Development and validation of the body-related distress scale for obesity (BDS-O): a patient-reported outcome measure for Chinese population with obesity Authors: Wanyang Li, Hongyu Chen, Shuyao Dong, Hao Bai, QingLu Wu, Dan Wang, Wenbin Yu, Liyong Chen Published: 2026 ABSTRACT Body dissatisfaction has significant negative effects on Chinese individuals with obesity. However, tools used for measuring body-related distress that are culturally relevant to the condition of obesity remain underdeveloped. To address this, we developed and validated the Body-Related Distress Scale for Obesity (BDS-O), a culturally appropriate patient-reported outcome (PRO) scale for Chinese adults with BMI ≥ 28.0. An initial item pool was developed through qualitative semi-structured interviews, guided by the Symptom Experience Model and international PRO standards. These items were further refined through cognitive interviews and a Delphi procedure involving ten experts in the medical field. Following multiple revisions based on these cognitive interviews and expert feedback, the final scale was established to measure 25 individual statements. The factorial analysis process involved 415 participants to determine the constructs. The model analysis revealed that these constructs explained 70.45% of the variance. The constructs also demonstrated good model fit. The reliability measures for the body image and obesity constructs were also highly consistent, with composite reliability 0.86–0.93.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s40337-026-01685-9. Introduction Obesity has been recognized as a significant threat to international public health, with broad-ranging consequences in terms of population-level rates of illness and mortality [1, 2]. According to national representative surveys in China, over half of the Chinese population suffers from obesity. Although the biomedical impacts of obesity (e.g., its direct link to type 2 diabetes mellitus, cardiovascular disorders, and certain cancers) are well-documented in the literature, its psychosocial correlates and their far-reaching consequences remain neglected in the literature. On an individual level, people living with obesity face stigmatization and social ostracism on the grounds of their weight, as well as shame and self-blame, all of which tend to decrease their self-esteem and contribute to low psychological well-being. Body-related distress Body-related distress, including body dissatisfaction, perceived weight stigma, social evaluation, and emotional burden, can negatively affect Chinese individuals with obesity [3]. Body dissatisfaction is defined as having negative thoughts and feelings about one’s body and is a key component of the psychosocial impact of obesity [4]. Body dissatisfaction often emerges in early childhood [5, 6]. It becomes more pronounced as individuals start engaging more in social interaction and are increasingly exposed to social media and frequently persists into later adulthood [7]. Body dissatisfaction is a widespread, global issue affecting both Western and non-Western populations, with particularly high rates among adolescents and young women [8, 9]. It also tends to worsen as young people progress through adolescence and early adulthood [10]. Body dissatisfaction is linked to a variety of adverse consequences, including low self-esteem, elevated psychological distress (e.g., anxiety and depression), disordered eating, excessive exercise, and reduced social functioning [5, 11–13]. Patient-reported outcomes (PROs) Under the current paradigm of patient-centered care, body image concerns—like many other subjective experiences in patient health—are best measured using PROs [14, 15]. According to the United States’ Food and Drug Administration (FDA), PROs refer to measurements based on reports directly from the individual being inquired of that describe the status of the individual’s health condition, without modification or interpretation by anyone else, and only concerning the specific individual’s condition [16]. By focusing on patients’ own reports of their experiences (e.g., the symptoms they might exhibit, the functions they might no longer undertake through their own bodily capabilities, and their perceived level of wellness), PROs are an important tool in medical research [17]. PROs play an irreplaceable role by providing an objective dimension to current patient assessment processes. Several scales have been developed to assess body dissatisfaction, including the Body Shape Questionnaire (BSQ), the Eating Disorder Inventory-3 (EDI-3), the Body Dissatisfaction subscale of the Eating Attitudes Test-26 (EAT-26), the Body Dissatisfaction scale of the Eating Disorder Inventory (EDI-BD), and the Adolescent Body Image Assessment Scale [18]. Among these, the Body Shape Questionnaire (BSQ) is a classic self-report scale for assessing patients’ concerns around their weight and body shape, with comprehensive measurement dimensions covering the core symptoms of body image disturbance and a wide scope of application [19]. The EDI-3 is a comprehensive multidimensional self-reporting scale with excellent psychometric properties and a structured framework, which can provide a basis for clinical case conceptualization and treatment planning [20]. The EAT-26 is a classic, brief self-report scale for eating disorder risk screening that features streamlined items and ease of administration and interpretation. It is suitable for the preliminary screening of large-scale populations and cross-age use, and capable of quickly identifying core high-risk factors such as obsessive dieting and body dissatisfaction [21]. However, these scales all have significant inherent limitations, as most were originally designed for the general population or patients with diagnosed eating disorders, with development samples largely excluding individuals with obesity. Consequently, the adaptation and validation work for this specific group is very limited. More specifically, the applicability of the BSQ for male populations has not been sufficiently validated [22]. It is not specific to individuals with obesity, and it adopts a single-dimensional scoring method. Similarly, the EDI-3 relies heavily on Western female samples, resulting in poor generalizability to male and non-Western populations [23]. The EAT-26 focuses exclusively on eating-related attitudes and behaviors without incorporating any accompanying psychological and emotional characteristics [24]. In addition to these scale-specific flaws, all the assessment tools share common critical shortcomings when used to evaluate body dissatisfaction in individuals with obesity. Specifically, they overemphasize aesthetic aspects of body image (e.g., “I think my thighs are too large”) while ignoring other obesity-specific issues that are highly relevant to this group (e.g., weight-related social discrimination and stigmatization, the physical discomfort caused by obesity, limited physical functionality, and reduced