Back to Journals » International Journal of General Medicine » Volume 17

Eating Behaviors and Adiposity Indicators – Observations from the 2016/17 Examination of the Population Study of Women in Gothenburg

Authors Mehlig K , Sliwa P, Blomstrand A , Farhan L, Hällström T, Hange D 

Received 3 January 2024

Accepted for publication 23 April 2024

Published 13 May 2024 Volume 2024:17 Pages 2101—2111

DOI https://doi.org/10.2147/IJGM.S457731

Checked for plagiarism Yes

Review by Single anonymous peer review

Peer reviewer comments 3

Editor who approved publication: Professor Satish Chandrasekhar Nair



Kirsten Mehlig,1 Pawel Sliwa,1 Ann Blomstrand,1 Lena Farhan,1 Tore Hällström,2 Dominique Hange1,3

1School of Public Health and Community Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden; 2Neuropsychiatric Epidemiology Unit, Department of Psychiatry and Neurochemistry, Institute of Neuroscience and Physiology, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden; 3Research, Education, Development & Innovation, Primary Health Care, Region Västra Götaland, Gothenburg, Sweden

Correspondence: Kirsten Mehlig, School of Public Health and Community Medicine, Institute of Medicine, University of Gothenburg, Box 453, SE-405 30, Gothenburg, Sweden, Tel +46 766 186844, Email [email protected]

Purpose: To determine the cross-sectional association between eating behavior in terms of the three-factor eating questionnaire (TFEQ) and adiposity measures.
Methods: The TFEQ-R21 was administered to 573 women aged 38 and 50 who participated in the population study of women in Gothenburg 2016/17. Three domains, emotional eating (EE), uncontrolled eating (UE), and cognitive restraint (CR) were examined as outcomes as well as predictors of adiposity outcomes. Multiple linear and logistic regression models were adjusted for age, education, lifestyle factors, and dieting behavior.
Results: All TFEQ domains were positively associated with dieting. EE and UE were associated with higher consumption of sweets and CR with lower consumption of sweets. Wellbeing was negatively associated with EE and UE. In mutually adjusted models, EE and CR but not UE were positively associated with BMI, waist circumference (WC), waist-to-hip ratio, waist-to-height ratio, and skinfold. One standard deviation higher EE was associated with obesity, BMI ≥ 30 kg/m2, OR = 1.62 (1.26, 2.10), and abdominal fatness, WC > 88 cm, OR = 1.57 (1.26, 1.95). Former and current dieting were positively associated with these outcomes, too, but UE and CR were not associated in mutually adjusted models.
Conclusion: This study shows that emotional eating behavior is associated with adiposity in a population of middle-aged women, over a large range of values for body fatness, and independent of dieting behavior. The results imply that treatment of obesity should include psychological support to restrict the risk for emotional eating in response to states of negative mood.

Keywords: three-factor eating questionnaire, eating behavior, dieting, obesity, adiposity, women, population-based study

Introduction

The prevalence of overweight and obesity worldwide continues to be high.1 In Sweden, trends are still increasing2–4 and so are the projected costs for society.5 The prevalence of overweight and obesity is about 50%, with higher values for men than for women, but women have overtaken men with regard to severe obesity.2–4 Poor dietary habits and lack of physical activity lead to a positive energy balance and accumulation of excess body fat.6 Underlying reasons are societal and economic changes that are often beyond the control of a single individual, eg a growing availability of cheap energy-dense and nutrient-depleted foods, as well as industrialization and urbanization.7,8 The impact of these changes on the individual may vary according to education, socioeconomic status, and genetic predisposition to obesity,2,9,10 and may be mediated by mental health problems including depression, anxiety, eating disorders, low self-esteem, and social isolation.11 The relationship between obesity and mental health is complex and often of bidirectional nature, and may differ by sex. A recent study showed that overweight and obesity correlated with lower quality of life in women but not men, compared to normal weight.12

Understanding food intake biology and weight management is crucial for effective obesity treatment and weight loss strategies. Dietary manipulation, like calorie restriction, is central to treatment, but long-term studies show modest results,13,14 and non-adherence to interventions is common.15–17 To strengthen adherence to lifestyle intervention programs and ascertain long-term maintenance of beneficial changes, it is increasingly recognized that the intervention should include aspects of behavioral treatment leading to a healthier attitude towards eating.18–20 A small intervention study in Finnish adults with obesity showed that weight loss was larger the more the participants learned to maintain control over their food consumption.21 A study of patients who underwent bariatric surgery showed that those with good control over food consumption had better weight maintenance and quality of life than those with poor control.22 Both studies used short versions of the Three-Factor-Eating-Questionnaire (TFEQ-R18, TFEQ-R21) that measure three domains of eating behavior: emotional eating that describes overeating during periods of negative mood, uncontrolled eating that describes loss of control over food intake when hungry or exposed to external stimuli, and cognitive restraint that describes the conscious restriction of food intake to control body weight and shape.23–25 The latter is one of the most fundamental and well-researched variables in studies of human eating behavior.14 It refers to the tendency to constantly control food intake instead of following natural physiological cues such as hunger and satiety. This behavior must not be confounded with what is commonly known as dieting, when people restrict food intake to reach a negative energy balance and lose weight. Instead, restrained eaters limit their food consumption while keeping a constant energy balance.26

