Artigos

INTRODUCTION
Nowadays, obesity has risen to the top of the global public health agenda. It is one of the main metabolic risk factors for non-communicable chronic diseases (NCDs)1. These diseases have a high prevalence globally, including Argentina, and are among the leading causes of death2. In Argentina, the prevalence of overweight and obesity is 36.2% and 25.4%, respectively3, and this shows a clear and sustained increase of both through time4. In Córdoba, the prevalence of obesity was 17%, according to the Córdoba Obesity and Diet Study (CODIES I), the study that preceded the current one5.
Given the complexity and inherent diversity of obesity, identifying the main factors associated with its development and evolution in an integrated way is a challenge. Scientific evidence highlights the individual aetiologies of obesity, while numerous contextual factors -often social, economic, political or cultural6- have directly or indirectly influenced the issue. Each influences dietary habits and physical activity, shaping obesogenic behaviours, conditions and lifestyles at both societal and individual levels7.
The Social Determinants of Health (SDH) approach proposed by the WHO offers a conceptual framework for this challenging problem7. This model postulates that people's health status depends on the social determinants that shape it8. A WHO report in 2008 defined the SDH as ‘‘the conditions in which people are born, grow up, live, work and grow old, and the structural factors of those conditions, that is, the distribution of power, money and resources’’9. Thus, there are structural and intermediate SDHs that act synergistically6, 8. The intermediate SDHs are distributed according to social stratification and are linked to individual-level influences, which explain differences in exposure and vulnerability to health and disease conditions7. The main categories of intermediate SDH are material circumstances, psychosocial circumstances, behavioural and biological factors, social cohesion, and the health system itself7. Some behavioural factors recognised as intermediate SDH are nutrition, physical activity (PA), tobacco, drug, and alcohol consumption, and biological factors such as age, sex, and genetic characteristics of individuals7. Besides, scientific evidence has shown that socioeconomic determinants such as education, wealth, social status, and household income influence nutritional status10. Addressing obesity from a SDH perspective offers a more comprehensive understanding of the processes and interactions contributing to its development11. Structural and intermediate SDHs encompass many aspects of individuals’ micro- and macro-social environments, some of which may not be empirically measurable.
The WHO has provided a classification system for overweight and obesity and their comorbidities12. However, as mentioned before, the effect of certain factors on obesity is likely not uniform across the entire BMI distribution. In light of these considerations, we sought to capture these variations by using BMI as a continuous variable to explore subpopulations among people with obesity and their shaping characteristics, by using Finite Mixture Models (FMM). These models are being used increasingly in medical literature13 because of their flexibility, which allows us to identify latent variables and recognise possible configurations of the SDH related to obesity that are not visible at first glance.
To investigate this health phenomenon, it could be useful to adopt a holistic approach that considers the mechanisms through which different determinants impact the exacerbation and advancement of obesity. Thus, this study aims to ascertain the specific SDH that contributes to the BMI increment in individuals with obesity in the city of Córdoba, Argentina.
METHODS
Study participants and design
A population-based observational study with a cross-sectional design was conducted from June 2020 to March 2022 in Córdoba, the second most populated Argentinian city14. All individuals over 18 years old with at least two years of residence in Córdoba were potential candidates for our sample. The exclusion criteria involved physical or mental disabilities and pregnancy or breastfeeding at the time of the interview.
All participants provided their written informed consent. The research was approved by the relevant Ethics Committee (RePIS No. 4060) and was conducted following the tenets of the Declaration of Helsinki and the respective national laws.
Sample size
To calculate the sample size, an obesity prevalence of 35% was hypothesised and a Type I probabilistic error of 0.05 was set. Furthermore, we imposed an excess of no more than 10% of the hypothetical prevalence. Finally, 1186 eligible subjects were selected using multistage random sampling (Supplementary Material 1). First, neighbourhood quintiles were calculated according to the population percentage with at least one unsatisfied basic need15. Considering the homogeneity criterion, within each stratum (quintile), a simple random sample was used to select the neighbourhoods, and then the blocks within each neighbourhood were included in the sample. Last, in each block, the houses included in the final sample were systematically selected (one house every three houses), and then one person per house was interviewed.