social engagement) [25]. Few of these scales assess the emotional distress caused by body dissatisfaction, and several lack validation for populations with obesity. Consequently, they fail to reflect the current research understanding of body dissatisfaction in this group [26]. Given these limitations of existing measurement tools, there is an urgent need to develop new PRO scales specifically tailored to assess body dissatisfaction in individuals with obesity. This would provide a more reliable and targeted assessment tool for clinical practice and scientific research in this field. The symptom experience model To overcome these measurement shortcomings, the Symptom Experience Model serves as an appropriate theoretical framework for designing a new PRO instrument [27]. It views symptoms as complex subjective processes through the co-constructions developed by an individual’s perception processes rather than as individual biomedical events [1]. It conducts a comprehensive measurement of symptoms by examining four interrelated domains, namely [28, 29]: (1) the frequency of symptoms (incidence rate); (2) the severity of the symptoms (intensity); (3) disturbances in emotional function caused by the symptoms (bother); and (4) interpretations of personal identity due to the symptoms (meaning). By incorporating these four domains into the instrument’s measurement process within this theoretical framework (which is distinct from the biomedical paradigm), the concerned individual’s role within their social context becomes more prominent [28, 30, 31]. Although these four domains are conceptually distinguishable, they are theoretically intertwined and jointly constitute an individual’s subjective experience of their symptoms. In this study, these domains were referenced to guide the development of the scale’s items and measurements, rather than as separate factors in the final factor structure. Study aims The aim of this study is to develop and validate a new PRO measure—the Body-Related Distress Scale for Obesity (BDS-O). By following the best international practices for scale development, namely, the FDA’s Guidance for Industry: Patient-Reported Outcome Measures (2009) and the COSMIN checklist [32], the proposed scale will provide adequate content validity, translational acceptability, and psychometric properties. This scale will help in: (1) psychosocial risk assessment for individuals; (2) designing interventions; (3) assessing the success of interventions through body dissatisfaction measures rather than weight-related measures; and (4) policy-making related to the alleviation of weight-related stigmatization. Methods This trial protocol was reviewed for approval by the Medical Ethics Committee of the Qilu Hospital of Shandong University (Approval No.: KYLL-2025-07-053-1), and aligns with the ethical standards outlined in the Declaration of Helsinki (2013 revision). All participants provided informed consent prior to taking part in the survey. Study procedure Phase 1 Semi-structured interviews were conducted using thematic analysis to extract core themes and construct an initial item pool for the scale. The items were then preliminarily refined through cognitive interviews and a pilot test. Phase 2 A 10-member expert panel conducted a review to screen each item and form a preliminary scale. This was followed by item analysis to eliminate inadequate items and ultimately determine the final 25-item scale. Phase 3 A cross-sectional survey was conducted with 415 participants. Exploratory factor analysis, confirmatory factor analysis, as well as reliability, validity and measurement invariance tests were performed to validate and evaluate the scale. Study population The cross-sectional survey was administered at the Qilu Hospital of Shandong University from September 2025 to November 2025. The subject population consisted of Chinese adults with obesity, recruited from the same hospital. This population represents a group at high risk for obesity-related complications and is the target group for clinical weight management interventions. Validating the questionnaire in this critical population is essential for developing evidence-based strategies to guide weight loss and improve clinical outcomes for individuals. Chinese individuals aged 18 years or older with a Body Mass Index (BMI) of 28.0 kg/m² or greater are defined as individuals with obesity, according to the Guidelines for Prevention and Control of Overweight and Obesity in Chinese Adults (2003) [33]. The relevant data were collected online. Inclusion Criteria: At least 18 years of age.Body mass index (BMI) of at least 28.0 kg/m², calculated from height and weight measures. Height and weight were measured by trained clinical staff using standardized procedures and calibrated instruments. Immediately after measurement, participants entered their height and weight directly into the online questionnaire system as part of a paper‑less data collection process. All submitted data were subsequently checked and verified by research staff to ensure data accuracy and integrity.Literacy in Chinese, adequate for completing the survey independently.Eagerness to participate in the research (as proven by signing the consent form).Not undergoing any concurrent obesity-related pharmacotherapy or surgical interventions, except the receipt of routine lifestyle counseling (i.e., diet or exercise guidance) in the outpatient clinic during the study period. At least 18 years of age. At least 18 years of age. Body mass index (BMI) of at least 28.0 kg/m², calculated from height and weight measures. Height and weight were measured by trained clinical staff using standardized procedures and calibrated instruments. Immediately after measurement, participants entered their height and weight directly into the online questionnaire system as part of a paper‑less data collection process. All submitted data were subsequently checked and verified by research staff to ensure data accuracy and integrity. Body mass index (BMI) of at least 28.0 kg/m², calculated from height and weight measures. Height and weight were measured by trained clinical staff using standardized procedures and calibrated instruments. Immediately after measurement, participants entered their height and weight directly into the online questionnaire system as part of a paper‑less data collection process. All submitted data were subsequently checked and verified by research staff to ensure data accuracy and integrity. Literacy in Chinese, adequate for completing the survey independently. Literacy in Chinese, adequate for completing the survey independently. Eagerness to participate in the research (as proven by signing the consent form). Eagerness to participate in the research (as proven by signing the consent form). Not undergoing any concurrent obesity-related pharmacotherapy or surgical interventions, except the receipt of routine lifestyle counseling (i.e., diet or exercise guidance) in the outpatient clinic