Internalized Shame in Treatment-Seeking Adults with Obesity Class II-III and Its Association with Quality of Life, Body Image, and Self-Esteem

Obesity is associated with several medical comorbidities, and living with obesity often has a negative impact on health-related quality of life (HRQoL) [1‒3]. A history of psychopathology, in particular depression [4, 5], is common among treatment-seeking individuals with obesity [6‒9].

A major source of distress for individuals living with obesity is the pervasive weight stigma they encounter [10]. Research has shown that individuals living with obesity are perceived as lazy, unintelligent, lacking self-discipline, and poorly motivated compared to those with normal weight [10]. This widespread stigmatization can underpin various forms of psychopathology [11, 12]. Exposure to obesity stigma can impair HRQoL [13] and negatively affect body image and increase symptoms of depression and anxiety [14, 15]. Long-term exposure to obesity stigma may lead to internalized weight stigma [16‒18]. Internalized stigma occurs when individuals adopt negative societal attitudes about themselves, leading to emotional distress, independent of external judgment. This self-directed stigma often creates a cycle of psychological suffering. Stigmatizing experiences often trigger feelings of shame [19, 20], which are central to the experience of internalized shame. Internalized weight stigma and internalized shame partly overlap, as both include profound negative self-evaluation. However, internalized weight stigma is more specifically related to weight and obesity, whereas internalized shame is a more general and broader concept related to global self-worth [21].

Shame is an emotion tied to a negative internal state, marked by feelings of being fundamentally flawed, exposed, and powerless, often accompanied by a self-view of worthlessness [22, 23]. The information threat theory of shame, an evolutionary-based framework, suggests that social devaluation and the loss of social status are perceived as existential threats. In this context, shame acts as a warning signal for potential social rejection, prompting adaptive behaviors such as withdrawal and hiding to mitigate the risk of exclusion [24]. Shame is associated with a range of psychopathologies, including social anxiety [25], eating disorders [26], low self-esteem [27], and depression [28]. It is also a recurrent theme in qualitative studies of individuals living with obesity [29‒31]. An abundance of shame-evoking experiences forms the foundation of internalized shame, which is primarily centered around feelings of inferiority. Internalized shame is often rooted in early life experiences and relationships [32, 33]. Memories of shame related to attachment figures (a person a child or infant forms a strong emotional bond to, typically a parent) correlate with internal shame and predict depressive symptoms [34].

Preoperative internalized shame has been associated with the persistence of psychiatric comorbidities, negative body image, and reduced physical activity levels 1 year after bariatric surgery [33, 34]. Consequently, identifying individuals with high levels of internalized shame in a treatment setting may be valuable for targeting those in need of additional support, as suggested by Lier et al. [34]. Limited knowledge exists regarding internalized shame among treatment-seeking individuals living with obesity class II-III and how internalized shame is related to HRQoL, body image, and self-esteem. The aim of this study was therefore to investigate internalized shame and its associations with HRQoL, body image, and self-esteem in treatment-seeking people with obesity.

In this cross-sectional questionnaire study, adult patients (≥18 years of age) referred to a tertiary obesity clinic at Sahlgrenska University Hospital (Gothenburg, Sweden) were recruited from the waiting list for specialist treatment (both nonsurgical and surgical) of class II (BMI 35.0–39.9) and class III (BMI ≥40) obesity. Sahlgrenska University Hospital serves approximately two million inhabitants in the Västra Götaland region. Most patients were referred by general practitioners or hospital specialists, but patients could also self-refer. Patients in need of an interpreter were excluded from the study. The study was approved by the Swedish Ethical Review Authority in Gothenburg, Sweden, reference number 897-18.

Between October 2019 and March 2020, all patients on the waiting list were sent an invitation to participate in the study. Invitations were sent by regular mail to a total of 601 patients. Together with the invitation, the potential participants received multiple questionnaires to fill out and return in a prepaid envelope. Reminders were sent twice to non-responders by regular mail.