Data collection
The subjects' socioeconomic, demographic, lifestyle and environmental characteristics were collected through a face-to-face interview conducted by a trained interviewer, using a semi-structured questionnaire. Subsequently, a food frequency questionnaire (FFQ) validated for the population of Córdoba16 was used, coupled with a photographic food atlas17, to determine portion sizes of foods and preparations. Also, the International Physical Activity Questionnaire (IPAQ)18 was administered to calculate the metabolic equivalent of tasks (MET)19.
For the assessment of nutritional status, body mass index [BMI=weight(kg)/height(m)2] was used12. Weight and height were measured using an Omron scale and a wall stadiometer, respectively.
Measurements
Exposure variables. The variables included were: sex (female/male) and age as continuous and categorical (<25, 25-34, 35-44, 45-54, 55-64, and 65 or over) variables; marital status (single/married/divorced/widowed); and health coverage (presence/absence) according to the presence of health insurance, prepaid system or emergency services. Another variable included was the Charlson Index20, which is an indicator of long-term mortality based on age and a wide range of comorbidities (19 in total). Higher values indicate a higher burden of comorbidities and, consequently, a higher risk of mortality. The inclusion of this index in the analysis, as one of the biological aspects of the intermediate SDH, allows us to consider multiple coexisting conditions in individuals jointly. Besides, socio-economic status (SES) was built, considering the occupation and the level of formal education attained by the "head of household". Physical activity (PA) level was categorised into low (<600 METs), moderate (600-1500 METs) and high (>1500 METs). Food quality score was estimated by the healthy eating index (HEI)21, adapted to the Argentinean Dietary Guidelines Recommendations (GAPA, by its Spanish acronym)22.
Outcome variable. The outcome of interest was BMI, treated as a continuous scale.
Statistical analysis
An exploratory analysis of the data collected was performed to describe the sample studied in terms of their socio-economic, demographic and lifestyle characteristics by absence or presence of obesity (BMI?30kg/m2); Student's t-tests and Chi-square tests were performed, as appropriate.
FMMs were fitted exclusively among individuals with obesity (n=273) to identify latent subgroups based on intermediate SDH behaviour across increasing BMI. In this analysis, BMI was modelled as a continuous variable to capture within-group variability and detect severity gradients, rather than assuming obesity as a homogeneous condition. This approach allows us to examine how different social determinants cluster and relate to increases in BMI, recognising that higher BMI values in already obese individuals reflect progressive metabolic dysregulation and increasing health risk. FMMs were applied to uncover latent hypothetical classes (unobservable empirically) within the obesity group. Different FMMs were fitted to determine the optimal number of latent classes based on latent class marginal probabilities. As FMMs allow the recognition of latent classes (subgroups) that can partly be explained by other observed variables13, we applied multiple linear regression within each latent class with an identical set of covariates. Two statistical methodologies were used to select the variables to be included in the final FMM: first, multiple linear regression models, starting with the full model (including all candidate predictor variables) and sequentially removing variables based on hypothesis tests (Supplementary Material 2); second, Lasso models comparing results from both methodologies. Furthermore, we applied a biological plausibility criterion to incorporate certain variables of interest regardless of the statistical performance. Finally, covariates included were age, sex, Charlson index, SES, HEI score, PA, health coverage, marital status, and an interaction term (sex*SES).
The FMM model was constructed considering three latent classes, according to the high probability of membership estimated in each of them (class 1, 0.38; class 2, 0.48; and class 3, 0.13, Supplementary Material 3). A fourth class would have resulted in a new subgroup with lower membership representation. In each class, a multiple linear regression was implemented, considering BMI as the response variable. Estimators are expressed as a coefficient (ß) and a 95% Confidence Interval (95% CI). The marginal probabilities of belonging to each class were estimated, and the average BMI within each was calculated. In addition, age was incorporated in quadratic terms as a covariate, allowing, a posteriori, the calculation of the age representing the maximum BMI peak in each latent class.
To evaluate the differences in the coefficients between the different latent classes of the model, the Wald test was applied. Finally, predictions were estimated for the BMI value of each subject and presented in graphs according to the variables of interest. Statistical analyses were performed using Stata v. 17.1 statistical software23.