While most studies focused on the treatment of obesity, the relationship between eating behavior and weight status in the general adult population is less explored. The overall aim of this study was to examine whether eating behaviors assessed by the TFEQ-R21 are associated with different measures of adiposity in a population-based sample of 38- and 50-year-old women living in Gothenburg, Sweden. The rationale for choosing this segment of the population was the increasing trend of severe obesity among women2–4 and the mental burden associated with it.12 Also, a previous study of PSWG with women examined 1968–2000 showed that perceived stress was highest in the age interval 38–54,27 which may also be the age range with high risk for the development of unhealthy eating behaviors. In a first step, we sought to determine correlates of eating behaviors, eg sociodemographic and lifestyle variables. Special attention was given to dieting behaviors that are common in contemporary female populations, and to the assessment of social desirability bias regarding self-reported eating behavior. In the main analysis, we aimed to determine the mutually adjusted associations between emotional, uncontrolled, and restricted eating behavior with different measures of body fatness and determine the impact of potential confounders including dieting. Affirmative findings may inform clinicians on how to improve obesity interventions by addressing the underlying eating behavior in addition to changes in diet and physical activity.

Subjects and Methods

Data Set

The population study of women in Gothenburg (PSWG) consists of population-based samples of women in specific age strata between 38 and 60 invited and examined every 12th year between 1968 and 2016. The age span was chosen to cover the age range around the menopausal transition, which is a major research aim of PSWG from the beginning.28 Subsequent examinations added newly recruited participants aged 38 and 50 in order to allow for analyses of trends in risk factors over time, which is another research aim of PSWG.29 The present study is based on the most recent examination that took place in 2016/17 and included 573 women aged 38 (n = 279) and 50 (n = 294), with participation rates given by 63% and 73%, respectively.30 It is also the first examination including the TFEQ questionnaire to assess eating behavior.

Exposure

Three types of eating behaviors were self-assessed using a revised 21-item version of the Three-Factor Eating Questionnaire (TFEQ-R21); emotional eating (EE, 6 items) that describes overeating during periods of negative mood, uncontrolled eating (UE, 6 items) that describes loss of control over food intake when hungry or exposed to external stimuli, and cognitive restraint (CR, 9 items), that describes the conscious restriction of food intake to control body weight and shape. The TFEQ-R21 was constructed in a Swedish population31 and shown to be valid in populations of varying ethnicity, age, and weight status.24,32–34 The items were measured on a 4-point Likert scale and domain scores for EE, UE, and CR were calculated as mean scores of their respective items excluding observations with missing items. The scores were standardized to values between zero and 100, with higher values indicating more problematic behavior. Validated cut-offs for excessive eating behaviors are lacking, however, dichotomization at value 50 was proposed by Chacko et al35 and we adopted this definition. For association analyses, the three TFEQ scores were standardized to zero mean and unit standard deviation (SD) to be able to compare the magnitude of their associations with the different health outcomes.

Anthropometric Variables

The physical examination included measurements of weight and height, waist and hip circumferences, and skinfolds at three different positions. For this study, the sum of skinfolds at the biceps, triceps, and subscapular skinfold was used. Body mass index was calculated as weight/height2 (in units of kg/m2) as well as the dimensionless quantities waist-to-hip ratio (WHR) and waist-to-height ratio (WHtR). BMI was further categorized into underweight (BMI < 18.5), overweight (BMI ≥ 25) and obesity (BMI ≥ 30). Markers involving waist circumference were dichotomized at cut points for increased metabolic risk, according to the American Heart Association and the WHO, waist circumference at 88 cm,36 WHR at 0.85,37 and WHtR at 0.5.38 The sum of skinfolds was dichotomized at the highest quintile, given by 64 mm.

