Introduction

Social anxiety disorder (SAD) is a fear of being embarrassed or humiliated in social or performance-based situations, resulting in significant functional impairment or distress regarding the anxiety (American Psychiatric Association 2013). Given the fears of being negatively evaluated by others (American Psychiatric Association 2013) and being judged negatively by others based on their appearance (“social appearance anxiety”; Levinson and Rodebaugh 2011), it is understandable that SAD and obesity often are comorbid. Some studies have indicated a prevalence rate of SAD of 6–9% among obese individuals (Kalarchian et al. 2007; Mather et al. 2008; Petry et al. 2008), and rates of SAD within samples of bariatric surgery candidates range from 3.2 to 9.4% (Dalrymple et al. 2011; Kalarchian et al. 2007; Mauri et al. 2008; Muhlhans et al. 2009). For example, SAD was the second most prevalent disorder in a sample of nearly 2000 bariatric surgery candidates (after specific phobia), with a lifetime prevalence rate of 7.5% (Dalrymple et al. 2011).

The presence of comorbid social anxiety could have potential implications for bariatric surgery candidates, as some evidence suggests that the presence of psychological problems can negatively affect outcomes in bariatric surgery (de Zwaan et al. 2011; Lier et al. 2013; White et al. 2010). Individuals with social anxiety-obesity comorbidity may be prone to engaging in a vicious cycle of overeating to manage emotions, which may lead to weight gain and resulting feelings of shame, and further overeating to manage feelings of shame. Preoperative levels of shame and self-perceived inadequacy were found to be significant predictors of the presence of postoperative psychiatric disorders in a sample of bariatric surgery patients (Lier et al. 2013), and the presence of SAD was significantly associated with emotional eating in a cross-sectional sample (Ostrovsky et al. 2013).

Prior research has suggested a potential link between comorbid SAD and emotional eating in bariatric surgery candidates (Dalrymple et al. 2017), finding that 62.2% of surgery candidates with SAD endorsed emotional eating compared to 24.0% of surgery candidates with no lifetime history of psychiatric disorders. Furthermore, a greater percentage of surgery candidates with SAD and “modified SAD” (social anxiety related to weight only) reported emotional eating as the primary reason for the weight problem compared to surgery candidates with no lifetime history of psychiatric disorders.

It is important to continue to examine the implications of this potential link between social anxiety and emotional eating, as it may affect bariatric surgery outcomes. In general, emotional eating is a common behavior found in bariatric surgery candidates (occurring in up to 40% of surgery candidates) and often is associated with depression and anxiety symptoms (Miller-Matero et al. 2014). Pretreatment emotional eating severity has been found to be significantly associated with poorer weight loss following three types of bariatric procedures (gastric bypass, laparoscopic adjustable gastric band, biliopancreatic diversion; Castellini et al. 2014). It also has mediated the relationship between a neurotic predisposition and weight loss, as well as the relationship between a neurotic predisposition and quality of life, following bariatric surgery or a weight loss program (Canetti et al. 2009). Neurotic predisposition is a higher-order construct of negative affectivity that is correlated with self-esteem, neuroticism, locus of control, fear of intimacy, and generalized self-efficacy, factors that also are associated with social anxiety (Allan et al. 2017; Iancu et al. 2015; Ritter et al. 2013). However, one study found that emotional eating was associated with increased odds of postsurgical weight loss success (defined as ≥ 50% excess weight loss 2 years postoperatively; Wedin et al. 2014).

It may be helpful to expand understanding of psychological processes that may be associated with emotional eating, to further inform treatment development efforts and improve surgical outcomes for these patients. One such process is mindfulness. Studies from bariatric surgery samples, patients with diabetes, college students, and a nonclinical sample have found that higher levels of mindfulness were negatively associated with emotional eating (Lattimore et al. 2011; Ouwens et al. 2015; Pidgeon et al. 2013; Tak et al. 2015). More specifically, one study found that higher levels of acting with awareness, describing, and nonjudgment were predictive of less emotional eating, with the acting with awareness facet being the strongest predictor (Tak et al. 2015). Other studies have found that greater difficulty identifying and describing feelings is associated with more emotional eating in men (Larsen et al. 2006), and a prospective study found that higher levels of observing predicted greater emotional eating over the course of a 6-month period (Sala and Levinson 2017).

