The impact of body mass index (BMI) on satisfaction with work life: An international BODY‐Q study

Summary Obesity is a global health issue known to have a major influence on health‐related quality of life (HR‐QOL). HR‐QOL is a concept evaluating physical and psychological health. Work life can impact HR‐QOL in people with obesity. The aim of this study was to measure the association between body mass index (BMI) and satisfaction with work life. This study included participants from an international multicenter field‐test study of BODY‐Q scales. Recruitment took place at hospitals in Denmark, The Netherlands and USA between June 2019 and January 2020. The BODY‐Q Work Life scale was used to measure work life satisfaction. The difference between BMI groups and work life satisfaction was examined using one‐way analysis of variance. Multivariable linear regression analysis was used to examine the association between BMI and work life satisfaction, adjusted for significant confounders. Of 4123 participants, 2515 completed the BODY‐Q Work Life scale. BMI groups showed significant difference in work life satisfaction (p < .0001). The Work Life scale mean score was 77.6 for the normal BMI group, 78.5 for the overweight group and 75.0, 68.9 and 63.8 for Class 1, 2 and 3 obesity, respectively. Furthermore, BMI was significantly associated with satisfaction with work life (adjusted regression coefficient −.962, p < .0001). Higher BMI was associated with lower work life satisfaction. This finding suggests that a reduction in BMI may have a positive influence on work life satisfaction in people with obesity.


What this study adds
• Increasing BMI is shown to have a direct negative correlation to work-related quality of life.
• Work-related quality of life is shown to be better in people with obesity after they have had BS.
• This research suggests that BS could be beneficial for work-related quality of life in people with obesity.

| BACKGROUND
Obesity is an important and increasing global health issue. At least 2.8 million people die from obesity-related medical conditions every year. 1,2 The correlation between obesity and chronic diseases has been well-documented. 3,4 In addition, obesity is known to have major impact on people's psychological well-being. [5][6][7] A review examining how health-related quality of life (HR-QOL) is related to body mass index (BMI) has shown that higher BMI is associated with diminished HR-QOL. 8 An important aspect of HR-QOL of people with obesity is work-related problems, especially job security. 9 Obesity can result in weight-related discrimination and has been shown to be associated with increased absenteeism and reduced work performance, which may impose an economic liability to the workplace. 10,11 Furthermore, it has been suggested that obesity causes higher unemployment rates and lower job satisfaction. [12][13][14] Bariatric surgery (BS) is known to be the most effective treatment for obesity, resulting in weight loss and improved HR-QOL. [15][16][17][18] Thus, BS may be a key factor to alleviate some of the challenges that negatively influence people with obesity in their work life. Furthermore, it is known that BS is associated with a reduction in employment absence and an improvement in work productivity. 19 Patients also reported significant improvements in mental and physical function related to work after BS and it is suggested that BS may improve the employment rate in people with severe obesity. [19][20][21] Change and improvement related to specific treatments is best assessed by condition-specific well developed and psychometrically sound patient-reported outcome measures (PROMs). 22 PROMs are questionnaires that allow patients to report their own health condition and HR-QOL without a clinician's interpretation. 23 Systematic reviews and meta-analysis of randomised trials examining the relationship between obesity, weight loss, and HR-QOL described that a limitation of research is the inconsistent use of generic and diseasespecific PROMs. 24,25 In a review that addressed the quality of PROMs for BS and body contouring surgery (BC), de Vries et al. 26 found that the BODY-Q demonstrated the strongest validation evidence for use in BS and BC patients.
The BODY-Q is a condition-specific, validated PROM that measures HR-QOL, appearance and experience of health care for use in obesity, BS and BC. The BODY-Q was developed according to guidelines for adequate development of PROMs. 27 The development process included a literature review, 63 patient interviews, 22 cognitive patient interviews and input from 9 experts. 28 An international sample of 734 participants completed the BODY-Q, and Rasch measurement theory analysis was used to examine its psychometric properties. 29,30 Recently, five new BODY-Q scales were developed and field-tested in an international sample of 4004 participants. 31 These scales were designed to measure eating-related concerns for people seeking treatment for weight loss and included a scale measuring weight-related work life issues. Each BODY-Q scale is independently functioning, which enables the user to select the scales that suit their research or clinical purpose. The new scales were developed following the same rigorous guidelines as the original scales, and evidence of their validity and reliability is reported elsewhere. 31 The aim of this study was to measure the association between BMI and satisfaction with work life using the BODY-Q Work Life scale.

| METHODS
This study used the BODY-Q Work Life scale field-test study dataset.
The BODY-Q Work Life scale will be referred to in the rest of the arti- The sample also included participants in Canada and USA recruited from the online research platform Prolific Academic.
Each site, except Denmark, received approval from the respective ethics committee for the collection of data. In Denmark, as this study was based on questionnaire data, approval from The Regional Com-  and was based on the Danish BODY-Q database including patients pre-and post-BS who were currently, or had earlier been, in a treatment course at a Danish hospital. These patients have previously filled out the BODY-Q and thereby contributed data for earlier studies on bariatric patients. 32,33 The patients included in the database who consented to be contacted again were sent the new BODY-Q scales for testing. Up to three reminder emails were sent to nonrespondents.