Potential Confounders

All regression models included age (50 vs 38 years) which is an important design variable and is known to be related to weight status. The self-reported questionnaire included a validated question about maximally attained education, dichotomized into university education vs lower education. The women’s and their partners’ socioeconomic status (SES) was derived from their occupation,39 and low family SES was defined as maximum SES dichotomized into low vs medium/high. The validated question about general wellbeing was answered on a 7-step Likert scale,40 and was used as a linear predictor, with higher values indicating better wellbeing. The question about marital status was dichotomized into “living with a partner” vs “living alone”. Women were asked whether they had been on a weight-loss diet previously and whether they were currently dieting. Because all except one woman currently dieting had been on a diet previously, we defined three mutually exclusive categories: never, previous only, and current dieting irrespective of previous dieting. Consumption of sweets was measured in four categories: never, a few sweets/day, consumption several times/day, is very fond of sweets and consumes a lot almost every day; due to small numbers, the latter two categories were combined for this analysis. The self-reported frequency of habitual alcohol consumption from beer, wine and spirits was converted into total ethanol intake in grams/day using assumptions about portion size and ethanol content as described in.41 Habitual coffee consumption was included as the number of cups/day irrespective of type of coffee or preparation method. Smoking status was defined as current use of any tobacco (cigarette, cigar, pipe, e-cigarette, and moist snuff) versus none. Habitual leisure-time physical activity (LTPA) was measured in four categories, including sedentary behavior, moderate activity ≥ four hours/week, regular training, and competitive sports. Eating behavior was assessed using a validated instrument, but it cannot be excluded that the women underreported behavior that is considered undesirable, such as emotional and uncontrolled eating.42 For this reason, we also adjusted for a control variable derived from the Eysenck Personality Inventory assessing the tendency to give socially desirable answers, commonly known as the lie scale.43,44

Ethical Approval

Ethical approval was granted by the Regional Ethics Committee Gothenburg (258–16). The study complies with the Declaration of Helsinki and written informed consent was obtained from all participants.

Statistical Methods

Between group comparisons were performed using the Mann–Whitney U-test (continuous variables), χ2-test (binary variables), and Cochran-Armitage test of trend (categorical variables). Pearson correlation coefficients were calculated for EE, UE, and CR. Multiple linear regression was used to examine the associations between TFEQ scores and sociodemographic and lifestyle variables as well as age, and the Eysenck lie scale. TFEQ scores per se were used as outcome variables because residuals were normally distributed despite some positive skewness of the distributions, 0.8 (EE), 0.7 (UE), and 0.1 (CR). Logistic regression with stepwise forward selection of variables was applied to identify predictors of excessive eating behaviors. Furthermore, we assessed the mutually adjusted associations between TFEQ scores and indicators of adiposity. All anthropometric variables were positively skewed and logarithmized values were used in linear regression models to ensure normally distributed residuals. As a result, the exponentiated value of the beta coefficient was given that describes the relative change in outcome per one-unit change in predictor according to (exp(beta) – 1) × 100%. Regression models were adjusted for age, university education, living with a partner, wellbeing, coffee consumption, ethanol consumption, current tobacco use, LTPA, and the Eysenck lie scale. University education was chosen over family SES because of fewer missing values. Because of the large conceptual overlap between dieting behavior as well as sweet consumption with eating behaviors, association results were reported for models with and without further adjustment for these variables. Dichotomized markers for adiposity were related to TFEQ scores via logistic regression. While all models included EE and CR, we applied stepwise forward selection of covariates to account for the restricted number of event outcomes. Model check included the assessment of correlations between beta-values of predicting variables, with variance inflation factors <2.0 for all models presented here. Model fit was assessed in terms of adjusted coefficient of determination (Radj2) for linear regression models and area under the ROC curve (AUROC) for logistic regression models. The statistical significance level was set at 0.05 (2-sided tests). Analyses were performed using SAS v.9.4 and MATLAB R2022a.

Results

Definition of the Analytic Sample

Primary exclusion criteria were missing answers to any item of the TFEQ (n = 27). After the exclusion of five women, who were pregnant at the time of the examination, the final sample size was given by 541 women.

Psychometric Properties of TFEQ-R21

The three domains of TFEQ-R21 showed very good internal validity, Cronbach’s alpha = 0.93 (EE), 0.88 (UE) and 0.84 (CR). Pairwise correlations between eating scores were positive, with larger correlation between EE and UE, Pearson correlation coefficient r = 0.65, and somewhat smaller values for EE and CR, r = 0.22, and UE and CR, r = 0.18 (p < 0.0001 for all three estimates). Some domains reached the minimum value of zero (EE, UE, and CR) or the maximum of 100 points (EE), but the skewness of the distributions was small (<0.8). Defining excessive eating behaviors as values >50 points, we found that 87 women presented with excessive EE behavior (16%), 83 women with excessive UE behavior (15%), and 140 women with excessive CR eating behavior (26%), cf. Table 1.

Table 1 Basic Characteristics of Women Participating in the Population Study of Women in Gothenburg 2016–17

Basic Characteristics of the Study Sample (Table 1)

About half of the women were 38 or 50 years old. Most of them had university education (68%) and only a few had economic difficulties (14%). More than two-thirds of the women had been dieting previously, and among those currently dieting, all except one had dieted previously. Eleven women were underweight (2%), while 201 had overweight with a BMI ≥ 25 (37%) and 73 women were obese (14%).