during the study period. Not undergoing any concurrent obesity-related pharmacotherapy or surgical interventions, except the receipt of routine lifestyle counseling (i.e., diet or exercise guidance) in the outpatient clinic during the study period. Exclusion Criteria: Since severe psychiatric disorders or significant cognitive impairment may affect the authenticity and reliability of questionnaire responses, participants with a history of major psychiatric or cognitive disorders were excluded. This information was obtained and verified by medical record reviews when available to ensure that the participants had the capacity to understand and complete the questionnaire. Sample Size: A total of 415 participants were divided into two subsamples, namely, 207 for the exploratory factor analysis (EFA) and 208 for the confirmatory factor analysis (CFA). These sample sizes were determined based on the requirements for EFA (≥ 100 cases) and CFA (≥ 200 cases) [34–36]. Study instrument The scale was developed based on the PRO framework and the Symptom Experience Model, and employs a five-point Likert scale response format ranging from 1 (‘Does Not Apply At All’) to 5 (‘Applies Completely’). Item pool development based on qualitative inquiry In-depth, semi-structured qualitative interviews were conducted to inform the item generation process. Purposive sampling was employed to identify and recruit participants with self-perceived weight-related concerns. A total of 14 participants were interviewed, with each interview lasting an average of 34.5 min. The interview data were thematically analyzed, with two researchers independently identifying recurring core concepts and integrating them into main themes through discussion until consensus was reached. The key extracted themes included: weight-related awareness and negative affective responses; worrying about weight gain; perceived stigma, judgment, and negative evaluation from family, peers, and significant others; body comparison and dissatisfaction with one’s physical shape; social avoidance, loneliness, and perceived reduced attractiveness due to weight; weight-loss intentions, behaviors, and perceived lack of self-control; and self-attribution for weight gain and emotional distress related to perceived failure to achieve an ideal body shape. An item pool was developed based on these themes. The final items were determined following expert review by 10 experts and pilot testing with five participants. Each item directly reflected the core content of the qualitative themes, as shown in Supplementary Table S1. Expert review for content validity Each item was reviewed by a 10-person expert panel consisting of two clinical nutritionists, a general surgeon, three public health researchers, two psychometrists, and two clinical nurses. Each item was rated independently by the panel members for relevance, clarity, and representativeness on a four-point Likert scale (1 = ‘not relevant’, 2 = ‘somewhat relevant’, 3 = ‘relevant’, 4 = ‘highly relevant’). The item-level CVI (I-CVI) was performed using the average method (S-CVI/Ave). Any item attaining an I-CVI of ≥ 0.78 and a S-CVI of ≥ 0.80 or higher was selected. This resulted in the generation of the 28-item scale for field testing [28]. Statistical analysis Item analysis Item discrimination was performed via the critical ratio method. Participants who scored in the top 27% of the total scores were assigned to the high-scoring group, and those who scored in the bottom 27% of the total scores were assigned to the low-scoring group. Independent samples t-tests were used to evaluate both groups’ scores on each individual item. Only items with P < 0.05 were retained. The pearson correlation coefficient was used to calculate the correlation between each item score and the total scale score. Items with an item-total correlation of r ≥ 0.40 were retained. Reliability assessment The internal scale consistency for the overall scale and its subscales was estimated using Cronbach’s α and McDonald’s ω. An index value of 0.80 or higher was considered adequate [26, 37, 38]. Validity assessment Since the four domains reflect complementary aspects of a unified symptom experience construct, a total sum score was calculated to represent the overall symptom experience. Although this composite score is parsimonious, clinically meaningful, and facilitates interpretation and cross-study comparison, it may potentially mask domain-specific patterns. Construct validity A two-phase method was used to assess the validity of the constructs. EFA was used to determine the factorial structure of the initial sub-sample, employing principal component analysis (PCA) with varimax rotation [39, 40]. EFA was performed if the Kaiser-Meyer & Olkin (KMO) value was > 0.80 and the significance for the likelihood of sphericity (as determined via Bartlett’s test) was < 0.05. Factors included in the analysis were determined by eigenvalues > 1 (Kaiser criterion), scree testing, and meaningful interpretation. Items were retained if they loaded ≥ 0.40 on a single factor. No cross-loading was allowed (i.e., items loading < 0.20 on different factors) [41, 42]. To examine the robustness of the factor structure, a sensitivity analysis with Direct Oblimin oblique rotation was conducted. CFA CFA was conducted on a separate subsample to assess the factor model fit as specified through EFA [43]. The estimation was conducted via the maximum likelihood method. Model fit was indexed based on the criteria: χ²/df < 3 (indices < 5 were indicative of good fit), comparative fit index (CFI) > 0.90 (indices > 0.95 were indicative of excellent fit), Tucker-Lewis index (TLI) > 0.90, root mean square error of approximation (RMSEA) < 0.08 (indices < 0.06 were indicative of good fit), and standardized root mean square residual (SRMR) < 0.08. Meeting these criteria indicated good construct validity [44, 45]. The CFA results enabled the estimation of the average variance extracted (AVE) and the composite reliability of each latent variable. If the AVE was > 0.50 and the composite reliability was > 0.80, the scale was assumed to accurately represent the constructs for each field of study [45]. The Fornell-Larcker criterion was used to assess the model’s discriminant validity. According to the criterion, the model was taken to have discriminant validity if the square root of the AVE values for each construct was greater than the correlation between the constructs. In addition, the Heterotrait-Monotrait ratio of correlations (HTMT) was employed as a stricter criterion. As suggested in prior research, discriminant validity is confirmed if all the HTMT values are below the 0.85 threshold [46]. Measurement invariance across gender Multigroup confirmatory factor analysis was performed to test gender-based measurement invariance. The validated four-factor structure was established separately for male and female samples. We sequentially constructed nested constrained models including configural, metric, structural