RESULTS
The sociodemographic and lifestyle characteristics for the entire sample, according to the presence of obesity, are shown in Table 1. The sample included subjects aged 18 to 95 years, of whom 59.11% were women. Concerning BMI, the mean (SD) was 26.72 (5.47) kg/m2, and 23% (95%CI 20.6, 25.5) of the subjects had obesity, while 33.3% (95%CI 30.7, 36.1) had pre-obesity (BMI?25 and <30kg/m2). Significant differences were identified between age groups (p<0.001). The proportion of people with obesity was higher with increasing age, indicating a higher prevalence of obesity in the 55-64 age group and a decline thereafter (Table 1). Low and medium SES presented similar prevalence, around 25% (27.5% and 25.3%, respectively), while the highest level presented a lower proportion of people with obesity (13.8%) (Table 1). The prevalence of obesity was higher in PA low level (42.8%, p<0.001), widows or married (31.4% and 29.4%, respectively p<0.001), people without health coverage (27.6%, p<0.001), and in the tertile 3 of Charlson index (32.6%, p<0.001), indicating a higher risk of mortality. Besides, the average HEI score was significantly higher for those without obesity (65.8 vs. 63.6, p<0.001), indicating a better quality of diet (Table 1).

Table 1. Socio-demographic and biological characteristics of the entire sample and according to presence/absence of obesity. Córdoba. Argentina 2020-2022.

Tab.1

Figure 1 shows the 3 latent classes identified by fitting FMM. According to the marginal probability of belonging to each class, 38% of the subjects belonged to class 1, 48% to class 2 and 13% to class 3. In addition, the mean BMI for each class were 31.33 (95%IC 31.09;31.56), 35.33 (95%IC 34.67;35.99), and 40.89 kg/m2 (95%IC 40.45;41.34), respectively. Based on these results, each class was considered as a profile of increased BMI in people with obesity in grades I, II, and III, respectively, as the mean BMI of each class corresponded to the cut-off points defined by the WHO for obesity classification12. The profiles were statistically different from each other (p<0.001). In addition, the age at which peak BMI occurred was 26 years for class 1, 54 years for class 2, and 56 years for class 3 (Supplementary Material 3).
In the first profile, the interaction between being female and having a high SES (ß 2.02, 95%CI 0.23;3.82), as well as having health coverage (ß 0.54, 95%CI 0.01;1.06), have a direct effect on BMI, according to their reference categories. On the other hand, in this profile, widows showed an inverse effect (ß -0.9, 95%CI -1.71;-0.09). In profile II, according to men, women presented a direct effect with BMI (ß 1.81, 95%CI 0.18;3.43). However, females belonging to a high SES have an inverse effect (ß -4.74, 95%CI -7.82;-1.66) on their BMI concerning males of low SES. There was an inverse effect for married and divorced marital status with BMI (ß 2.06, 95%CI -3.42;-0.70 and ß 3.96, 95%CI -6.23;-1.68, respectively). In the third profile, age and Charlson index were directly associated with BMI (ß 2.29, 95%CI 1.95;2.64 and ß 1.65, 95%CI 1.22;2.09, respectively). The same effect was observed among females (ß 6.06, 95%CI 4.05;8.06), widows (ß 5.8, 95%CI 3.15;8.45), and high SES participants (ß 2.41, 95%CI 0.43;4.39). On the other hand, the determinants that had an inverse effect on BMI included HEI score (ß -0.21, 95%CI -0.26;-0.17), moderate PA levels (ß -6.64, 95%CI -7.77;-5.50), being married or divorced (ß -5.83, 95%CI -6.94;-4.72 and ß -8.33, 95%CI -9.95;-6.72, respectively), and the interaction of women with a high SES (ß -4.13, 95%CI -6.77;-1.48).
Figure 2 shows the distribution of individuals belonging to each profile according to their BMI for age (A) and diet quality score (B). Subjects were included in each class in terms of their probability of belonging to each profile (greater than 0.5). According to Figure 2A, in class 1, there is a slight increase in the BMI values with increasing age. This increase is more pronounced in class 3, which has the highest BMI. On the other hand, in class 2, we observe that the BMI values tend to decrease slightly with increasing age. Figure 2B shows that in class 3, the BMI values tend to decrease sharply as the diet quality score rises. In classes 1 and 2, these values remain stable throughout the HEI score.