Correlates of Eating Behavior

Table 2 shows the results from multiple linear regression of TFEQ scores on potential confounders of their association with adiposity measures. Consumption of sweets was most strongly associated with all three factors, in a positive way for EE and UE, and negatively for CR. The associations showed a dose–response relationship, ie significant differences in beta values for low-intermediate and intermediate-high consumption of sweets. Dieting behavior was positively associated with UE, EE, and CR, in order of increasing effect size. Current dieting showed stronger associations with TFEQ scores than former dieting. Living with a partner and better wellbeing were associated with lower EE, the latter also with lower UE. Higher CR scores were associated with more physical activity and less smoking. Coffee consumption was positively associated with CR, but alcohol consumption was not associated with eating behavior. A higher propensity to give socially desirable answers was negatively associated with EE and UE, and positively with CR. The small number of cases of excessive eating behavior limited the number of beta-values estimable in logistic regression, but stepwise variable selection confirmed the results for predictors most strongly associated with the continuous eating scores (Table S1). Sensitivity analyses showed that dieting behavior and CR did not share risk factors except for coffee consumption (Table S2). In contrast to women with higher level of CR, there were no differences in sweet consumption or LTPA among the different dieting categories. While former and especially current dieters reported lesser wellbeing, there was no association between wellbeing and CR (Table 2).

Table 2 Associations Between Sociodemographic and Lifestyle Variables as Well as the Eysenck Lie Scale with Individual TFEQ Scores (N = 506) a

Associations with Anthropometric Markers

Table 3 shows the results for associations between TFEQ items and adiposity indicators adjusted for potential confounders as described in Table 2. Higher scores of EE and CR but not UE were associated with higher anthropometric values in mutually adjusted models. One SD higher EE was associated with 4.2% higher BMI, and 3.4% higher BMI for one SD higher CR. Further adjustment for dieting attenuated the associations between TFEQ scores and BMI to 2.7% per SD for emotional eating, and 1.8% per SD for cognitive restraint. Current and former dieting per se were associated with 17.4% and 8.9% higher BMI, respectively (Table S3). Comparing effect sizes, the strongest associations were observed for BMI and total skinfold, and the weakest for WHR. Interaction analyses showed that the associations between TFEQ domains and BMI were generally stronger among younger vs older women, among never-dieters compared to former or current dieters, and among women not currently smoking (compared to current smokers, Table S4). The interaction terms were not statistically significant except for age class, where associations with EE and UE were observed in younger women only (p-values for interaction = 0.04 and 0.02, respectively). The association pattern between EE and continuous anthropometric measures was confirmed with respect to dichotomized outcomes (Figure 1). One SD higher EE increased the risk for all markers of adiposity by 30–60%, except for skinfold. CR was not associated with dichotomized outcomes in regression models that also included dieting behavior, except for a weak positive association with WHR ≥ 0.85 (Table S5). Higher values for EE and CR were associated with lower odds for underweight but not significantly so (not shown). Associations between TFEQ and anthropometry were not confounded by the Eysenck lie scale that was not associated with any anthropometric measure (Tables S3 and S5).

Table 3 Cross-Sectional Associations Between TFEQ Scores and Anthropometric Measuresa

Figure 1 Associations between EE, CR and dichotomized adiposity markers, OR (95% CI). Stepwise variable selection from covariates listed in Table 2, forcing the variables for eating behavior into the regression model (Table S5). aHighest quintile vs lower values of the sum of skinfolds (biceps, triceps, and subscapular skinfold).

Discussion

This population-based study of middle-aged women showed that emotional eating (EE) and cognitive restraint (CR) were positively associated with continuous measures of adiposity. EE rather than CR was associated with the dichotomized outcome measures obesity and abdominal fatness based on waist circumference. Uncontrolled eating was not associated with adiposity measures in models including EE and CR. The associations between eating behaviors and adiposity indicators were independent of lifestyle and social factors measured in this study. The associations were also independent of dieting behavior, which was common in this sample of women, and per se positively correlated with TFEQ domains and with adiposity indicators. Women with a higher propensity to give socially desirable answers in the Eysenck lie scale scored lower on EE and UE and higher on CR, but this bias did not affect the associations between TFEQ scores and objectively measured weight status.