The association between emotion regulation and emotional eating also has been explored. Emotion regulation is defined as a multidimensional construct that includes emotional awareness, distress tolerance, and behavioral control (Gratz and Roemer 2004). Cultivating mindfulness may be one way of enhancing emotion regulation skills. Emotion regulation has mediated the relationship between mindfulness and symptoms of depression and anxiety (Freudenthaler et al. 2017), as well as the relationship between mindfulness and uncontrolled eating (e.g., loss of control over eating when hungry or exposed to food stimuli; Fisher et al. 2017). Difficulties in emotion regulation have predicted emotional eating in college students and a general adult population in Australia (Crockett et al. 2015; Stapleton and Whitehead 2014). More specifically, deficits in goal-directedness, emotional awareness, and impulse control were uniquely associated with poor regulation of overeating (including “discomfort overeating”; Kerin et al. 2017). In addition, emotion suppression in adults leads to a greater intake of comfort foods, and difficulties in emotion regulation predicted emotional overeating above and beyond negative affect (Evers et al. 2010; Gianini et al. 2013). In bariatric surgery candidates, emotion dysregulation has mediated the relationship between weight bias internalization and emotional eating (Baldofski et al. 2016), as well as the relationship between attachment insecurity and emotional eating (Taube-Schiff et al. 2015). Difficulty identifying emotions also was associated with more emotional eating among female bariatric surgery candidates (Zijlstra et al. 2012). Specific to social anxiety, it was found that the relationship between social anxiety and disordered eating was fully mediated by expressive suppression in undergraduate women (McLean et al. 2007).

In a study of 820 bariatric surgery candidates that examined the relationship between specific mindfulness facets and problematic eating behaviors, it was found that higher levels of acting with awareness and describing were associated with less emotional eating above and beyond depression severity, with a trend toward a similar relationship with nonreactivity (Levin et al. 2014). Another study showed an association between SAD and emotional eating in bariatric surgery candidates (Dalrymple et al. 2017). However, the impact of the interaction between social anxiety symptoms and mindfulness/emotion regulation on emotional eating has not yet been examined. Perhaps one reason for the lack of relationship between some facets of mindfulness or emotion regulation processes and emotional eating could be that the interaction between symptoms and reaction to symptoms is most important, rather than either variable independently.

The current report from the Rhode Island Bariatric Surgery (RIBS) study of the Methods to Improve Diagnostic Assessment and Services (MIDAS) Project seeks to extend this work by examining the impact of the interaction between mindfulness facets/emotion regulation and social anxiety symptoms on the presence of emotional eating within bariatric surgery candidates. Based on our prior studies and findings from other studies, we hypothesized that the interaction of social anxiety symptoms with facets of mindfulness/difficulties in emotion regulation would be associated with emotional eating above and beyond the independent effects of these variables.

Method

Participants

Participants included 1088 individuals seeking bariatric surgery in Rhode Island for whom data was recorded on the presence or absence of lifetime emotional eating. Data was collected from July 2004 through October 2009. The decision was made to examine social anxiety continuously rather than only those with a SAD diagnosis, given prior findings that individuals with more “subthreshold” levels of SAD (e.g., those for whom the social anxiety was related to weight only) closely resembled the group diagnosed with DSM-IV SAD in terms of both social anxiety severity and bariatric surgery characteristics (Dalrymple et al. 2011, 2017). The majority of participants were Caucasian, female, and married and had graduated high school (Table 1). Average BMI for this sample was 47.4 and ranged from 45.9 to 63.2, indicating that all of the participants were obese based on criteria from the National Institutes of Health (1998). The most common current axis I disorders across the entire sample were specific phobia (9.0%) and SAD (7.9%). Because of the potential confound of having an eating disorder, participants with a lifetime history of anorexia nervosa, bulimia, and binge eating disorder were not included in the sample of 1088 participants.