DISCUSSION
The prevalence of obesity in Córdoba has increased compared to previous years5, following national and global trends24. Changes in various lifestyle factors25, combined with Argentina's economic and political context in recent years, have had a major influence on this increase. In response, both in Córdoba and nationwide, initiatives have been implemented to prevent obesity; however, their impact has not yet been sufficient, as rates continue to rise5. These programmes largely focus on individual risk factors such as diet and physical activity, and include general interventions targeting NCDs, with an emphasis on hypertension and diabetes. As Pou et al. point out26, it would therefore be essential for the decision-makers to generate comprehensive and intersectoral actions that transform environments and address obesity from a macrosocial perspective.
The main results of this study recognise the intermediate SDH associated with increasing BMI in people with obesity, identifying three profiles. In the first profile (grade I obesity), characterised by a BMI peak reached at 26 years old, health coverage was associated with a higher BMI, as well as high-SES women. This result contrasts with traditional expectations where lower SES is typically associated with higher obesity risk in women, particularly27. Rather than contradicting this broader pattern, our findings may indicate emerging heterogeneity in weight gain during early adulthood in people with obesity grade I. In particular, the higher BMI observed among young high-SES women could reflect specific behavioural or sociocultural factors affecting this subgroup - such as heightened academic or work-related stress, changing dietary or physical activity patterns, or body weight fluctuations linked to transitions typical of early adulthood28. This interpretation aligns with research showing that gender and SES interact in complex ways to shape BMI increase over the life course. In addition, previous research has shown that young women (19-30 years old) often have a stronger preference for foods high in sugar29. Therefore, consuming these energy-dense foods with a high glycaemic response could lead to a consequent increase in body weight. Also, high-SES young women may have greater access to fast and ultra-processed foods, which are often expensive in Argentina. Moreover, being a widower was associated inversely with BMI in the first profile. This finding is consistent with other research30, which highlights that the death of a spouse is a stressful life event that can lead to loss of appetite and skipping regular meals in the grieving phase31, leading to a decrease in body weight.
Some similarities in profiles 2 and 3 were observed. In both groups, the female sex was directly associated with BMI increase, in line with the literature32. Sociocultural pressures to adhere to a thin ideal substantially influence body dissatisfaction and health-related behaviours. Women living with obesity may experience heightened concern about physical appearance, and when such concerns interact with thin-ideal contemporary cultural models33, they can further exacerbate vulnerability to disordered eating behaviours34. In this way, sociocultural demands not only reinforce dissatisfaction among women with obesity but may also trigger maladaptive behavioural responses. However, when considering the interaction between SES and sex, an inverse association with BMI was found for high-SES women. To understand this difference with the observed results in the first profile, it is necessary to consider the ages at which the maximum BMI peak is reached. While the first profile peaks at younger ages, the second and third profiles peak at ages 54 and 56, respectively. Thus, in profiles 2 and 3, adult women with more favourable SES could be involved in more self-care practices than men and may be covering health and care costs to improve their quality of life. This is consistent with a Finland study, which found that high SES women aged 51-65 were more interested in healthcare35. This is mainly because SES is strongly associated with people's level of education, which underlies the strong effect of this indicator on care practices as well as food selection. Besides, when it comes to health, high SES women appear to be more engaged, involved, attentive and seemingly better informed when they have to make decisions35. In addition, another aspect to consider is that women with a more favourable economic situation are better positioned to access costly weight-loss interventions to a comprehensive treatment of obesity, such as bariatric surgery or drug treatment, particularly relevant for those with higher grades of obesity.
On the other hand, in profiles 2 and 3, marriage or cohabitation and divorce were inversely associated with BMI. However, these findings likely arise from distinct processes. Some studies identified that at the beginning of marriage, there is an increase in weight36,37. Nevertheless, evidence from cross-sectional studies indicates that married older adults often show better anthropometry, compared with non-married counterparts38. Marriage may foster emotional stability, shared routines and social support, which can reduce stress and encourage adherence to medical care or structured eating patterns39, mechanisms particularly relevant in individuals with severe obesity. In this sense, previous longitudinal research shows that decreasing social isolation could reduce obesity-related risks of all-cause mortality40. On the other hand, divorce may lead to behavioural changes motivated by identity reconstruction or re-entry into social and affective relationships. In some contexts, divorced individuals may engage more actively in weight management or physical activity as part of lifestyle renegotiation37. The coexistence of protective effects in two contrasting marital conditions suggests that social role transitions, not the status itself, may shape BMI trajectories41. This complexity highlights the need to interpret marital status as an indicator of behavioural restructuring and social support dynamics rather than solely as a demographic descriptor.
