for this study. Outcome measures were completed in person at weeks 1 and 12, and patients then received the measures by mail with a postage paid envelope 3-months following termination of group. Outcome measures included: (1) The Outcome Questionnaire-45 (OQ-45), a 45-item validated scale that asks respondents to rate difficulties across different life areas (e.g., personal distress, relationships, and responsibilities) on a 5-point Likert scale (0 = never; 4 = almost always); (2) The Emotional Overeating Questionnaire (EOQ), a 9-item validated scale that asks respondents to rate how likely they were to overeat over the past month in response to different emotions on a 7-point Likert scale (0 = not likely at all; 6 = extremely likely); (3) The Difficulties with Emotion Regulation Scale (DERS), a 36-item validated scale that asks respondents to indicate how closely they relate to various statements on emotion regulation, awareness, understanding, and acceptance of emotions, using a 5-point Likert scale (1 = almost never; 5 = almost always); and, (4) The Five Facet Mindfulness Questionnaire (FFMQ), a 39-item scale that asks respondents to rate their experience of the five facets of mindfulness (e.g., observing, describing, acting with awareness, non-judging of inner experience, and non-reactivity to inner experience) along a 5-point Likert scale (1 = never or very rarely true; 5 = very often or always true). Total score was computed for these four measures and higher scores reflected greater difficulties, except for FFMQ, whereby higher scores reflected greater experiences of mindfulness. Demographic information was collected at baseline, and attendance, collected throughout. The Client Satisfaction Questionnaire (CSQ-3), a 3-item measure that evaluates satisfaction along a 4-point Likert scale, was administered at the end of the final session; patients were also asked what they liked most about this group, what could have been better, and whether there were other services needed.

Results

Of the 18 study participants, three were male and 15 were female. Eleven participants were in the surgical treatment stream and seven were in the medical stream. The mean age was 46.3 years (range 25–67). The majority reported their maternal language as English (72.2%) and as having been born in Canada (88.9%). Most reported having graduated from college/university (66.7%) or having completed some college/university (22.2%). Over a third were divorced or separated (38.9%), a third single (33.3%), and the remaining were married or common law (27.8%). 38.9% were working full-time, 33.3% on disability leave, and the remaining were retired, studying, or working part-time.

Patients attended on average 9.7 sessions (range 2–15; SD = 3.0).Footnote 1 Three patients discontinued prior to the group completion. Using the CSQ-3, satisfaction with the group was high (average of 3.71 out of 4 for the three item total). Patient’s comments about what they liked most related to three main themes. The first theme, support from the group, spoke to participant experiences of non-judgement and feeling cared for by group members and leaders. They also appreciated the information and tools learned, such as skills for regulating emotions, and the experience of sharing and not being alone in their struggles. In terms of what could have been better about the group, suggestions clustered across three main themes. Many expressed a preference for more programming, for example: a longer duration for group or more follow-up beyond the group. A second theme related to group rules around attendance—with a few members suggesting that group leaders could have been more firm to facilitate greater cohesion. Some commented on the group discussion related to homework, indicating that less time on this portion of the group would have been preferred in favor of more time for education. When asked whether there were other services needed, many participants indicated that nothing was lacking, while some others expressed to needing additional psychological services; one participant requested medical monitoring for weight.

Data were prepared for analysis in SPSS by first verifying accuracy for approximately 20% of the data and subsequently computing a weighted mean for missing values within measures containing more than 75% of the items. A comparison of the three therapy groups involving a one-way ANOVA was run to investigate for baseline differences in dependent measures; no significant differences were found. Independent sample t tests were run to assess for possible differences in baseline dependent measures between participants who attended eight or more sessions to those who attended fewer than 8 sessions; no significant differences were found.