Several different questionnaires were used to examine aspects of shame, health-related quality of life, body image, and self-esteem within the population. Specifically, internalized shame was assessed using the Internalized Shame Scale (ISS), HRQoL was evaluated on a general level with the RAND-36 and on an obesity-specific level with the Obesity-related Problem scale (OP-scale), body image was measured with the Body Shape Questionnaire (BSQ), and self-esteem was investigated using the Self-Concept Questionnaire (SCQ) and the self-esteem subscale called ISSSELFE. A more detailed description of each questionnaire follows below.

In the present study, a Swedish translation of the 5th edition of the ISS was used to measure internalized shame [35, 36]. The ISS is a validated self-report questionnaire with sound psychometric properties [37, 38]. It consists of 30 items rated on a five-point Likert scale with the following anchors: 0 (= Never), 1 (= Seldom), 2 (= Sometimes), 3 (= Often), and 4 (= Almost always). The scale is designed to assess internalized shame, adhering to the “trait” approach to shame, which emphasizes global, pervasive, and negative self-evaluations, in contrast to situational shame, which is more transient.

The ISS comprises two subscales: the Shame Subscale and the Self-Esteem Subscale (ISSSELFE). The Shame Subscale includes 24 core items that measure internalized shame through statements such as “I feel like I am never quite good enough” and “I think people look down on me.” To mitigate response set bias, six additional positively worded items, adapted from the original 10-item Rosenberg Self-Esteem Scale, were included by the developers. These six items form the ISSSELFE. Total scores for the 24 core shame items range from 0 to 96, with higher scores indicating greater levels of internalized shame.

A total score ≥50 is used as a cut-off to indicate problematic or clinical levels of shame. This cut-off score is based on research involving individuals with various mental health disorders, including depression, anxiety, and eating disorders, who typically exhibit mean shame scores around this value [37, 39]. In nonclinical populations, ISS usually yields mean/median values between 23 and 34 [38, 40‒42]. In the present study, Cronbach’s α = 0.97, indicating excellent internal consistency for ISS, for ISSSELFE Cronbach’s α = 0.90.

A self-constructed demographic questionnaire collected data on age (years), height (cm), body weight (kg), educational level, housing conditions, and civil status. Respondents were also asked “Have you ever sought and received professional help for eating disorder, depression, or other mental health problems?” and could select “Yes” for multiple categories. One question addressed binge eating by asking “Have you ever had periods of binge eating where you consumed large amounts of food and felt like you had no control over your eating?” and giving the following response options: “No - never,” “Yes - previously,” and “Yes - currently.” Two questions regarding who provided comfort and safeness in childhood were also included. Possible answers were mother, father, brother/sister, grandmother, grandfather, other relative, other person, and no one (multiple answers allowed).

The Obesity-related Problem scale was used to assess obesity-specific health-related quality of life. The OP-scale is tailored to capture the unique challenges faced by people with obesity focusing on psychosocial impairment [43, 44]. Responders rated how concerned they were by their weight and body shape in different social situations, e.g., “bathing in public,” “intimate relations with partner” or “going to a restaurant.” Raw scores were transformed to a 0–100 scale, with higher scores indicating greater impairment. The questionnaire (OP-scale version 3) consists of 14 items on a four-point scale ranging from major discomfort to no discomfort [45]. The OP-scale has shown good reliability and validity [46].

The RAND-36 was used as a measure of generic health-related quality of life [47]. It is a 36-item self-report questionnaire with eight domains of functioning and well-being. Answers are rated on a Likert-like scale and the raw score is transformed to a scale score of 0–100, with higher scores indicating better health. RAND-36 has been validated in the Swedish population and shows good reliability, validity, and clinical utility [48].

The Self-Concept Questionnaire was used to assess self-esteem [49]. In the SCQ, self-esteem is conceptualized as social acceptability/appearance, significance, worthiness, competence, resilience and determination, control over personal destiny, and value of existence. It consists of items such as “I’m easy to like” and “It would be boring if I talked about myself.” The 30 items are responded to on an eight-point Likert scale ranging from strongly disagree (0) to strongly agree (7), with a possible total score ranging from 0 to 210 and higher scores indicating better self-esteem [49]. With a Cronbach’s α of 0.91, the scale demonstrated excellent internal consistency.