Table 1 Demographic characteristics of bariatric surgery candidates with versus without emotional eating
Full size table

Procedure

Surgeons referred participants for a comprehensive presurgical psychiatric diagnostic evaluation as part of the surgical clearance process. The evaluations were conducted at the outpatient psychiatry practice of Rhode Island Hospital. The RIBS program is an integrated clinical research program, in which evaluations are conducted as part of a clinical service to determine psychiatric clearance for bariatric surgery, and informed consent is obtained to use the information from the evaluation for research purposes. Informed consent was obtained prior to administering the diagnostic interview. All research procedures were approved by the institutional review committee at Rhode Island Hospital. Details on the development of the RIBS study of the MIDAS Project have been presented elsewhere (Zimmerman et al. 2007).

The diagnostic evaluations were conducted by research assistants with a bachelor’s degree in biological or social sciences and doctoral-level or licensed psychologists. Interviewers received extensive training and were monitored throughout the project to minimize rater drift. Research assistants observed at least 20 interviews before administering at least 20 interviews while being observed, and psychologists observed 5 interviews and administered 15 to 20 interviews while being observed. Interviewers then demonstrated exact or near-exact diagnostic reliability with a senior interviewer on five consecutive evaluations.

Measures

Structured Clinical Interview for DSM-IV (SCID; First et al. 1996)

The SCID was used to diagnose current and lifetime axis I disorders. Lifetime axis I disorders were assessed retrospectively. Joint interview reliability evaluations were conducted throughout the project on 63 bariatric surgery candidates, with good to excellent agreement on diagnoses (e.g., major depressive disorder kappa = 1.0; agoraphobia without panic disorder = 1.0; specific phobia = 0.82; SAD = 0.79; obsessive compulsive disorder = 1.0; generalized anxiety disorder = 1.0).

Rhode Island Bariatric Surgery Interview (RIBSI; Zimmerman et al. 2007)

The RIBSI is a semistructured interview added to the SCID evaluation designed as a bariatric surgery assessment module. The RIBSI is clinician-administered and assesses past and current eating behaviors, expectations for surgery, knowledge of the surgery, dieting history, current stressors, and social support. For the current study, the variable of emotional eating from the RIBSI was used for analyses (“non-binge eating that has as its primary goal the reduction of emotional upset or distress”). Participants were asked: “For some people, eating is a way to cope with stress or help manage their emotions. Do you turn to food to comfort yourself when you are upset, anxious, depressed, or angry?” This item was rated as present (either currently or in the past) or absent by the clinical interviewer based on the response to the question above. Agreement between interviewers on this item was high (κ = 0.83). A prior study (Levin et al. 2014) examined emotional eating as a continuous variable of frequency of emotional eating in the past 30 days. However, some candidates were already engaged in the presurgical process and reported having made dietary changes in the past month, and therefore, these individuals were excluded from analyses in that study. It was believed that a history of emotional eating might be associated with poorer mindfulness and emotion regulation skills even if candidates denied emotional eating currently, and therefore, the decision was made to examine a lifetime history of emotional eating in the current sample.

Five Facet Mindfulness Questionnaire (FFMQ; Baer et al. 2006)

The FFMQ is a self-report questionnaire composed of 39 items and consists of five subscales assessing different facets of mindfulness: observing, describing, acting with awareness, being nonjudgmental, and being nonreactive. Items are rated on a 5-point ordinal scale ranging from 1 (never or very rarely true) to 5 (very often or always true). The FFMQ has demonstrated reliability and validity (Baer et al. 2006), and each of the subscales in the current study showed good internal consistency (Cronbach’s alpha ranging from 0.79 to 0.87).