covariance and measurement residual invariance models, with the configural model served as the baseline. Equality constraints were gradually imposed on factor loadings, latent factor covariances and item residuals. Beyond traditional chi-square difference test, we adopted changes in practical fit indices including ΔCFI,ΔRMSEA and ΔSRMR as core evaluation criteria. Invariance was deemed acceptable if model fit did not deteriorate obviously after constraint imposition, generally with ΔCFI ≤ 0.01 and ΔRMSEA ≤ 0.015. Given that chi-square difference test is easily affected by sample size, we interpreted measurement invariance synthetically based on both chi-square differences and variations in fit indices [47]. All the statistical analysis work was performed using R version 4.4.2 and GraphPad Prism 10. Statistical significance was set at α = 0.05 for all two-tailed tests. Results Data overview A total of 415 valid respondents were retained for analysis. Using the random number generator function, the respondents were split equally into two separate subsamples—one for EFA (Subsample A, consisting of 207 respondents) and the other for CFA (Subsample B, consisting of 208 respondents). As shown in Table 1, the subsamples were compared based on all the selected demographic variables including age, BMI, gender, marital status, educational attainment, occupation, dietary behavior, smoking habits, alcohol consumption, hours of sleep per night, levels of physical activity, and secondary complications, all of which exhibited values of P > 0.05 (indicating no statistically significant differences). Content validity To determine the content validity of the BDS-O scale, the scale items were evaluated from the perspectives of ten experts in terms of their relevance to the construct and clarity. The item-level content validity index (I-CVI) and the scale-level content validity index average (S-CVI/AVE) values were calculated. Items with I-CVI values ≥ 0.78 and S-CVI/AVE values ≥ 0.80 were retained. Table 1Demographic and clinical characteristics of the study sampleVariableTotal Sample(n = 415)Age, M (SD)27.92 ± 6.96BMI*, M (SD)32.34 ± 11.34Mild obesity (28.0 ≤ BMI<32.5 kg/m²), n (%)205 (49.40%)Moderate obesity(32.5 ≤ BMI<37.5 kg/m²), n (%)142 (34.22%)Severe obesity(BMI ≥ 37.5 kg/m²), n (%)68 (16.38%)Sex Male, n (%)247 (59.52%) Female, n (%)168 (40.48%)Marital status Single, n (%)258 (62.17%) Married, n (%)141 (33.98%) Divorced/Widowed, n (%)16 (3.85%)Education level High school or below, n (%)40 (9.64%) Bachelor’s degree, n (%)360 (86.75%) Graduate degree and above, n (%)15 (3.61%)Occupation Civil servant, n (%)82 (19.76%) Teacher, n (%)32 (7.71%) Military Personnel, n (%)3 (0.72%) Retired, n (%)1 (0.24%) Commercial and service industry personnel, n (%)179 (43.13%) Workers and professional technical personnel, n (%)86 (20.72%) None, n (%)32 (7.72%) Other, n (%)0 (0.00%)Dietary habits* Light taste, n (%)26 (6.27%) Normal taste, n (%)128 (30.84%) Heavy taste, n (%)261 (62.89%)Smoking status Never smoked, n (%)239 (57.59%) Occasional smoker (<10/day), n (%)100 (24.10%) Daily smoker (10–20/day), n (%)57 (13.73%) Heavy smoker (≥ 20/day), n (%)15 (3.61%) Quit smoking, n (%)4 (0.97%)Alcohol consumption Never, n (%)110 (26.51%) Occasionally, n (%)163 (39.28%) 1–2 times/week, n (%)70 (16.87%) 3–5 times/week, n (%)46 (11.08%) ≥5 times/week, n (%)26 (6.26%)Sleep duration <6 h, n (%)46 (11.08%) 6–9 h, n (%)334 (80.48%) >9 h, n (%)35 (8.44%)Physical exercise No, n (%)189 (45.54%) Yes, n (%)226 (54.46%)Complications Hypertension, n (%)68 (16.39%) Hyperlipidemia, n (%)51 (12.29%) Diabetes, n (%)45 (10.84%) Others, n (%)5 (1.20%) Multiple complications, n (%)82 (19.76%) No complications, n (%)164 (39.52%)*Body Mass Index, BMI= weight (kg)/height (m)2**Light taste (less salt, less oil); Heavy taste (high salt, high oil); Multiple complications refer to two or more comorbidities. Continuous variables were compared between groups using independent samples t-tests, and categorical variables were analyzed using chi-square tests; a p-value < 0.05 was considered statistically significant Table 1Demographic and clinical characteristics of the study sampleVariableTotal Sample(n = 415)Age, M (SD)27.92 ± 6.96BMI*, M (SD)32.34 ± 11.34Mild obesity (28.0 ≤ BMI<32.5 kg/m²), n (%)205 (49.40%)Moderate obesity(32.5 ≤ BMI<37.5 kg/m²), n (%)142 (34.22%)Severe obesity(BMI ≥ 37.5 kg/m²), n (%)68 (16.38%)Sex Male, n (%)247 (59.52%) Female, n (%)168 (40.48%)Marital status Single, n (%)258 (62.17%) Married, n (%)141 (33.98%) Divorced/Widowed, n (%)16 (3.85%)Education level High school or below, n (%)40 (9.64%) Bachelor’s degree, n (%)360 (86.75%) Graduate degree and above, n (%)15 (3.61%)Occupation Civil servant, n (%)82 (19.76%) Teacher, n (%)32 (7.71%) Military Personnel, n (%)3 (0.72%) Retired, n (%)1 (0.24%) Commercial and service industry personnel, n (%)179 (43.13%) Workers and professional technical personnel, n (%)86 (20.72%) None, n (%)32 (7.72%) Other, n (%)0 (0.00%)Dietary habits* Light taste, n (%)26 (6.27%) Normal taste, n (%)128 (30.84%) Heavy taste, n (%)261 (62.89%)Smoking status Never smoked, n (%)239 (57.59%) Occasional smoker (<10/day), n (%)100 (24.10%) Daily smoker (10–20/day), n (%)57 (13.73%) Heavy smoker (≥ 20/day), n (%)15 (3.61%) Quit smoking, n (%)4 (0.97%)Alcohol consumption Never, n (%)110 (26.51%) Occasionally, n (%)163 (39.28%) 1–2 times/week, n (%)70 (16.87%) 3–5 times/week, n (%)46 (11.08%) ≥5 times/week, n (%)26 (6.26%)Sleep duration <6 h, n (%)46 (11.08%) 6–9 h, n (%)334 (80.48%) >9 h, n (%)35 (8.44%)Physical exercise No, n (%)189 (45.54%) Yes, n (%)226 (54.46%)Complications Hypertension, n (%)68 (16.39%) Hyperlipidemia, n (%)51 (12.29%) Diabetes, n (%)45 (10.84%) Others, n (%)5 (1.20%) Multiple complications, n (%)82 (19.76%) No complications, n (%)164 (39.52%)*Body Mass Index, BMI= weight (kg)/height (m)2**Light taste (less salt, less oil); Heavy taste (high salt, high oil); Multiple complications refer to two or more comorbidities. Continuous variables were compared between groups using independent samples t-tests, and categorical variables were analyzed using chi-square tests; a p-value < 0.05 was considered statistically significant Demographic and clinical characteristics of the study sample *Body Mass Index, BMI= weight (kg)/height (m)2 **Light taste (less salt, less oil); Heavy taste (high salt, high oil); Multiple complications refer to two or more comorbidities. Continuous variables were compared between groups using independent samples t-tests, and categorical variables were analyzed using chi-square tests; a p-value < 0.05 was considered statistically significant Initial item analysis An initial item analysis was performed on the pilot measurement instrument using the critical ratio method and the item-total correlation. The study participants were split according to the preliminary BDS-O scores into a high-scoring group (top 27%) and a low-scoring group (bottom 27%). The results of the independent