A Finnish study of almost 3000 women aged 17–20 years reported positive associations between BMI and both EE and CR, but not UE.45 Our study extends these results to other adiposity indicators, and to a population of middle-aged women. The positive association between EE and adiposity is plausible given that EE correlates with sweet consumption, which suggests a positive energy balance and risk for weight gain. Uncontrolled eating was not associated with adiposity in regression models adjusting for emotional eating, indicating that EE has a larger impact on adiposity than UE, at least in women. Previous studies showing associations between UE and BMI used older versions of TFEQ that did not explicitly assess emotional eating.46 Based on our results we hypothesize that the association between UE and BMI might have been reduced if it had been adjusted for EE. Another Swedish population study showed that recent negative life events were associated with weight gain over 13 years suggesting that overeating may be a common way to cope with adversities.47 The importance of EE relative to the other components of TFEQ is also demonstrated in the present study by the finding that only EE was consistently associated with dichotomized adiposity measures. The largest associations were observed for EE in relation to the more severe outcomes obesity and abdominal adiposity defined as WC > 88 cm.

The positive association between CR and obesity may be unexpected because conscious food restriction, as suggested by lesser sweet consumption, may be expected to correlate negatively with obesity. Indeed, a previous study among 3000 Czech adults found that BMI and waist circumference were negatively related to restraint, but positively to disinhibition (a factor from the original 51-item TFEQ that corresponds to uncontrolled eating in TFEQ-R21).48 Lowe argues that the eating behavior exhibited by restrained eaters stems from frequent dieting and overeating in the past rather than from their current state of dietary or cognitive restraint.49 In this sense, we could interpret the positive association between CR and adiposity as a sign of a vicious circle with dieting, overeating, and eating restraint alternating and reinforcing each other, which is supported by positive correlations observed between TFEQ scores, dieting behavior, and adiposity measures.

About 70% of the women had a history of dieting behavior. Both former and current dieting were positively associated with all adiposity measures, and the association with former dieting may show the futility of previous dieting attempts. Associations of dieting and higher values for EE and UE with lesser wellbeing suggest that weight dissatisfaction is a widespread mental health problem for women in these ages. The strong association between emotional eating and adiposity suggests that weight-loss interventions should include psychological support on how to cope with negative mood states, and how to avoid overeating in these situations. For example, cognitive behavioral therapy (CBT) targeting eating behavior could be an effective approach to treat obesity within primary care settings. The method utilizes a short-term cognitive group treatment with elements from CBT and psycho-education in a group therapy format.18,20 Surgical or drug treatment may be a good solution for some patients, but even then psychological support is highly needed to avoid getting back to old eating habits and regaining weight.50

Because eating behaviors may be misreported by some individuals, we included a validated score for the propensity to give socially desirable answers. This so-called lie scale by Eysenck & Eysenck43,44 was negatively associated with overeating (EE, UE) and positively with restrained eating (CR), indicating under- and over-reporting of eating habits deemed undesirable or desirable, respectively. However, inclusion of the lie scale in the regression models for anthropometric variables did not confound the associations with TFEQ scores, because the lie scale was not associated with these objectively measured outcome measures. This confirms the robustness of the association results despite manifest reporting bias for self-reported eating behavior.

Fat distribution is strongly linked to metabolic risk, with differential risk for various fat depots.51 While BMI is a marker for overall body fat, waist circumference alone or in relation to hip circumference or height has been found to pose more health risks than fat elsewhere.52 Skinfold is both a marker for the less risky subcutaneous fat but correlates with total body fat as well.53 In the present study, eating behaviors are consistently associated with all five measures of adiposity. Larger effect sizes were obtained for BMI and skinfold compared to measures based on waist circumference, which may be partly due to lower measurement accuracy for waist and hip circumference. It shows that the most common measures of weight and height may be sufficient to monitor the risk of unhealthy eating behavior in women.

Strengths and Limitations

The examination of different adiposity measures in a contemporary population-based sample of middle-aged women is a strength of this study. Accounting for dieting behavior and social desirability bias is a further strength. The fact that data were collected eight years ago may be a limitation as living conditions may have changed since then. However, the prevalence of obesity is still increasing among Swedish women3 while eating habits deteriorated54 and socioeconomic imparities increased,55 suggesting that our results and conclusions are still relevant. A related limitation is the restriction to cross-sectional data. Prospective data are needed to establish time order and causality among the different risk factors, as a basis for improvement of current lifestyle interventions for obesity.

Conclusions

In this study of middle-aged women, we showed that emotional eating and, to a lesser extent, cognitive restraint correlated positively with different adiposity indicators, while no associations were observed for uncontrolled eating, in mutually adjusted models. The consistency of associations with adiposity measures beyond BMI as well as their independence of dieting behavior and social desirability bias are important novel aspects of this study. The negative associations between the propensity to give socially desirable answers with overeating and the positive association with eating restraint are interesting on own account and support the validity of both the TFEQ and the Eysenck lie scale.

Data Sharing Statement

The data sets used and analysed during the current study are available from the corresponding author upon reasonable request.

Funding

This work was supported by the Swedish State under the agreement between the Swedish Government and the county council, the ALF agreement (GBG-68771) as well as by The Health & Medical Care Committee of the Region Västra Götaland (VGFOUREG-586771).