Furthermore, age, Charlson index, SES, food quality score (HEI) and physical activity were associated with BMI for the third profile. The results in this group are consistent with the scientific literature and the recognised biological model in obesity32. Both a higher diet quality score and moderate PA level were found to be inversely associated with BMI in this profile, and while these factors may be beneficial for all people with obesity regardless of grade, they appear to have a significant impact on those with higher weight. Regarding diet quality, it is important to note that the index used (HEI) was developed based on the GAPA guidelines42, which define recommended dietary patterns for our population. Consequently, the HEI reflects both the quantity and frequency of food consumption according to these nutritional standards. In particular, the guidelines emphasise increasing the intake of fruits and vegetables, foods rich in micronutrients and fibre and characterised by low energy density. Accordingly, HEI scores decrease when consumption of these foods is low, and intake of products high in simple carbohydrates and saturated fats is elevated, such as pasta, bread, refined flours, ultra-processed foods (including cold cuts, sausages, and snacks), and sugar-sweetened beverages. Thus, a low-quality diet is associated with higher total energy intake, favouring a positive energy balance. It has also been noted that, due to the high basal and total energy expenditure characteristic of individuals with higher degrees of obesity, even relatively small dietary changes combined with exercises that increase the PA level may help generate a negative energy balance and thereby promote greater weight loss43. Moreover, several studies44, 45 have shown that weight loss achieved solely through calorie restriction can lead to a decline in lean mass and bone mineral density. Hence, incorporating PA, specifically progressive resistance training, alongside aerobic training, is essential to enhance improvements in body composition and physical function in individuals with obesity46.
In the Argentine population, a study based on the National Household Expenditure Survey showed that diets of low nutritional quality -high in fats, sugars and sodium, and low in fibre and micronutrients- were associated with increased obesity47. Similarly, data from the National Risk Factors Survey 2018 indicate that about 25% of Córdoba's population perceives their diet as unhealthy or very unhealthy. Argentina has taken important steps to improve population health. A recent example is the enactment of National Law 27,642 on the Promotion of Healthy Eating. It aims to guarantee the right to health, to prevent malnutrition and reduce NCDs by promoting adequate nutrition through clear front-of-pack labelling on food and non-alcoholic beverage packaging, and by implementing restrictions in school settings and public procurement. As it has only recently been implemented (it was regulated in 2022), its impact on health could be observed in the coming years. In this sense, it is particularly concerning that the province of Córdoba has not yet adhered to this Law. The absence of full provincial implementation weakens the capacity of the law to modify obesogenic environments and reduces opportunities for prevention at the population level.
Some methodological issues need to be considered. As the sample was randomly chosen, the probability of selection bias is almost non-existent. Misclassification bias, instead, may be present in the study because people could have given socially desirable answers to the questions covered by the research. However, if present, it bears no importance, as it is a non-differential misclassification bias, and the bias introduced would likely lead the estimates towards the null hypothesis. Finally, unknown or non-observable confounding factors not controlled during the design or analysis stages may be present.
We performed FMM because of its flexibility, not only to identify subpopulations but also to uncover structures within the data. Moreover, the heterogeneous effect of an exposure on the complete distribution of a variable has been frequently raised in the scientific literature48. In our opinion, FMM overcomes the problem of categorisation of a continuous variable with the consequent loss of information. This is particularly useful when exposure effects have different values and magnitudes on the dependent variable (i.e. age, sex and SES). Furthermore, post-estimation features are useful to estimate latent class marginal means, probabilities and model-comparison statistics (Stata Manual). In this context, biological issues find the optimal scenario as recently defined to characterise pre-clinical and clinical obesity on the basis of objective clinical manifestations49. This approach provides a clinical identity to each one of the obesity stages and offers a conceptual framework to formulate health policies for management strategies for preclinical and clinical obesity.
It is important to note that BMI was treated as a continuous variable to capture severity gradients within the obese population, rather than assuming homogeneity among individuals exceeding the standard cutoff. This analytical decision acknowledges that obesity is not a categorical state but a progressive condition, in which incremental increases in BMI are associated with disproportionate rises in metabolic dysfunction, systemic inflammation, and cardiovascular risk. Therefore, modelling BMI as a continuous variable allowed us to identify SDH profiles that not only coexist with obesity but also differentiate levels of severity within it, highlighting factors that may contribute to weight maintenance versus further progression. These findings reinforce the importance of recognising intra-group heterogeneity in obesity research and designing interventions tailored to the specific determinants operating at each severity gradient.