Difficulties in Emotion Regulation Scale (DERS; Gratz and Roemer 2004)

The DERS is a 36-item self-report questionnaire that assesses the ability to effectively handle emotions, with higher scores indicating greater difficulty in regulating emotions. Items are rated on a 5-point ordinal scale ranging from 1 (almost never) to 5 (almost always). The DERS contains six subscales: lack of awareness of emotional responses (awareness), lack of clarity of emotional responses (clarity), nonacceptance of emotional responses (nonacceptance), limited access to emotion regulation strategies (strategies), difficulties controlling impulses when experiencing negative emotions (impulse), and difficulties engaging in goal-directed behaviors when experiencing negative emotions (goals). The DERS has demonstrated high internal consistency, good test-retest reliability, and adequate construct and predictive validity (Gratz and Roemer 2004). The internal consistencies for the total score and subscales in the current study were high (ranging from 0.76 to 0.88).

Clinically Useful Social Anxiety Disorder Outcome Scale (CUSADOS; Dalrymple et al. 2013)

The CUSADOS is a 12-item self-report measure of social anxiety severity. Items are rated on a 5-point ordinal scale ranging from 0 (not at all true) to 4 (almost always true). The CUSADOS has demonstrated excellent internal consistency, good discriminant and convergent validity, and high item-total correlations and test-retest reliability within an outpatient psychiatric sample (Dalrymple et al. 2013). Internal consistency in the current sample also was excellent (Cronbach’s alpha = 0.95).

Clinically Useful Depression Outcome Scale (CUDOS; Zimmerman et al. 2008)

The CUDOS is an 18-item self-report questionnaire of depression, tied to DSM-IV criteria. Items are rated on a 5-point ordinal scale, ranging from 0 (not at all true) to 4 (almost always true). The last two items assess interference of depression symptoms and quality of life; therefore, a total score is derived by summing the first 16 items. The CUDOS has demonstrated excellent reliability and discriminant/convergent validity (Zimmerman et al. 2008). Internal consistency in the current sample also was excellent (Cronbach’s alpha = 0.89).

Data Analyses

Preliminary analyses were conducted to compare those who endorsed versus denied lifetime emotional eating on demographic and other clinical characteristics such as social anxiety and depression scores. Next, initial analyses were conducted to examine whether bariatric patients who did versus did not report engaging in emotional eating differed on mindfulness facets and emotional regulation subscale scores, using independent t tests. Cohen’s d effect sizes also were calculated (0.3 = small effect; 0.5 = medium effect; 0.8 = large effect). Pearson correlations also were calculated between emotional eating, the CUDOS, the CUSADOS, the FFMQ, and the DERS. A series of hierarchical logistic regression analyses were conducted to examine the relationship between the interaction of each FFMQ facet/DERS subscale and social anxiety symptoms on the presence of emotional eating. Gender and depression severity (via the CUDOS) were entered into the first block of each regression analysis because these variables were significantly associated with emotional eating in the preliminary analyses. It also was decided a priori to control for depression symptoms in the first block because they often are comorbid with social anxiety symptoms. Social anxiety severity (via the CUSADOS) and the individual FFMQ facet/DERS subscale were entered into the second block to test for their independent associations with emotional eating. The CUSADOS by FFMQ facet/DERS subscale interaction was entered into the third block to test for its association with emotional eating above and beyond the independent effects of those variables. Two final hierarchical logistic regression analyses were conducted: one to test the relative effects of FFMQ facets and their interactions with social anxiety and another to test the relative effects of DERS subscales and their interactions with social anxiety. In these two analyses, all FFMQ facets or DERS subscales were entered individually into the second block, and all FFMQ facet or DERS subscales by CUSADOS interactions were entered into the third block.

Results

As shown in Table 1, a greater percentage of candidates endorsing emotional eating were female. Therefore, gender also was entered into the first block of the subsequent regression analyses as a covariate. There were no other demographic differences between those endorsing versus not endorsing emotional eating. There were significant differences between the two groups on depression and social anxiety symptoms, with emotional eaters reporting greater depression and social anxiety symptoms compared to nonemotional eaters. Chi-square analyses showed that a greater percentage of candidates endorsing emotional eating met current criteria for an adjustment disorder, major depression, dysthymia, SAD, and an impulse control disorder compared to candidates denying emotional eating (Table 2). For lifetime diagnoses (Table 3), a greater percentage of emotional eaters met criteria for a past or partial remission diagnosis of a major depressive disorder, panic disorder, and alcohol or drug use disorders compared to nonemotional eaters.