The Body Shape Questionnaire was used to assess body image. The BSQ is a self-report measure consisting of 34 items on a six-point Likert-like scale ranging from “never” to “always,” with a possible score range from 34 to 204 [50]. The BSQ assesses the phenomenological experience of “feeling fat,” which is in turn connected to body image concerns. It was developed for people suffering from eating disorders and measures the extent of psychopathology regarding concerns about body shape. In the present study, Cronbach’s α was 0.96, indicating excellent internal consistency.

The statistical analyses were performed using SPSS Statistics software (version 26.0; IBM Corp., NY, USA) and SAS^® v9.4 (SAS-Institute, Cary, NC, USA). Participants were stratified into two groups according to the score on the ISS: A “low shame group” (LSG; ISS total score <50) and a “high shame group” (HSG; ISS total score ≥50). All variables were compared with ISS total score as well as LSG and HSG. Categorical variables were described as frequency and percentages, while continuous variables were shown as the mean and SD, or median and range. To compare between the LSG and the HSG, mean differences were calculated between the groups with 95% confidence interval. Fisher’s exact test (lowest 1-sided p value multiplied by 2) was used to compare between groups for dichotomous variables, while Fisher’s nonparametric permutation test was used for continuous variables. The Mantel-Haenszel chi-square test was used for ordered categorical variables and the chi-square test was used for non-ordered categorical variables. The confidence interval for dichotomous variables is the unconditional exact confidence limits. If no exact limits can be computed, the asymptotic Wald confidence limits with continuity correction are calculated instead. The confidence interval for the mean difference between groups is based on Fisher’s nonparametric permutation test. Effect size (ES) was computed with Cohen’s d [51], mean-difference/pooled SD; an ES ≥0.2 is considered small, ≥0.5 medium, and ≥0.8 large. Correlations were calculated using the Spearman rank correlation coefficient. All the tests were two-tailed and conducted at the 5% significance level.

The missing data analysis assumed data were missing at random. If a participant responded to fewer than 50% of the items on the ISS, they were excluded from the study. For the other questionnaires, if fewer than 50% of the items were completed, the questionnaire was excluded, but the participant was retained. For the included questionnaires, person-mean substitution was chosen for imputation of missing values [52]. Age, gender, and BMI were compared between non-responders and responders.

Of the 601 individuals invited to participate in the study, 262 (43.6%) responded fully or partially to the questionnaires. Five patients were excluded due to too many missing items on the ISS. Twenty-nine patients only returned demographic data and were therefore excluded from analysis. Thus, 228 patients (37.9%) with obesity class II-III referred for obesity treatment were included in the final analyses. The majority of patients were female (73.7%). The mean age was 46 years (SD = 13.0) and 32% of the participants were between 40 and 50 years of age. Almost one-third (30.7%) had a university education. Of the participants, 155 (68.0%) shared a household with a spouse/partner. On average, the study participants had a BMI of 41.8 (SD = 5.1). Demographic data are presented in Table 1.

Comparisons between the patient group that did not return the questionnaires (nonresponders; n = 339) and the study participants (responders) revealed no significant differences regarding sex or BMI. However, nonresponders were significantly younger (mean 43.5 years) compared to the responders (mean 46.1 years; p = 0.02).