To date, no studies in Argentina have addressed the differential effects of sociodemographic and other factors on BMI increase among people with obesity. The three SDH profiles identified here offer an opportunity to advance more tailored and context-specific public health strategies for obesity prevention and management. First, the profile characterised by younger women with high socioeconomic status highlights a group rarely prioritised in traditional obesity prevention agendas. Although access to resources is presumed to be protective, our findings suggest that this population may still experience BMI increases driven by behavioural and lifestyle patterns linked to convenience food consumption, frequent food delivery use, and weak engagement in preventive health practices. In this sense, early interventions include clearer front-of-package labelling, reduced exposure to ultra-processed foods in educational settings, and increased nutrition education. These measures, combined with routine health monitoring and the development of healthier food environments in schools and workplaces, may support weight stabilisation before progression to more severe forms of obesity.
For individuals with more severe obesity (profiles II and III), our results highlight the need for more intensive and sustained public health actions. Beyond addressing comorbidities, diet quality, and PA, this group requires improved access to obesity treatment and expanded healthcare coverage. In addition, family and social context, particularly marital status, play an important role and should be considered in care strategies through counselling, social protection policies, and community-based programmes aimed at reducing social isolation and reinforcing daily support networks.
At the broader policy level, the limited progress in Argentina in implementing structural measures to curb obesity is notable. In this sense, our findings reinforce the need for provinces to adopt and operationalise National Law 27,642 on the Promotion of Healthy Eating and to complement it with local actions that align with the specific SDH profiles identified. From a collective health perspective, integrating these approaches could enhance the effectiveness of obesity prevention and treatment while reducing health inequalities. Ultimately, we aim to improve the quality of life for individuals and populations affected by obesity, reducing the risk of associated diseases. These findings reinforce the value of adopting a life-course perspective to the study of obesity, enabling a more comprehensive analysis of weight-gain trajectories and the differential influence of key determinants across ages and stages of obesity.
Acknowledgements
The authors would like to thank the participants and research team members who participated in the data collection.
Funding
This work was supported by the National Agency for Scientific and Technological Promotion (grants PICT-2020-A-03283, PIIDTA-IA RESOL-2024-61-E-UNC-SECYT#ACTIP) and the Science and Technology Department of the University of Córdoba (RESOL-2023-266-E-UNC-SECYT#ACTIP).
Declaração de Disponibilidade de Dados
Os dados de pesquisa estão disponíveis mediante solicitação ao autor de correspondência.
REFERENCES
1. Aballay LR, Eynard AR, Díaz MdelP, Navarro A, Muñoz SE. Overweight and obesity: a review of their relationship to metabolic syndrome, cardiovascular disease, and cancer in South America. Nutr Rev. 2013; 71(3):168–79.
2. World Health Organization. The top 10 causes of death. [online]. 2020. https://www.who.int/es/news-room/fact-sheets/detail/the-top-10-causes-of-death
3. Instituto Nacional de Estadísticas y Censos y el Ministerio de Salud de la Nación. 4° Encuesta Nacional de Factores de Riesgo para enfermedades no transmisibles Argentina. [online]. 2019. https://www.indec.gob.ar/ftp/cuadros/publicaciones/enfr_2018_resultados_definitivos.pdf
4. Instituto Nacional de Estadísticas y Censos y el Ministerio de Salud de la Nación. 4° Encuesta Nacional de Factores de Riesgo para enfermedades no transmisibles Argentina. [online]. 2015. https://www.indec.gob.ar/indec/web/Nivel4-Tema-4-32-68
5. Aballay LR, Osella AR, De La Quintana AG, Diaz MdelP. Nutritional profile and obesity: results from a random-sample population-based study in Córdoba, Argentina. Eur J Nutr. 2015;55(2):675-685.
6. World Health Organization. A conceptual framework for action on the social determinants of health. World Health Organization. [online]. 2010. https://iris.who.int/handle/10665/44489
7. Pan-American Health Organization. Salud en las Américas edición de 2012. Panorama regional y perfiles de país. [online]. 2012. https://iris.paho.org/handle/10665.2/3272
8. Solar O, Irwin A. Social determinants, political contexts and civil society action: a historical perspective on the Commission on Social Determinants of Health. Health Promot J Austr. 2006 Dec;17(3):180-5.