Table 2 Current axis I diagnoses in bariatric surgery candidates with versus without emotional eating
Full size table
Table 3 Lifetime axis I diagnoses in bariatric surgery candidates with versus without emotional eating
Full size table

Initial Pearson correlations conducted between the variables showed small to moderate significant correlations between variables, with the highest correlations occurring between some subscales of the DERS (see Table 5). Initial t tests on mindfulness facets showed that surgery candidates endorsing emotional eating reported significantly lower scores on the describing, acting with awareness, and being nonjudgmental facets compared to candidates denying emotional eating, with a small effect for each. There were no significant differences between these two groups on the observing and being nonreactive facets (Table 4).

Table 4 Symptom severity, mindfulness, and emotion regulation in bariatric surgery candidates with versus without emotional eating
Full size table

Hierarchical logistic regressions then were conducted to examine the interaction between mindfulness facets and social anxiety severity and the association of these interactions with emotional eating. The final model was significant for the observing, describing, and acting with awareness facets, with the interaction between those facets and social anxiety symptoms being significantly associated with emotional eating above and beyond depression severity, gender, and the independent effects of mindfulness facets and social anxiety (Table 5). For the being nonjudgmental and being nonreactive facets, only depression severity and gender were significantly associated with emotional eating (Table 6). The final hierarchical logistic regression analysis testing the relative effects of all mindfulness facets showed that only the observing by social anxiety interaction was associated with emotional eating, relative to the other mindfulness facets and mindfulness × social anxiety interactions (overall model χ2 = 111.68, − 2LL = 1063.71, p < 0.01, Nagelkerke R2 = 0.16; observe × CUSADOS p = 0.02).

Table 5 Correlations between emotional eating, depression, social anxiety, mindfulness, and emotion regulation
Full size table
Table 6 The interaction between mindfulness facets or emotion regulation and social anxiety symptoms and its association with emotional eating
Full size table

The initial t tests on emotion regulation showed that candidates endorsing emotional eating reported poorer emotion regulation on all subscales of the DERS compared to candidates denying emotional eating (Table 4), with small to medium effects. Similar hierarchical logistic regressions were conducted to test the interaction of emotion regulation and social anxiety symptoms and its association with emotional eating. Significant results were obtained for the goals, impulse, strategies, and clarity subscales, indicating that the interaction between these subscales and social anxiety symptoms was significantly associated with emotional eating after controlling for depression severity, gender, and the emotional regulation and social anxiety variables independently. For the regression on nonacceptance, only depression severity, gender, and nonacceptance were significantly associated with emotional eating. For the regression on the DERS awareness subscale, only depression severity, gender, and awareness were significantly associated with emotional eating (Table 6). The final regression testing the relative effects of emotion regulation subscales showed that only the nonacceptance by social anxiety interaction was significantly associated with emotional eating relative to other emotion regulation subscales (overall model χ2 = 120.29, − 2LL = 1167.97, p < 0.01, Nagelkerke R2 = 0.16; nonaccept × CUSADOS p = 0.03).

Discussion

Prior research has established a link between psychopathology and deficits in mindfulness and emotion regulation (Cisler and Olatunji 2012; Joormann and Gotlib 2010; Sloan et al. 2017). The current study has expanded upon this literature by examining how the interaction of psychopathology and mindfulness/emotion regulation deficits impacts upon problematic eating behaviors, especially within a bariatric surgery population. Given the increased utilization of bariatric surgery over the years (Buchwald et al. 2004) and the percentage of individuals who experience difficulties following surgery (Beck et al. 2012; Sarwer et al. 2011), it is important to examine these relationships to inform treatment development efforts to enhance outcomes from bariatric surgery. Overall, results from the current study suggest that certain mindfulness or emotion regulation skills may be particularly important to target either pre- or postsurgery, especially for surgical candidates experiencing social anxiety symptoms.