Disclosure

The authors report no conflicts of interest in this work.

References

1. Ng M, Fleming T, Robinson M, et al. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the global burden of disease study 2013. Lancet. 2014;384(9945):766–781. doi:10.1016/S0140-6736(14)60460-8

2. Hemmingsson E, Ekblom O, Kallings LV, et al. Prevalence and time trends of overweight, obesity and severe obesity in 447,925 Swedish adults, 1995–2017. Scand J Public Health. 2021;49(4):377–383. doi:10.1177/1403494820914802

3. Statistics on overweight and obesity in adults. Available from: https://www.folkhalsomyndigheten.se/livsvillkor-levnadsvanor/mat-fysisk-aktivitet-overvikt-och-fetma/overvikt-och-fetma/statistik-om-overvikt-och-fetma/overvikt-och-fetma-hos-vuxna/. Accessed April 10, 2024.

4. Overweight and obesity. Available from: https://www.folkhalsomyndigheten.se/the-public-health-agency-of-sweden/living-conditions-and-lifestyle/obesity/. Accessed April 10, 2024.

5. Andersson E, Eliasson B, Steen Carlsson K. Current and future costs of obesity in Sweden. Health Policy. 2022;126(6):558–564. doi:10.1016/j.healthpol.2022.03.010

6. Hill JO, Wyatt HR, Peters JC. Energy balance and obesity. Circulation. 2012;126(1):126–132. doi:10.1161/CIRCULATIONAHA.111.087213

7. Hruby A, Hu FB. The Epidemiology of Obesity: a Big Picture. Pharmacoeconomics. 2015;33(7):673–689. doi:10.1007/s40273-014-0243-x

8. Bricas N. Urbanization issues affecting food system sustainability. In: Brand C, Bricas N, Conaré D, editors. Designing Urban Food Policies. Cham, Switzerland: Springer; 2019.

9. Devaux M, Sassi F. Social inequalities in obesity and overweight in 11 OECD countries. Eur J Public Health. 2013;23(3):464–469. doi:10.1093/eurpub/ckr058

10. Lissner L, Wijnhoven TM, Mehlig K, et al. Socioeconomic inequalities in childhood overweight: heterogeneity across five countries in the WHO European childhood obesity surveillance initiative (COSI-2008). Int J Obes Lond. 2016;40(5):796–802. doi:10.1038/ijo.2016.12

11. Sarwer DB, Polonsky HM. The psychosocial burden of obesity. Endocrinol Metab Clin North Am. 2016;45(3):677–688. doi:10.1016/j.ecl.2016.04.016

12. Zhang J, Xu LZ, Li JJ, et al. Gender differences in the association between body mass index and health-related quality of life among adults: a cross-sectional study in Shandong, China. BMC Public Health. 2019;19 doi:10.1186/s12889-019-7351-7

13. Adeola OL, Agudosi GM, Akueme NT, et al. The Effectiveness of Nutritional Strategies in the Treatment and Management of Obesity: a Systematic Review. Cureus. 2023;15(9):e45518. doi:10.7759/cureus.45518

14. Johnson F, Pratt M, Wardle J. Dietary restraint and self-regulation in eating behavior. Int J Obes Lond. 2012;36(5):665–674. doi:10.1038/ijo.2011.156

15. Wadden TA, Webb VL, Moran CH, Bailer BA. Lifestyle modification for obesity: new developments in diet, physical activity, and behavior therapy. Circulation. 2012;125(9):1157–1170. doi:10.1161/CIRCULATIONAHA.111.039453

16. Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB. Changes in diet and lifestyle and long-term weight gain in women and men. N Engl J Med. 2011;364(25):2392–2404. doi:10.1056/NEJMoa1014296

17. Galani C, Schneider H. Prevention and treatment of obesity with lifestyle interventions: review and meta-analysis. Int J Public Health. 2007;52(6):348–359. doi:10.1007/s00038-007-7015-8

18. Stahre L, Tarnell B, Hakanson CE, Hallstrom T. A randomized controlled trial of two weight-reducing short-term group treatment programs for obesity with an 18-month follow-up. Int J Behav Med. 2007;14(1):48–55. doi:10.1007/BF02999227

19. Burgess E, Hassmen P, Welvaert M, Pumpa KL. Behavioural treatment strategies improve adherence to lifestyle intervention programmes in adults with obesity: a systematic review and meta-analysis. Clin Obes. 2017;7(2):105–114. doi:10.1111/cob.12180

20. Stahre L, Blomstrand A, Hällström T. Effectivness of cognitive behaviour therapy targeting eating behaviour for patients with abdominal obesity in an ordinary primary health care setting. In: Gomaa RS, editor. Novel Research Aspects in Medicine and Medical Science. B P International; 2023.