9. World Health Organization. Closing the gap in a generation: health equity through action on the social determinants of health [online]; 2008. whqlibdoc.who.int/publications/2008/9789241563703_eng.pdf
10. Caldwell AE, Sayer RD. Evolutionary considerations on social status, eating behaviour, and obesity. Appetite. 2019;132:238-248.
11. Cockerham WC. Theoretical Approaches to Research on the Social Determinants of Obesity. Am J Prev Med. 2022;63(1):S8–S17.
12. World Health Organization. Obesity: preventing and managing the global epidemic: report of a WHO consultation. World Health Organ Tech Rep Ser. 2000;894:i–253.
13. McLachlan GJ, Lee SX and Rathnayake SI. Finite Mixture Models. Annu. Rev. Stat. Appl. 2019. 6:355–78.
14. Instituto Nacional de Estadísticas y Censos. Censo Nacional de Población, Población, Hogares y Viviendas 2022: Resultados definitivos. [online]. 2022. https://censo.gob.ar/index.php/datos_definitivos_total_pais/
15. Dirección General de Estadística y Censos de la Provincia de Córdoba. Total de hogares según condición de Necesidades Básicas Insatisfechas -NBI- desagregado por tipo de NBI por barrio. [online]. 2010. https://datosestadistica.cba.gov.ar/dataset/f2ee0b23-7d07-42a5-ad52-515af75b2a25/resource/3e725589-5fbf-4b61-8077-91ea56d01978/download/desd_cn2010_muncba_barrios-1.xlsx
16. Navarro A, Osella AR, Guerra V, Muñoz S, Lantieri M, Eynard AR. Reproducibility and validity of food-frequency questionnaire in assessing dietary intakes and food habits in epidemiological cancer studies in Argentina. J Exp Clin Cancer Res. 2001;20(3):365-370.
17. International Life Sciences Institute Argentina. Guía Visual de porciones y pesos de alimentos (1ª ed.). Ciudad Autónoma de Buenos Aires: ILSI Argentina; 2018. ISBN: 978-987-21507-5-4.
18. Hurtig-Wennlof A, Hagstromer M, Olsson LA. The International Physical Activity Questionnaire modified for the elderly: aspects of validity and feasibility. Public Health Nutr. 2009;13(11):1847–1854. doi.org/ 10.1017/S1368980010000157
19. Serón P, Muñoz S, Lanas F. Nivel de actividad física medida a través del cuestionario internacional de actividad física en población chilena. Rev Med Chil. 2010;138:1232-1239.
20. Charlson ME, Charlson RE, Peterson JC, Marinopoulos SS, Briggs WM, Hollenberg JP. The Charlson comorbidity index is adapted to predict costs of chronic disease in primary care patients. J Clin Epidemiol. 2008;61(12):1234–1240.
21. Krebs-Smith M, Pannucci T, Subar A, Kirkpatrick S, Lerman J, Tooze J, et al. Update of the Healthy Eating Index: HEI-2015. J Acad Nutr Diet 2018; 118 (9): 1591-1602.
22. Ministerio de Salud de la Nación. Guías Alimentarias para la Población. Argentina Buenos Aires: 2016.
23. StataCorp. Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC; 2021.
24. World Health Organisation. Obesity and overweight. [online]. 2021. https://www.who.int/es/news-room/fact-sheets/detail/obesity-and-overweight
25. Popkin BM. Nutrition Patterns and Transitions. Population and Development Review. 1993;19:138-157.
26. Pou SA, Wirtz Baker JM, Aballay LR. Epidemia de obesidad: evidencia actual, desafíos y direcciones futuras [Obesity epidemic: current evidence, challenges and future directions]. Medicina (B Aires). 2023;83(2):283-289.
27. Christine PJ, Diez Roux AV, Wing JJ, Alazraqui M, Spinelli H. Temporal trends in BMI in Argentina by socio-economic position and province-level economic development, 2005-2009. Public Health Nutr. 2015;18(5):817–826.
28. Fernández MS, Pilatti A, Pautassi RM. Eating-to-Cope Motives and Uncontrolled Eating as Mediators Between Negative Emotional States and Food Addiction Among Argentinean Young Adults. Int J Ment Health Addiction. 2024;22:1433–1451.