Bariatric surgery candidates who endorsed past or current emotional eating were more likely to be female. This is consistent with prior research that has examined the relationship between emotional eating and depressive symptoms, finding a stronger effect in women compared to men (Rawana et al. 2016; van Strien et al. 2016). In addition, they reported greater depression and social anxiety severity and were more likely to meet current criteria for an adjustment disorder, major depression, dysthymia, SAD, and an impulse control disorder compared to surgery candidates not endorsing emotional eating. Prior research also has shown that emotional eating is significantly associated with higher levels of anxiety and depression in bariatric surgery candidates (Miller-Matero et al. 2014; Sevincer et al. 2017). A similar pattern was found for lifetime diagnoses, with emotional eaters being more likely to meet criteria for a past or partial remission major depressive disorder and panic disorder, as well as an alcohol or drug use disorder compared to nonemotional eaters. In particular, the finding on substance use disorders is consistent with research that has examined processes underlying overconsumption of food and substances, showing that higher trait reward sensitivity is associated with binge eating, hazardous drinking, and use of illicit substances (Loxton and Dawe 2006; Loxton and Tipman 2017). Further research in this area is needed, as presurgical eating behaviors such as food addiction, nocturnal eating, and environmental responsiveness to food cues have been associated with substance misuse following Roux-en-Y gastric bypass surgery (Reslan et al. 2014).

Overall, surgery candidates endorsing emotional eating reported poorer mindfulness skills for the facets of describing, acting with awareness, and being nonjudgmental, consistent with prior studies (Larsen et al. 2006; Levin et al. 2014; Tak et al. 2015). When examining the impact of social anxiety interacting with mindfulness facets, it was found that the interaction between social anxiety severity and observing, describing, and acting with awareness was significantly associated with emotional eating. Interestingly, the observing facet was associated with emotional eating only in the context of the interaction with social anxiety, and not independently. Previous research has found that the observing facet often is either not a significant predictor or significantly associated with greater, rather than fewer, problems (Baer et al. 2006; Cash and Whittingham 2010; Desrosiers et al. 2013). Higher levels of observing were associated with higher levels of social anxiety, supporting the hypothesis that observing in and of itself may be more indicative of a hypersensitivity or attention bias effect that often occurs in social anxiety and other anxiety disorders (Allan et al. 2017; Levin et al. 2017; Michel et al. 2016). The opposite occurred with the describing and acting with awareness facets, in which higher levels of these skills were associated with lower social anxiety severity. Examination of items for these facets shows that deficits in these areas are consistent with clinical characteristics of individuals experiencing social anxiety (e.g., for describing: “I have trouble thinking of the right words to express how I feel about things”; for awareness: “I find it difficult to stay focused on what’s happening in the present”). Interestingly, the nonreact facet was not associated with emotional eating, but this is consistent with other studies finding a lack of a relationship especially when analyzed in relation to the other facets (Levin et al. 2014). Items on this subscale assess the degree to which one is able to notice thoughts and feelings without responding to them (e.g., “I perceive my feelings and emotions without having to react to them”). Future research should continue to examine possible reasons for this lack of relationship. When examining the relative effects of mindfulness facets in a single regression analysis, only the observing by social anxiety interaction was significantly associated with emotional eating. Therefore, for individuals with higher levels of social anxiety who engage in emotional eating, it may be particularly helpful to address deficits in nonjudgmental observation of experiences.

As expected based on prior research (Baldofski et al. 2016; Kerin et al. 2017; McLean et al. 2007; Zijlstra et al. 2012), surgery candidates endorsing emotional eating reported greater deficits in emotion regulation compared to nonemotional eaters with respect to lack of awareness of emotional responses, lack of clarity of emotional responses, nonacceptance of emotional responses, limited access to emotion regulation strategies, difficulties controlling impulses when experiencing negative emotions, and difficulties engaging in goal-directed behaviors when experiencing negative emotions. When examining the impact of the interaction between social anxiety symptoms and emotion regulation, it was found that the interaction between social anxiety and goals, impulse control, strategies, and clarity was significantly associated with emotional eating. Items on these subscales tend to reflect characteristics described by individuals with social anxiety concerns, such as difficulty engaging in goal-directed behaviors when upset (“When I’m upset, I have difficulty getting work done”), experiencing lower awareness of or difficulty describing feelings (“I have no idea how I am feeling”), and a sense of losing control over oneself when upset (e.g., “When I’m upset, I lose control over my behaviors”). Interestingly, the final regression analysis testing all emotion regulation strategies found that only the nonacceptance by social anxiety interaction was significantly associated with emotional eating relative to the other emotion regulation subscales. Items on this subscale reflect a tendency to become upset with oneself for experiencing emotions (e.g., “When I’m upset, I become embarrassed for feeling that way”). Some of the higher correlations occurred between the goals, impulse control, and strategies subscales, suggesting a high degree of overlap between these subscales that may have canceled out effects when added into the analysis together. Therefore, nonacceptance may be a particularly unique and important aspect of emotion regulation to target in surgery candidates experiencing social anxiety who are engaging in emotional eating.