samples t-tests for all 28 questions revealed statistically significant differences (P < 0.01) between the two groups for 25 of the questions. This indicates that these 25 statements had good discrimination ability. The item-total correlation analysis also revealed that these 25 statements were significantly associated with the total scale score (P < 0.01). Based on the item analysis results, several items were removed, including one item with an insignificant critical ratio, one item with an item-total correlation coefficient close to zero, and one item because its removal resulted in an increase in the Cronbach’s alpha coefficient. Thus, the 25-item scale was developed. Internal consistency for the full 25-item scale was high (α = 0.925, ω = 0.926). EFA KMO sampling adequacy was 0.925. The significance of the Barlett Test of Sphericity was very high (chi-square = 3657.451, freedom = 300, P < 0.001). This indicated that the data were suitable for performing factor analysis. Principal component analysis was used to extract the factors. Initially, the factors that fulfilled the eigenvalue > 1 rule were extracted. Four factors explained 70.45 of the variance. During the item refinement process, the criteria for the deletion of items included: (1) factor loading < 0.40; (2) cross-loading ≥ 0.40 on more than one factor, with the difference between the largest and the second largest loading < 0.30; (3) fewer than three items per factor; and (4) items with conceptually inconsistent item-to-factor alignment. Following iterative analysis, all 25 items from the initial pool were retained. Through parallel analysis, the squared multiple correlation (SMC) values were used to determine the appropriate number of factors to extract. The analysis revealed the eigenvalues for the first four factors to be larger than those of the randomly generated data (Fig. 1), suggesting that the fourth factor should be retained. Taken together, these results provide strong evidence for the robustness of the fourth-factor model. The item alignments for each proposed factor are summarized in Table 2. Fig. 1Scree plot from parallel analysis. The horizontal axis shows factor numbers ranging from 1 to 25. The vertical axis represents eigenvalues, indicating the amount of variance explained by each factor Scree plot from parallel analysis. The horizontal axis shows factor numbers ranging from 1 to 25. The vertical axis represents eigenvalues, indicating the amount of variance explained by each factor Four factors emerged from the exploratory factor analysis that explained 70.45% of the variance. Factor 1 was named “Body Image and Self-Perception,” and included five items. Factor 2 was named “Social Comparison and Perceived Evaluation by Others,” consisting of 11 items. Factor 3 was named “Emotional Response and Psychological Burden,” including 5 items. Factor 4 was named “Weight Management Behaviors and Sense of Control,” and consisted of 4 items. Table 2Results of exploratory factor analysisItemFactor1Factor2Factor3Factor4CommunalityItem 10.7580.674Item 30.8360.722Item 160.8090.718Item 200.7710.679Item 240.8010.759Item 50.7900.709Item 60.7940.676Item 70.8060.700Item 80.8030.69Item 90.8120.700Item 100.8320.727Item 110.8390.735Item 120.8200.724Item 150.8240.716Item 170.8280.714Item 130.8150.679Item 20.8010.689Item 40.7730.647Item 180.7810.675Item 190.7890.679Item 250.8540.768Item 140.7810.722Item 210.7810.694Item 220.7830.689Item 230.7960.725Variance Explained30.60614.28114.15011.408Cumulative Variance Explained30.60644.88759.03770.445 Table 2Results of exploratory factor analysisItemFactor1Factor2Factor3Factor4CommunalityItem 10.7580.674Item 30.8360.722Item 160.8090.718Item 200.7710.679Item 240.8010.759Item 50.7900.709Item 60.7940.676Item 70.8060.700Item 80.8030.69Item 90.8120.700Item 100.8320.727Item 110.8390.735Item 120.8200.724Item 150.8240.716Item 170.8280.714Item 130.8150.679Item 20.8010.689Item 40.7730.647Item 180.7810.675Item 190.7890.679Item 250.8540.768Item 140.7810.722Item 210.7810.694Item 220.7830.689Item 230.7960.725Variance Explained30.60614.28114.15011.408Cumulative Variance Explained30.60644.88759.03770.445 Results of exploratory factor analysis A sensitivity analysis using direct oblimin oblique rotation replicated the 4-factor structure, with all items loading on the same factors as in the orthogonal solution (see Supplementary Table S2 for the pattern matrix and Supplementary Table S3 for the factor correlation matrix). Two items (B9 and C5) showed minor secondary cross-loadings, but their primary loadings on the intended factors remained substantially higher, indicating no meaningful impact on the factor structure. Confirmatory factor analysis Structural validity Confirmatory factor analysis indicated that the four-factor model had adequate model fit (Table 3). Table 3Model fit indices for the structural equation modelIndexX2/dfRMSEACFIIFITLIValue1.4800.0480.9570.9570.952Standard< 3< 0.05> 0.90> 0.90> 0.90 Table 3Model fit indices for the structural equation modelIndexX2/dfRMSEACFIIFITLIValue1.4800.0480.9570.9570.952Standard< 3< 0.05> 0.90> 0.90> 0.90 Model fit indices for the structural equation model Convergent validity As shown in Table 4, the outcome of the convergent validity test showed that the model fulfilled all the psychometric requirements: each item demonstrated a standardized loading value > 0.5 (Table 4; Fig. 2), the extracted average variance values for all the constructs were > 0.5 (ranging between 0.537 and 0.690), and the values for the composite reliability values were all > 0.7 (ranging from 0.861 to 0.927). Table 4Factor LoadingsPathEstimateAVEComposite ReliabilitySEItem 1<---F10.7640.5550.861-Item 3<---F10.6720.097Item 16<---F10.7930.11Item 20<---F10.8400.101Item 24<---F10.6370.115Item 5<---F20.6710.5370.927–Item 6<---F20.7410.121Item 7<---F20.7810.125Item 8<---F20.7300.12Item 9<---F20.7700.122Item 10<---F20.7890.106Item 11<---F20.6550.118Item 12<---F20.7520.12Item 13<---F20.7290.118Item 15<---F20.8040.119Item 17<---F20.6160.121Item 2<---F30.7800.6030.883–Item 4<---F30.8440.085Item 18<---F30.7450.091Item 19<---F30.7590.093Item 25<---F30.7500.097Item 14<---F40.9770.6900.897–Item 21<---F40.9270.039Item 22<---F40.6440.056Item 23<---F40.7300.058AVE, Average Variance Extracted; SE, standardized error Table 4Factor LoadingsPathEstimateAVEComposite ReliabilitySEItem 1<---F10.7640.5550.861-Item 3<---F10.6720.097Item 16<---F10.7930.11Item 20<---F10.8400.101Item 24<---F10.6370.115Item 5<---F20.6710.5370.927–Item 6<---F20.7410.121Item 7<---F20.7810.125Item 8<---F20.7300.12Item 9<---F20.7700.122Item 10<---F20.7890.106Item 11<---F20.6550.118Item 12<---F20.7520.12Item 13<---F20.7290.118Item 15<---F20.8040.119Item 17<---F20.6160.121Item 2<---F30.7800.6030.883–Item 4<---F30.8440.085Item 18<---F30.7450.091Item 19<---F30.7590.093Item 25<---F30.7500.097Item 14<---F40.9770.6900.897–Item 21<---F40.9270.039Item 22<---F40.6440.056Item 23<---F40.7300.058AVE, Average Variance Extracted; SE, standardized error Factor