| Outcome variables
Participants who had worked in a job with coworkers 3 months prior to completing the BODY-Q survey were eligible to complete the Work Life scale. The Work Life scale includes 10 items with a series of statements that ask about the following issues in the workplace; feeling accepted, listened to, treated equally, standing up for yourself, having equal opportunities, having confidence, eating around coworkers, confidence at social events, feeling good and comfortable about weight. Each item has four response options that measure agreement (i.e., definitely disagree to definitely agree). The raw score for this scale was converted into a Rasch transformed score (0-100), with higher scores indicating higher work life satisfaction. In addition to the Work Life scale, data for the following demographic variables were collected: age, gender, education level, part-time or full-time employment status, time since any weight loss treatment, medical or surgical weight loss treatment, type of weight loss surgery, country of recruitment and clinical or non-clinical participant group. Demographic characteristics are shown in Table 1.

| Statistical analysis
Data analysis was performed using IBM SPSS Statistics for Windows. 38 BMI was calculated from reported weight and height.
Patient characteristics were described as percentages or the mean ± standard deviation (SD). Distribution of the Work Life scale (normality) was evaluated by histograms. To examine the difference in Work Life scale outcome in relation to the demographic variables of the study population, we performed independent t-test for dichotomous variables or one-way analysis fo variance for categorical variables. A multivariable linear regression model was used to assess the association between BMI and Work Life scale outcome which was the main objective of this study. Before analysis, normality of BMI, age and time since treatment were evaluated with histograms, and continuous variables that were not normally distributed were rescored as categorical variables. We evaluated whether all assumptions for regression analysis were met for further analysis. In addition to the crude analysis, the model was adjusted for baseline variables that were significant confounders. The continuous potential confounders included BMI and age. The categorical variables that were used as dummies included education (high school, college, masters), country (USA/Canada, the Netherlands, Denmark) and time since treatment (pre-operative, first year, 1-2 years, 3 years or more). The dichotomous potential confounders included gender (male vs. female), medical or surgical weight loss treatment, type of weight loss surgery (sleeve gastrectomy vs. gastric bypass) and clinical or non-clinical participant group. Findings were considered statistically significant for a two-tailed significance level of p < .05.
Confounders were identified by a 10% or greater difference between the values of the regression β in the crude and adjusted analysis of association.
To detect effect modification of the clinical or non-clinical group, further analyses were performed with the interaction between clinical or non-clinical group and BMI (a p < .05 indicated an interaction effect). If there was a significant interaction effect, a stratified analysis was performed following the same method described above.

| RESULTS
Of the participants invited to complete the survey 64% agreed to be enrolled in the study. Overall, a total of 4123 people participated in the study. Of the enrolled participants, 2515 (61%) were working full or part-time in the past 3 months and filled out the Work Life scale.
The enrollment rate for the invited participants varied by recruitment country as follows: Denmark 59% (n = 620), The Netherlands 62% T A B L E 1 Sample characteristics and results from one-way ANOVA and t-test analysis (n = 256) and USA 73% (n = 1607). Figure 1 illustrates the study enrollment process.

| BODY-Q Work Life scale
The overall mean Work life scale score for all participants was 74.1 (0-100).  Life scale mean scores with the lowest score found in the preoperative group, whilst the highest score was found in the group recruited 1-2 years after surgery (p < 0.0001). The mean score on the Work Life scale was higher in the BS group than the medical weight loss group (p < 0.0001). By type of surgery within the BS group, patients who had gastric bypass reported the highest mean score.
There was no difference found in the Work Life scale mean score for employment status. Education showed a significant difference between groups (p < .0001) but seemed to differ arbitrarily and did not show a clear tendency associated with higher or lower education level.

| Multivariable linear regression analysis
In the multivariable linear regression model presenting the association between BMI and work life satisfaction, the crude regression coefficient was À.539, 95% confidence interval (CI) À0.677 to 0.497, p < .0001. After adjusting the multivariable linear regression model for significant confounders (age, country [dummy variables USA/Canada, Netherlands, Denmark] and pre-operative vs. post-operative), the regression coefficient was À0.736, 95% CI À0.892 to À0.580, p < .0001. Effect modification of clinical versus non-clinical participant group was detected, so we performed further stratified analysis. For the clinical group only we found the crude regression coefficient was À0.983, 95% CI À1.098 to À0.869, p < .0001 and adjusted for significant confounders (pre-operative vs. post-operative) the regression coefficient was À0.872, 95% CI À1.004 to À0.740, p < .0001. For the non-clinical group only we found the crude regression coefficient was À0.611, 95% CI À0.762 to À0.459, p < .0001 and adjusted for significant confounders (age) the regression coefficient was À0.699, 95% CI À0.0853 to À0.546, p < 0.0001.  Therefore, it is relevant to consider the potential effect bariatric treatment alone could have on body image and self-perception and that this could possibly have affected the reported BMI. However, BMI is commonly used as it is the most accessible choice when collecting data from a questionnaire survey. 49

| CONCLUSION
Overall, our findings confirm that BMI is related to work life satisfaction. Our analysis showed that a lower work life satisfaction was associated with increased BMI. This finding supports that a decrease in BMI, alongside other HR-QOL improving qualities, might have the potential to positively influence the work life of people with obesity.
This result can be useful in the clinical guidance of people seeking BS and benefit shared decision-making.