The mean score for internalized shame in the total sample was 41.6 (SD = 22.5; see Fig. 1). Based on the clinical cutoff of an ISS score ≥50, 149 patients (65.4%) were classified in the LSG, while 79 patients (34.6%) were in the HSG. The mean ISS score was 28.4 (SD = 14.1) in the LSG and 66.5 (SD = 11.5) in the HSG (see Fig. 1). Within the entire cohort, 38.7% of women and 23.3% of men were classified in the HSG. The proportion of women was significantly higher in the HSG (82.3%) compared to the LSG (69.1%, p = 0.044). Participants in the HSG were significantly younger than in the LSG (mean difference 6.7 years, p < 0.0002) and had a somewhat higher BMI (42.9 vs. 41.3, p = 0.027). However, the difference in body weight in absolute numbers was only 2.1 kg. Nevertheless, the HSG reported significantly more obesity-related psychosocial problems compared to the LSG as measured by the OP-scale (22.5-point difference, p < 0.001) and the ES was large (d = 0.89). A history of treatment for mental health problems was more common in the HSG (35.9% vs. 17.0%, p = 0.003), particularly for depression (59.0% vs. 29.9%, p < 0.0001). Self-reported current episodes of binge eating (27.8% vs. 14.3%, p = 0.023) were more frequent in the HSG.

In addition, there were significant differences between the groups regarding the participants’ perception of receiving comfort and protection during childhood. In the HSG, 25.3% answered that “no one” comforted them as a child, compared to 9.5% in the LSG (mean difference 15.8%, p = 0.0036). A similar difference was found when participants responded to the question “Who made you feel safe as a child?”, where 20.3% in the HSG responded “no one” compared to 6.8% in the LSG (mean difference 13.5%, p = 0.0061).

The differences in HRQoL, body image, and self-esteem between the LSG and the HSG are shown in Table 2. Apart from role physical, the HSG reported significantly lower generic HRQoL in all domains assessed with the RAND-36. The ESs indicated large differences between the groups in all mental health domains (range 0.85–1.36). Perceived body image, as assessed with the BSQ, was significantly more negative in the HSG compared to the LSG (p < 0.001) and the ES was large (d = 1.10). Self-esteem, as measured with SCQ, showed a large difference of 40.0 points between the groups (p < 0.001; d = 1.81).

Analysis of the correlation between ISS score and the other variables (e.g., physical functioning, role physical, etc.) revealed a strong negative correlation with self-esteem (ISSSELFE, r = −0.72 and SCQ, r = −0.78; Fig. 2). The largest negative correlation regarding HRQoL was for RAND-36 subscale Mental Health (r = −0.62). The physical dimensions of RAND-36 revealed weak correlations with shame. The correlation with OP scale (r = 0.55) showed a moderate correlation with internalized shame.

To the best of our knowledge, this is the first study to investigate internalized shame and its association with HRQoL, body image, and self-esteem in treatment-seeking adults with obesity. The majority of participants in our study reported similar levels of internalized shame as in a nonclinical population. However, approximately one-third of the participants reported shame levels exceeding clinical cut-off. High levels of shame were negatively associated with nearly every other variable assessed in the study, yet without any substantial difference in BMI between the HSG and LSG. Additionally, the HSG reported more lifetime and ongoing mental health problems as well as a reduced sense of comfort and protection during childhood.

Research on other conditions (e.g., post-traumatic stress disorder, obsessive-compulsive disorder, and eating disorders) has demonstrated that addressing internalized shame is crucial for achieving meaningful treatment outcomes [53]. Similarly, targeting internalized shame may be essential for improving mental health and HRQoL in treatment-seeking individuals with obesity. This reasoning may also apply to concerns with body image, given that a higher proportion of women was found in the HSG, possibly due to the increased body dissatisfaction observed within this group, previously shown by others [9].

ISS usually yields mean/median values between 23 and 34 in nonclinical populations [38, 40‒42]. In two Norwegian studies with candidates for bariatric surgery the mean ISS scores were 21.0 and 26.8 [54, 55], which might be explained by data collection during the preoperative evaluation phase, where underreporting could be assumed due to a perceived risk of being disqualified from surgery [56].

The LSG reported similar shame levels as in one of the Norwegian studies [55]. However, the overall mean ISS score in our cohort was 41.6, which is in line with the mean ISS score of 44.6 reported in an online lifestyle obesity intervention study from the USA [35]. In other clinical populations, the mean ISS score has been reported to be 47 for anxiety [39], 47 for PTSD [57], and 61 for individuals with depression and narcissistic personality disorder [58]. Therefore, we suggest that clinicians in obesity treatment should recognize internalized shame as a possible contributing factor to poor health-related quality of life in individuals living with obesity.