Loadings AVE, Average Variance Extracted; SE, standardized error Fig. 2Measurement Model of the Body-Related Distress Scale for Obesity (BDS-O). All factor loadings are standardized and statistically significant (p < 0 0.05).F1, Body Image and Self-Perception; F2, Social Comparison and Perceived Evaluation by Others; F3, Emotional Response and Psychological Burden; F4, Weight Management Behaviors and Sense of Control; Term 1, I pay attention to my body weight; Term 2, Weighing myself makes me feel discouraged; Term 3, When I think about having too much fat on my body, I feel repulsed; Term 4, I worry that I might gain more weight; Term 5, I worry that others will notice I have gained weight; Term 6, I feel uncomfortable or hurt when someone calls me “fat”; Term 7, When I compare my body shape to that of my peers, I am dissatisfied with my own; Term 8, I feel that only people who also struggle with their weight can truly understand me; Term 9, My parents think I am overweight; Term 10, Other family members besides my parents think I am overweight; Term 11, People who matter to me believe I am fat; Term 12, My peers think I am fat; Term 13, Others are embarrassed to be seen with me because of my weight; Term 14, I consider buying clothes and accessories that make me look slimmer; Term 15, I envy people who are thinner than I am; Term 16, Even when others say I look normal, I still feel I am too fat; Term 17, I avoid posting photos of myself on social media because I worry about comments on my weight; Term 18, I skip social activities because I am overweight; Term 19, I feel lonely because of my obesity; Term 20, I feel less attractive than others because I am overweight; Term 21, If circumstances allowed, I would consider weight-loss surgery; Term 22, I have tried to lose weight but gave up halfway; Term 23, I do not have enough self-control to maintain a healthy weight; Term 24, I believe I am responsible for my weight gain; Term 25, I become upset when I think about not being able to achieve my ideal body shape Measurement Model of the Body-Related Distress Scale for Obesity (BDS-O). All factor loadings are standardized and statistically significant (p < 0 0.05).F1, Body Image and Self-Perception; F2, Social Comparison and Perceived Evaluation by Others; F3, Emotional Response and Psychological Burden; F4, Weight Management Behaviors and Sense of Control; Term 1, I pay attention to my body weight; Term 2, Weighing myself makes me feel discouraged; Term 3, When I think about having too much fat on my body, I feel repulsed; Term 4, I worry that I might gain more weight; Term 5, I worry that others will notice I have gained weight; Term 6, I feel uncomfortable or hurt when someone calls me “fat”; Term 7, When I compare my body shape to that of my peers, I am dissatisfied with my own; Term 8, I feel that only people who also struggle with their weight can truly understand me; Term 9, My parents think I am overweight; Term 10, Other family members besides my parents think I am overweight; Term 11, People who matter to me believe I am fat; Term 12, My peers think I am fat; Term 13, Others are embarrassed to be seen with me because of my weight; Term 14, I consider buying clothes and accessories that make me look slimmer; Term 15, I envy people who are thinner than I am; Term 16, Even when others say I look normal, I still feel I am too fat; Term 17, I avoid posting photos of myself on social media because I worry about comments on my weight; Term 18, I skip social activities because I am overweight; Term 19, I feel lonely because of my obesity; Term 20, I feel less attractive than others because I am overweight; Term 21, If circumstances allowed, I would consider weight-loss surgery; Term 22, I have tried to lose weight but gave up halfway; Term 23, I do not have enough self-control to maintain a healthy weight; Term 24, I believe I am responsible for my weight gain; Term 25, I become upset when I think about not being able to achieve my ideal body shape Discriminant validity The inter-factor correlation values ranged from 0.316 to 0.348 (Table 5). The square roots of the extracted average variance for the factors ranged from 0.733 to 0.831 (all > 0.5), thus outperforming the inter-factor correlation values. These results provide support for the model’s discriminant validity. The results in Tables 5 and 6 indicate that the Fornell-Larcker criterion and the HTMT ratio both support adequate discriminant validity among the constructs in this study. Table 5Discriminant validityFactor 1Factor 2Factor 3Factor 4Factor 10.555Factor 20.3190.537Factor 30.3480.3160.603Factor 40.3420.3170.3390.690√AVE0.7450.7330.7770.831*Values on the diagonal are the square roots of the average variance extracted (√AVE), √AVE=square root of AVE Table 5Discriminant validityFactor 1Factor 2Factor 3Factor 4Factor 10.555Factor 20.3190.537Factor 30.3480.3160.603Factor 40.3420.3170.3390.690√AVE0.7450.7330.7770.831*Values on the diagonal are the square roots of the average variance extracted (√AVE), √AVE=square root of AVE Discriminant validity *Values on the diagonal are the square roots of the average variance extracted (√AVE), √AVE=square root of AVE Table 6Heterotrait-monotrait ratio of correlations (HTMT)Factor 1Factor 2Factor 3Factor 4Factor 11.000Factor 20.3451.000Factor 30.4520.2841.000Factor 40.2910.1380.2251.000*The diagonal elements represent the correlations within the same dimension, while the off-diagonal values represent the HTMT ratios between dimensions. Table 6Heterotrait-monotrait ratio of correlations (HTMT)Factor 1Factor 2Factor 3Factor 4Factor 11.000Factor 20.3451.000Factor 30.4520.2841.000Factor 40.2910.1380.2251.000*The diagonal elements represent the correlations within the same dimension, while the off-diagonal values represent the HTMT ratios between dimensions. Heterotrait-monotrait ratio of correlations (HTMT) *The diagonal elements represent the correlations within the same dimension, while the off-diagonal values represent the HTMT ratios between dimensions. Measurement invariance across gender test results The configural invariance model showed satisfactory fitting effects across male and female groups, with χ² = 756.094, df = 538, χ²/df = 1.405, CFI = 0.930, TLI = 0.922, RMSEA = 0.044 and SRMR = 0.065, which verified that the four-factor structure of the BDS-O scale was stable in both gender groups(Table 7). The factor loading invariance model also reached acceptable model fit (χ² = 783.695, df = 559, χ²/df = 1.402, CFI = 0.928, TLI = 0.923, RMSEA = 0.044, SRMR = 0.068). Compared with the configural invariance model, the changes in practical fitting indices were extremely small, with ΔCFI = -0.002, ΔRMSEA = 0.000 and ΔSRMR = 0.003. The chi-square difference test was not significant (Δχ² = 27.600, Δdf = 21, P = 0.152). In general, constraining factor loadings did not obviously reduce model fitting degree, and metric invariance was supported, indicating that items had consistent correlation degrees with