21. Nurkkala M, Kaikkonen K, Vanhala ML, Karhunen L, Keranen AM, Korpelainen R. Lifestyle intervention has a beneficial effect on eating behavior and long-term weight loss in obese adults. Eat Behav. 2015;18:179–185. doi:10.1016/j.eatbeh.2015.05.009

22. Engstrom M, Forsberg A, Sovik TT, Olbers T, Lonroth H, Karlsson J. Perception of control over eating after bariatric surgery for super-obesity--a 2-year follow-up study. Obes Surg. 2015;25(6):1086–1093. doi:10.1007/s11695-015-1652-4

23. Karlsson J, Persson LO, Sjostrom L, Sullivan M. Psychometric properties and factor structure of the three-factor eating questionnaire (TFEQ) in obese men and women. results from the Swedish Obese Subjects (SOS) study. Int J Obes Relat Metab Disord. 2000;24(12):1715–1725. doi:10.1038/sj.ijo.0801442

24. de Lauzon B, Romon M, Deschamps V, et al. The three-factor eating questionnaire-R18 is able to distinguish among different eating patterns in a general population. J Nutr. 2004;134(9):2372–2380. doi:10.1093/jn/134.9.2372

25. Cappelleri JC, Bushmakin AG, Gerber RA, et al. Psychometric analysis of the three-factor eating questionnaire-R21: results from a large diverse sample of obese and non-obese participants. Int J Obes Lond. 2009;33(6):611–620. doi:10.1038/ijo.2009.74

26. Lowe MR, Timko CA. What a difference a diet makes: towards an understanding of differences between restrained dieters and restrained nondieters. Eat Behav. 2004;5(3):199–208. doi:10.1016/j.eatbeh.2004.01.006

27. Hange D, Mehlig K, Lissner L, et al. Perceived mental stress in women associated with psychosomatic symptoms, but not mortality: observations from the population study of women in Gothenburg, Sweden. Int J Gen Med. 2013;6:307–315. doi:10.2147/IJGM.S42201

28. Bengtsson C, Blohme G, Hallberg L, et al. The study of women in Gothenburg 1968-1969--a population study. general design, purpose and sampling results. Acta Med Scand. 1973;193(4):311–318. doi:10.1111/j.0954-6820.1973.tb10583.x

29. Bjorkelund C, Andersson-Hange D, Andersson K, et al. Secular trends in cardiovascular risk factors with a 36-year perspective: observations from 38- and 50-year-olds in the population study of women in Gothenburg. Scand J Prim Health Care. 2008;26(3):140–146. doi:10.1080/02813430802088403

30. Rodstrom K, Weman L, Sandin L, Hange D, Bjorkelund C. Is it possible to investigate menopausal age? A comparative cross-sectional study of five cohorts between 1968 and 2017 from the population study of women in Gothenburg, Sweden. Menopause. 2020;27(4):430–436. doi:10.1097/GME.0000000000001476

31. Tholin S, Rasmussen F, Tynelius P, Karlsson J. Genetic and environmental influences on eating behavior: The Swedish young male twins study. Am J Clin Nutr. 2005;81(3):564–569. doi:10.1093/ajcn/81.3.564

32. Rosnah I, Mohd Zali MN, Noor Hassim I, Azmi MT. A tale of two construct validation analysis: rasch model and exploratory factor analysis approach for three-factor eating questionnaire (TFEQ-R21) among Malaysian male workers. Med J Malaysia. 2015;70(3):169–176.

33. Martin-Garcia M, Vila-Maldonado S, Rodriguez-Gomez I, et al. The Spanish version of the three factor eating questionnaire-R21 for children and adolescents (TFEQ-R21C): Psychometric analysis and relationships with body composition and fitness variables. Physiol Behav. 2016;165:350–357. doi:10.1016/j.physbeh.2016.08.015

34. Lin YW, Lin CY, Strong C, et al. Psychological correlates of eating behavior in overweight/obese adolescents in Taiwan: Psychometric and correlation analysis of the three-factor eating questionnaire (TFEQ)-R21. Pediatr Neonatol. 2021;62(1):41–48. doi:10.1016/j.pedneo.2020.08.006

35. Chacko SA, Chiodi SN, Wee CC. Recognizing disordered eating in primary care patients with obesity. Prev Med. 2015;72:89–94. doi:10.1016/j.ypmed.2014.12.024

36. American Heart A, National Heart L, Blood I, et al. Diagnosis and management of the metabolic syndrome. An American heart association/national heart, lung, and blood institute scientific statement. executive summary. Cardiol Rev. 2005;13(6):322–327. doi:10.1097/01.crd.0000380842.14048.7e

37. Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med. 1998;15(7):539–553. doi:10.1002/(SICI)1096-9136(199807)15:7<539::AID-DIA668>3.0.CO;2-S

38. Stegenga H, Haines A, Jones K, Wilding J, Guideline Development G. Identification, assessment, and management of overweight and obesity: summary of updated NICE guidance. BMJ. 2014;349:g6608. doi:10.1136/bmj.g6608

39. Statistics Sweden. Swedish socioeconomic classification. Stockholm; 1982. Available from: https://wwwscbse/dokumentation/klassifikationer-och-standarder/socioekonomisk-indelning-sei/.Assessed. Accessed August 23, 2023.