29. Van Langeveld AWB, Teo PS, de Vries JHM, Feskens EJM, de Graaf C, Mars M. Dietary taste patterns by sex and weight status in the Netherlands. Br J Nutr. 2018;119(10):1195-1206.
30. The NS, Gordon-Larsen P. Entry into romantic partnership is associated with obesity. Obesity. 2009;17(7):1441–7.
31. Umberson D. Family status and health behaviors: social control as a dimension of social integration. J Health Soc Behav. 1987;28(3):306-19.
32. Barbany M, Foz M. Obesidad: concepto, clasificación y diagnóstico. An Sist Sanit. 2002;25:7–16.
33. Coleman JC, Hendry LB. Psicología de la adolescencia. Madrid: Morata;2003.
34. Slevec J, Tiggemann M. Attitudes toward cosmetic surgery in middle aged women: Body image, aging anxiety, and the media. Psychology of Women Quarterly. 2010;34(1):65–74. doi.org/10.1111/j.1471-6402.2009.01542.x
35. Ek S. Gender differences in health information behaviour: a Finnish population-based survey. Health Promot Int. 2015;30(3):736-45. doi.org/10.1093/heapro/dat063.
36. Dinour L, Leung MM, Tripicchio G, Khan S, Yeh MC. The Association between Marital Transitions, Body Mass Index, and Weight: A Review of the Literature. J Obes. 2012;2012:294974.
37. Sobal J, Rauschenbach B, Frongillo EA. Marital status changes and body weight changes a US longitudinal analysis. Soc Sci Med. 2003;56(7):1543–55.
38. Hosseini Z, Veenstra G, Khan N, Conklin A. Social relationships andadiposity in older adults: population-based cross-sectional analysis of the Canadian Longitudinal Study on Aging. Curr Dev Nutr. 2020;4:nzaa063_040. doi:10.1093/cdn/nzaa063_040
39. Teachman J. Body Weight, Marital Status, and Changes in Marital Status. J Fam Issues. 2016;37(1):74-96.
40. Zhou J, Tang R, Wang X, Li X, Heianza Y, Qi L. Improvement of social isolation and loneliness and excess mortality risk in people with obesity. JAMA Netw Open. 2024;7(1):e2352824-e2352824. doi:10.1001/jamanetworkopen.2023.52824
41. Madani Civi R, Mehranfar S, Plunkett R, Veenstra G, Conklin AI. A systematic review of social connections as determinants of obesity: Longitudinal evidence limited to marital transitions. Obes Rev. 2024;25(11):e13819. doi: 10.1111/obr.13819
42. Ministerio de Salud de la Nación. Guías Alimentarias para la Población. [online] Buenos Aires, Argentina 2016. https://bancos.salud.gob.ar/sites/default/files/2020-08/guias-alimentarias-para-la-poblacion-argentina_manual-de-aplicacion_0.pdf 476-87.
43. Carneiro IP, Elliott SA, Siervo M, Padwal R, Bertoli S, Battezzati A, et al. Is Obesity Associated with Altered Energy Expenditure? Adv Nutr. 2016;7(3)
44. Batsis JA, Gill LE, Masutani RK, Adachi-Mejia AM, Blunt HB, Bagley PJ, et al. Weight loss interventions in older adults with obesity: a systematic review of randomized controlled trials since 2005. J Am Geriatr Soc. 2017;65:257–268
45. Jiang BC, Villareal DT. Weight loss-induced reduction of bone mineral density in older adults with obesity. J Nutr Gerontol Geriatr. 2019;38:100–114
46. Jiang BC, Villareal DT. The Benefits of Exercise Training in Combination With Weight Loss Therapies. Diabetes. 2025;74(12):2199–2206. doi.org/10.2337/dbi25-0001
47. Zapata ME, Rovirosa A, Carmuega E. Cambios en el patrón de consumo de alimentos y bebidas en las últimas dos décadas en Argentina. Salud Colectiva. 2016;12(4):473– 486. doi.org/10.18294/sc.2016.936.
48. Cameron AC, Trivedi PK. Microeconometrics using Stata. 2nd ed. College Station (TX): Stata Press; 2022.
49. Rubino F, Cummings DE, Eckel RH, Cohen RV, Wilding JPH, Brown WA, et al. Definition and diagnostic criteria of clinical obesity. Lancet Diabetes Endocrinol. 2025 Mar;13(3):221-62.