corresponding latent factors across genders. For structural covariance invariance model, the fitting indices were χ² = 805.081, df = 569, χ²/df = 1.415, CFI = 0.924, TLI = 0.920, RMSEA = 0.045 and SRMR = 0.094. Relative to the baseline model, ΔCFI = -0.006 and ΔRMSEA = 0.001 were within a narrow range, while ΔSRMR = 0.029 increased obviously. Meanwhile, the chi-square difference reached statistical significance (Δχ² = 48.987, Δdf = 31, P = 0.021). The measurement residual invariance model showed relatively poor fitting status (χ² = 956.884, df = 594, χ²/df = 1.611, CFI = 0.884, TLI = 0.883, RMSEA = 0.054, SRMR = 0.092). There was a prominent decrease in CFI (ΔCFI = -0.046) and a notable rise in SRMR (ΔSRMR = 0.027), though ΔRMSEA = 0.010 was still acceptable. The chi-square difference test was highly significant (Δχ² = 200.790, Δdf = 56, P < 0.001). Table 7Measurement invariance across genderTest TypeCMINDFCFITLIRMSEAΔCMINΔDFΔCFIΔRMSEAΔSRMR P Configural invariance756.094538.0000.9300.9220.044——————Measurement Weight Invariance783.525559.0000.9280.9230.04427.70021.000-0.0020.0000.0030.149Structural Covariance Invariance804.660569.0000.9240.9200.04548.83431.000-0.0060.0010.0290.022Measurement Residual Invariance956.864594.0000.8840.8830.054201.03956.000-0.0460.0100.0270.000*CMIN, Chi-square value; DF, Degrees of freedom;ΔCMIN, Chi-square difference;ΔDF, Degrees of freedom difference. CFI = comparative fit index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual. Δvalues were calculated relative to the configural invariance model. Because chi-square difference tests may be sensitive to sample size and model complexity, ΔCFI, ΔRMSEA, and ΔSRMR were also used to evaluate the practical deterioration in model fit after imposing equality constraints Table 7Measurement invariance across genderTest TypeCMINDFCFITLIRMSEAΔCMINΔDFΔCFIΔRMSEAΔSRMR P Configural invariance756.094538.0000.9300.9220.044——————Measurement Weight Invariance783.525559.0000.9280.9230.04427.70021.000-0.0020.0000.0030.149Structural Covariance Invariance804.660569.0000.9240.9200.04548.83431.000-0.0060.0010.0290.022Measurement Residual Invariance956.864594.0000.8840.8830.054201.03956.000-0.0460.0100.0270.000*CMIN, Chi-square value; DF, Degrees of freedom;ΔCMIN, Chi-square difference;ΔDF, Degrees of freedom difference. CFI = comparative fit index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual. Δvalues were calculated relative to the configural invariance model. Because chi-square difference tests may be sensitive to sample size and model complexity, ΔCFI, ΔRMSEA, and ΔSRMR were also used to evaluate the practical deterioration in model fit after imposing equality constraints Measurement invariance across gender *CMIN, Chi-square value; DF, Degrees of freedom;ΔCMIN, Chi-square difference;ΔDF, Degrees of freedom difference. CFI = comparative fit index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual. Δvalues were calculated relative to the configural invariance model. Because chi-square difference tests may be sensitive to sample size and model complexity, ΔCFI, ΔRMSEA, and ΔSRMR were also used to evaluate the practical deterioration in model fit after imposing equality constraints The BDS-O scale developed in this study (consisting of four domains with a total of 25 items) exhibits good psychometric properties. The complete scale is presented in Supplementary Table 1. It may be used in clinical settings for comprehensive obesity care or for conducting obesity research. Discussion This research, based on the PRO framework, developed and validated the BDS-O scale for the assessment of body-related distress in Chinese adult individuals with obesity. It consists of 25 question items across four domains, namely: body image and self-perception, social comparison and perceived evaluation by others, emotional responses and psychological burden, and weight management behaviors and sense of control. This factorial structure was defined through exploratory factorial analysis and confirmatory factorial analysis, and demonstrated good psychometric properties [48, 49]. The ‘Body Image and Self-Perception’ domain reflects the internalization of negative body image and self-blame among individuals with obesity. Items such as “Even if people say I am normal-looking, I think I am too fat” and “I think I am to blame for gaining too much weight” indicate that respondents often attribute obesity to personal factors, which is consistent with existing literature on the internalization of weight stigmatization among individuals with obesity [50]. The item “I think I am less attractive than others because I am overweight” also reflects the association between body weight and perceived attractiveness—a common concern among individuals with obesity [51, 52]. The ‘Social Comparison and Perceived Evaluations by Others’ domain is a core component of the scale that extends beyond conventional social functioning measures in health-related quality-of-life assessments [53, 54]. It includes items reflecting universal concerns such as “I worry that others notice that I have gained weight” and “I am embarrassed if people call me fat”, as well as perceptions of weight evaluations from significant others [55], such as “My parents think I am overweight” and “People who matter to me think I am fat.” These items were derived from patient interviews and reflect the impact of interpersonal weight evaluations on individuals with obesity [56, 57]. Additionally, the item “I avoid sharing pictures of myself on social media because I am worried people might say something negative about my weight” reflects the impact of social media on body image concerns in contemporary society [58], which may exacerbate social withdrawal tendencies among individuals with obesity [59]. The ‘Emotional Responses and Psychological Burden’ section reflects the psychological impacts of obesity [60]. Examples of question items in this section include: “I avoid social activities because of my weight” and “I feel lonely because of obesity.” This cycle of social avoidance due to stigmatization tends to perpetuate itself [50, 61]. Such concerns related to negative social evaluations cause individuals with obesity to socially isolate themselves to avoid the disapproval of others. However, social isolation further undermines the person’s social support network. Statements such as “Taking my weight measurement discourages me” and “I get upset because I am unable to achieve my ideal body shape” convey the frustration that manifests due to the discrepancy between reality and sociocultural expectations of ideal body types [62, 63]. The ‘Weight Management Behaviors and Sense of Control’ dimension is highly significant despite having fewer question items. Statements such as “I tried to lose weight but quit during the process” and “I do not have the self-control to maintain a healthy body weight” indicate marked self-doubts in terms of a person’s ability to manage their weight. The item “If the opportunity arose, I