40. Sullivan M, Karlsson J, Bengtsson C, Furunes B, Lapidus L, Lissner L. “The Göteborg quality of life instrument”--a psychometric evaluation of assessments of symptoms and well-being among women in a general population. Scand J Prim Health Care. 1993;11(4):267–275. doi:10.3109/02813439308994842

41. Mehlig K, Schult A, Bjorkelund C, Thelle D, Lissner L. Associations between alcohol and liver enzymes are modified by coffee, cigarettes, and overweight in a Swedish female population. Scand J Gastroenterol. 2022;57(3):319–324. doi:10.1080/00365521.2021.2009557

42. Kowalkowska J, Poinhos R. Eating behaviour among university students: relationships with age, socioeconomic status, physical activity, body mass index, waist-to-height ratio and social desirability. Nutrients. 2021;13(10):3622. doi:10.3390/nu13103622

43. Massey A. The Eysenck personality-inventory lie scale - lack of insight or. Irish J Psychol. 1980;4(3):172–174.

44. Eysenck HJ, Eysenck SBG. Psychoticism as a Dimension of Personality. London: Hodder and Stoughton; 1976.

45. Angle S, Engblom J, Eriksson T, et al. Three factor eating questionnaire-R18 as a measure of cognitive restraint, uncontrolled eating and emotional eating in a sample of young Finnish females. Int J Behav Nutr Phys Act. 2009;6:41. doi:10.1186/1479-5868-6-41

46. French SA, Epstein LH, Jeffery RW, Blundell JE, Wardle J. Eating behavior dimensions. Associations with energy intake and body weight. A review. Appetite. 2012;59(2):541–549. doi:10.1016/j.appet.2012.07.001

47. Mehlig K, Nehmtallah T, Rosvall M, Hunsberger M, Rosengren A, Lissner L. Negative life events predict weight gain in a 13-year follow-up of an adult Swedish population. J Psychosom Res. 2020;132:109973. doi:10.1016/j.jpsychores.2020.109973

48. Wagenknecht M, Hainer V, Kunesova M, et al. Relationships between the ”eating inventory” factors, socioeconomic status, anthropometric body adiposity indexes and health risks in Czech population. Cas Lek Cesk. 2007;146(3):284–291.

49. Lowe MR. The effects of dieting on eating behavior: a three-factor model. Psychol Bull. 1993;114(1):100–121. doi:10.1037/0033-2909.114.1.100

50. Lauti M, Kularatna M, Hill AG, MacCormick AD. Weight regain following sleeve gastrectomy-a systematic review. Obes Surg. 2016;26(6):1326–x. doi:10.1007/s11695-016-2152-x

51. Tchkonia T, Thomou T, Zhu Y, et al. Mechanisms and metabolic implications of regional differences among fat depots. Cell Metab. 2013;17(5):644–656. doi:10.1016/j.cmet.2013.03.008

52. Neeland IJ, Ross R, Despres JP, et al. Visceral and ectopic fat, atherosclerosis, and cardiometabolic disease: a position statement. Lancet Diabetes Endocrinol. 2019;7(9):715–725. doi:10.1016/S2213-8587(19)30084-1

53. Lohman TG. Skinfolds and body density and their relation to body fatness - a review. Hum Biol. 1981;53(2):181–225.

54. Statistics on adults’ eating habits. Available from: https://www.folkhalsomyndigheten.se/livsvillkor-levnadsvanor/mat-fysisk-aktivitet-overvikt-och-fetma/mat/statistik-om-mat/statistik-om-vuxnas-matvanor/. Accessed April 10, 2024.

55. Health--Europe, The Lancet Regional. Sweden’s economic inequality gap is widening and worrying. Lancet Reg Health-Eu. 2023;2023:100610 doi:10.1016/j.lanepe.2023.100610

Creative Commons License © 2024 The Author(s). This work is published and licensed by Dove Medical Press Limited. The full terms of this license are available at https://www.dovepress.com/terms.php and incorporate the Creative Commons Attribution - Non Commercial (unported, v3.0) License. By accessing the work you hereby accept the Terms. Non-commercial uses of the work are permitted without any further permission from Dove Medical Press Limited, provided the work is properly attributed. For permission for commercial use of this work, please see paragraphs 4.2 and 5 of our Terms.