Social anxiety is frequently present in bariatric surgery candidates, and it is commonly associated with emotional eating. The current study expanded this knowledge by examining potential mechanisms that may be associated with the link between emotional eating and social anxiety. It previously was hypothesized that it may not be the mere presence of social anxiety that is associated with emotional eating, rather one’s reaction toward the anxiety (Dalrymple et al. 2017). Results from the current study suggest that bariatric surgery candidates with higher levels of social anxiety who have deficits in certain mindfulness and emotion regulation skills may be more likely to engage in emotional eating, above and beyond the presence of social anxiety symptoms alone.

Some research has indicated that mental health problems resolve after bariatric surgery (Herpertz et al. 2003; Karlsson et al. 2007; Lier et al. 2013), yet other studies have found that anxiety severity did not decrease after surgery (Lier et al. 2013; Mathus-Vliegen 2007; Sarwer et al. 2008). It is possible that some individuals with social anxiety and poor mindfulness or emotion regulation skills may experience poorer long-term outcomes following surgery, especially if these individuals are likely to engage in emotional eating as a way to cope with negative affect. More recently, mindfulness- and acceptance-based treatments have been developed to address emotional eating and improve outcomes for bariatric surgery patients, showing that these interventions significantly reduced emotional eating (Chacko et al. 2016; Weineland et al. 2012). However, additional research is needed to examine surgical outcomes for individuals with social anxiety specifically, as well as potential mediators and moderators of outcomes.

Analyses were conducted with social anxiety only; therefore, results may be a reflection of psychopathology more broadly. However, analyses controlled for depression severity, and results indicated a significant interaction between mindfulness or emotion regulation deficits and social anxiety severity above and beyond depression severity. Nonetheless, future research should examine the interaction between mindfulness and emotion regulation deficits and depression, given that depression severity was significantly associated with emotional eating in all regression models. In addition, the amount of variance explained by each of the models is low, and future research should examine if other variables may explain a greater portion of variance.

The current study was conducted in a sample of bariatric surgery candidates seeking psychiatric clearance for surgery, and results may not generalize to individuals suffering from obesity in the general population or those seeking other types of weight loss treatment. As a result of seeking psychiatric clearance, they may have underreported the presence of emotional eating or psychiatric symptoms for fear that they would not get cleared. In addition, lifetime axis I disorders were assessed retrospectively, which may introduce recall biases. Finally, the majority of the sample was female and Caucasian; therefore, results may not be generalizable to surgery candidates with other demographic characteristics.

Results from the current study extend findings from previous studies examining a potential link between social anxiety and emotional eating in bariatric surgery candidates. Overall, surgery candidates endorsing emotional eating reported greater depression and social anxiety severity compared to nonemotional eaters, were more likely to meet criteria for several axis I disorders (including SAD), and reported greater deficits in mindfulness and emotion regulation skills. Results also suggested that individuals with higher levels of social anxiety experienced deficits in certain mindfulness and emotion regulation skills and that the interaction between social anxiety and these deficits was associated with emotional eating above and beyond depression severity, social anxiety severity, and mindfulness/emotion regulation deficits independently. This suggests that in some cases perhaps it is how one reacts to the social anxiety, rather than the mere presence of social anxiety, that is associated with emotional eating.