might seriously think about undergoing weight loss surgery” indicates marked dependence on external approaches rather than a strong focus on self-control. These statements point towards the paradox that, although self-discipline in weight control is advocated societally, facilities for receiving proper assistance in terms of psychological support and healthy eating options are woefully lacking [32, 64]. This research uses the PRO framework to provide a tool for the broader measurement of the psychosocial burden of obesity, beyond merely measuring the direct impact of its symptoms. Standard measurement methods in obesity research emphasize either the ‘level of distress’ or the ‘degree of functional limitation’ and the monitoring of individuals’ symptoms. The scale was developed to directly represent individuals’ own interpretations of their health status based on self-reporting. The scale captures the multidimensionality of the issue by allowing individuals to independently report negative reflections concerning their body image. This study verified configural and metric invariance of the BDS-O scale across genders, while structural covariance and measurement residual invariance were not fully established. Accordingly, the scale is applicable for gender comparisons at the factor structure and factor loading levels. Findings related to latent factor covariances and item measurement residuals across genders should be interpreted with caution to prevent overinterpretation. Clinical implications The BDS-O is designed as a screening and needs, to assessment tool for use in clinical settings, not as a diagnostic or judgmental instrument. When an individual with obesity completes the BDS-O, the total score and subscale patterns can help clinic staff identify specific psychosocial distress. The scale should be introduced to patients as a way to understand their feelings and struggles, not to evaluate their ‘willpower’ or ‘responsibility’ for their weight. Scores should be discussed in a supportive, non‑judgmental manner, using language that reduces shame (e.g., ‘Many people with obesity experience these feelings, and let’s see how we can help’ [65, 66]).The results should guide shared decision (making about psychosocial support, e.g., referral to counseling, body image interventions, stigma coping strategies, or group support [4, 67]) rather than simply intensifying weight loss prescriptions.Clinic staff should avoid using the scale to compare patients or to set weight‑loss targets based on scores. Its primary value lies in opening a dialogue about the patient’s internal experience and unmet psychological needs.” The scale should be introduced to patients as a way to understand their feelings and struggles, not to evaluate their ‘willpower’ or ‘responsibility’ for their weight. The scale should be introduced to patients as a way to understand their feelings and struggles, not to evaluate their ‘willpower’ or ‘responsibility’ for their weight. Scores should be discussed in a supportive, non‑judgmental manner, using language that reduces shame (e.g., ‘Many people with obesity experience these feelings, and let’s see how we can help’ [65, 66]). Scores should be discussed in a supportive, non‑judgmental manner, using language that reduces shame (e.g., ‘Many people with obesity experience these feelings, and let’s see how we can help’ [65, 66]). The results should guide shared decision (making about psychosocial support, e.g., referral to counseling, body image interventions, stigma coping strategies, or group support [4, 67]) rather than simply intensifying weight loss prescriptions. The results should guide shared decision (making about psychosocial support, e.g., referral to counseling, body image interventions, stigma coping strategies, or group support [4, 67]) rather than simply intensifying weight loss prescriptions. Clinic staff should avoid using the scale to compare patients or to set weight‑loss targets based on scores. Its primary value lies in opening a dialogue about the patient’s internal experience and unmet psychological needs.” Clinic staff should avoid using the scale to compare patients or to set weight‑loss targets based on scores. Its primary value lies in opening a dialogue about the patient’s internal experience and unmet psychological needs.” Strengths and limitations One major strength of this research is that the item generation process was based on the feedback of real individuals. However, this study also has limitations that should be acknowledged. The sample group was derived from a single hospital and was primarily composed of urban residents and highly educated participants. This demographic bias may limit the generalizability of our findings to rural populations or individuals with different educational backgrounds who are also affected by obesity. Therefore, caution should be exercised when extrapolating the conclusions of this study. Although we used a sum score of the four domains to represent the overall symptom experience, this aggregation may reduce the specificity of each individual dimension and limit their separate interpretation. Additionally, we did not collect external measures, which precluded examination of convergent and criterion-related validity. Future research should aim to expand the geographical coverage of sample sources, include participants with more diverse backgrounds, and incorporate such external measures to further validate the BDS-O. In addition, test–retest reliability was not assessed in the present study. Although body-related distress may fluctuate over time in response to changes in weight, treatment experiences, social interactions, and psychological status, temporal stability remains an important source of reliability evidence for a newly developed patient-reported outcome measure. Future studies should evaluate the test–retest reliability of the BDS-O under appropriate conditions, such as among clinically stable participants over a short interval without major weight-management interventions or substantial changes in health status. Conclusion In conclusion, this study developed and validated a patient-reported outcome measure for assessing body-related distress in Chinese adults with obesity, using a patient-centered item generation process and rigorous psychometric validation methods. The 25-item, four-factor scale captures the multidimensional psychological processes related to body weight/size concerns among individuals with obesity. These processes include body image and self-perception, social comparison and perceived evaluations by others, emotional responses and psychological burden, and weight management behaviors and a sense of control. This scale provides a reliable and valid assessment tool for clinical practice and research in obesity care and represents a step forward in patient-reported outcome measurement for the psychosocial burden of obesity. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1. Supplementary Material 1.