The effectiveness of the group was assessed through repeated measures ANOVAs across the three time points (pre-intervention, post-intervention and 3-month follow-up; see Table 1). Four repeated measures ANOVAs were run to assess for changes related to general distress (OQ-45), emotional overeating (EOQ), emotion regulation (DERS), and mindfulness (FFMQ). Changes in functioning as measured by the OQ-45 approached but did not reach statistically significant improvement (F(2,16) = 2.975; p = 0.080; see Fig. 1). Next, a repeated measures ANOVA with a Greenhouse-Geisser correction, identified a significant effect for EOQ across time points (F(1.134,10.21) = 6.909; p = 0.022; η2 = 0.434). Post hoc tests using Fisher’s LSD tests were conducted to assess for linear and quadratic trends. Results pointed to a significant linear trend (F(1,9) = 8.245; p = 0.018) which was indicative of decreasing EOQ scores over time, but a non-significant quadratic trend (F(1,9) = 4.127; p = 0.073), suggesting no leveling off of effect over time. Specifically, significant improvements were found between baseline (M = 1.5, SD = 1.2) and post treatment (M = 0.5, SD = 0.3; p = 0.034) and between baseline (M = 0.5, SD = 1.2) and follow-up (M = 0.4; SD = 0.5; p = 0.018; see Fig. 2). The third repeated measures ANOVA identified a significant effect for DERS across time points (F(2,18) = 7.284; p = 0.005; η2 = 0.447). Post hoc tests using Fisher’s LSD tests pointed to a significant linear trend (F(1,9) = 7.73; p = 0.021), suggesting reduced DERS scores over time and a significant quadratic trend (F(1,9) = 6.099; p = 0.0.036), suggesting a leveling off of or maintenance of effect over time. Significant improvements were found between baseline (M = 124.3, SD = 21.94) and post treatment (M = 89.7; SD = 32.09; p = 0.008) and between baseline and follow-up (M = 88.7, SD = 41.46; p = 0.021), suggesting that the treatment led to statistically significant improvements in regulating emotions that were maintained at follow-up (see Fig. 3). The final repeated measures ANOVA identified a significant effect for FFMQ across time points (F(2,14) = 5.179; p = 0.021; η2 = 0.425). Post hoc tests using Fisher’s LSD tests found a significant linear trend (F(1,7) = 6.474; p = 0.038), indicating increasing FFMQ scores across time, but a non-significant quadratic trend (F(1,7) = 2.984; p = 0.128) suggesting that FFMQ scores continued to increase over time. Significant improvements were found between baseline (M = 106.25, SD = 15.20) and follow-up (M = 134.25, SD = 29.27; p = 0.038); a non-significant trend was indicated between baseline and post treatment (M = 132.89; SD = 31.03; p = 0.062), suggesting statistically significant improvements in mindfulness at follow-up (see Fig. 4).

Table 1 Outcome measures from baseline to 3-month follow-up: examination of general distress, emotional overeating, mindfulness, and emotion regulation
Full size table
Fig. 1
figure 1

Mean score on OQ-45 across time

Full size image
Fig. 2
figure 2

Mean score on EOQ across time

Full size image
Fig. 3
figure 3

Mean score on DERS across time

Full size image
Fig. 4
figure 4

Mean score on FFMQ across time

Full size image

Discussion

The present study evaluated the impact of a 12-week, modified DBT skills group on measures of general distress, emotional eating, mindfulness, and affect regulation among a group of patients undergoing weight management treatment. Patients showed increased mindfulness, improved emotion regulation, and reduced emotional overeating after participating in this group, and these positive changes were maintained and/or further improved at follow-up. Although changes for the OQ-45 were not statistically significant, power was low and improvements were clinically meaningfully (i.e., greater than 14 points). The quadratic trends were at times non-significant and improvements maintained or furthered at the 3-month follow-up; this suggests promise for patients in terms of continued improvements beyond this treatment.

These findings suggest that a group-based psychological intervention that incorporates psychoeducational and skills building components aimed at addressing affective dysregulation may be a useful treatment for a weight management outpatient population. This study adds to other recent studies supporting the use of adapted DBT skills-training groups for weight management populations [9, 10]. Limitations of this study include the short-term study duration, small sample size, and non-experimental design. That two group members received an additional 4 sessions, in addition to the shorter duration of the third group may also have impacted study outcomes. Further research would be useful in confirming these results and in determining the pathway for improvements. The finding from this study that both EOQ and DERS scores were reduced from baseline to follow-up suggests that improved emotion regulation is related to reduced frequency of emotional eating to self-soothe. The low rate of program attrition would also suggest that it is acceptable to both surgical and non-surgical patients, a concern highlighted by previous research questioning whether low attrition rates for bariatric patients might be motivated by the promise of surgery [8]. Further study is needed to determine whether improvements in the constructs studied in this research would translate to improved weight management outcomes.