We found a significantly lower HRQoL in all aspects of mental health in the HSG compared to the LSG, but no or small differences in physical aspects. A substantial difference with large ES was seen between the groups regarding obesity-specific HRQoL, where the HSG reported more negative impairment in social situations due to body weight. Association between high shame and psychosocial impairment, but not necessarily physical impairment, mirrors the findings demonstrated for internalized stigma [18], indicating that both internalized stigma and internalized shame are related to psychosocial aspects of living with obesity.

We explored a possible association between attachment to important individuals early in life and internalized shame, as suggested by Schore [32]. Childhood experiences where no one provided comfort, security, or safety were twice as common in the HSG compared to the LSG. Additionally, patients’ relational patterns and experiences seemed to affect how the offered care was perceived, which is in line with previous research showing that securely attached patients with obesity perceive the interaction with health care providers more positively [59]. These findings warrant further exploration.

Another area requiring further investigation is the distinction between internalized stigma and internalized shame. Internalized shame is a broader construct that can arise independently of external stigma, as seen in early interactions between children and caregivers [60]. However, internalized stigma, particularly weight-related stigma, may also contribute to the development of internalized shame by reinforcing negative self-evaluations. Future research could examine the interplay between these constructs to clarify their boundaries, shared mechanisms, and unique pathways through which they influence psychological well-being.

Important strengths of this study include (1) the use of a validated assessment tool for internalized shame (ISS) that has previously been used in bariatric surgery populations [54, 55] and (2) participant responses not being recorded during the evaluation phase for bariatric surgery eligibility, which can be an incentive for more socially desirable responses [56]. A limited sample size and a low response rate limits the generalizability of findings. The sample may also not fully represent the demographic diversity of the treatment-seeking population. However, it is worth noting that nonresponders and responders had similar BMI and sex distributions.

Another limitation is the lack of a control group of individuals with similar levels of obesity (class II-III) but not actively seeking treatment for obesity where higher HRQoL scores have been reported [1]. As a cross-sectional study design cannot determine possible causality, future studies should explore the direction of the associations between internalized shame and HRQoL, concerns with body image and self-esteem.

Although the experience of shame varied substantially among treatment-seeking patients with obesity class II-III, approximately one-third of participants reported levels above a clinical threshold for internalized shame which indicates a need for psychological treatment. A high level of internalized shame was associated with low mental HRQoL, greater concerns about body image, and low self-esteem. Offering interventions to target shame in patients reporting high-internalized shame should be considered.

We are grateful for the help of the Regional Obesity Center in identifying and inviting potential participants to take part in the study.

The study was reviewed and approved by the Swedish Ethical Review Authority in Gothenburg, Sweden, reference number 897-18. The study was conducted in accordance with the Declaration of Helsinki [61]. Participation was anonymous; the information letter informed participants that by returning the questionnaires, participants gave their written informed consent to participate.

M.L. reports personal fees (lecturer and speaker fees) from Johnson & Johnson, Novo Nordisk, Baricol Bariatrics, and GB Obesitas. T.O. reports reimbursement for participation in advisory board and educational activities from Novo Nordisk and Johnson & Johnson directed to his institution. K.J. reports reimbursement for educational activities from Novo Nordisk and Johnson & Johnson directed to her clinical institution. P.J., B.O., and M.E. report no conflicts of interest.

This work was supported by the Healthcare Board, Region Västra Götaland, SE 462 80 Vänersborg, Sweden Grant No. VGFOUREG-967527’. The funder had no role in the design, data collection, data analysis, and reporting of this study.

M.L., M.E., T.O., and P.J. designed the study. M.L. performed the data collection and the statistical analyses. M.L. and M.E. drafted the article. M.L., M.E., T.O., K.J., P.J., and B.O. undertook revisions and contributed intellectually to the development of this paper. All authors have approved the final draft of the manuscript.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from the corresponding author M.L. upon reasonable request.