Abstracts

Introduction

Overweight and obesity are among the major global public health issues of the 21st century ¹1. Phelps NH, Singleton RK, Zhou B, Heap RA, Mishra A, Bennett JE, et al. Worldwide trends in underweight and obesity from 1990 to 2022: a pooled analysis of 3663 population-representative studies with 222 million children, adolescents, and adults. Lancet 2024; 403:1027-50.^,22. Muñoz FL, Pou SA, Diaz MDP. An empirically derived "prudent" dietary pattern is associated with lower obesity occurrence: modeling and mapping from a national nutrition survey. Nutr Res 2023; 109:26-34.. Despite progress in understanding the obesity epidemic, strategies to control and reduce it, many focusing on promoting physical activity and healthy lifestyles, have not yet produced the expected results. In recent decades, as the world’s population has undergone a nutrition transition, a decline in physical activity levels has been observed ³3. Popkin BM, Adair LS, Ng SW. Global nutrition transition and the pandemic of obesity in developing countries. Nutr Rev 2012; 70:3-21.. In line with this global trend, the prevalence of obesity in Colombia increased from 13.7% in 2005 to 18.7% in 2015, and leisure-time physical activity decreased significantly from 2005 to 2010 ⁴4. González S, Sarmiento OL, Lozano Ó, Ramírez A, Grijalba C. Physical activity levels among Colombian adults: inequalities by gender and socioeconomic status. Biomedica 2014; 34:447-59.. Interestingly, Colombia holds valuable information, rigorously collected via a nationwide survey called the Colombian National Survey of Nutritional Status (ENSIN; Encuesta Nacional de Situación Nutricional), which has not yet been fully used for a detailed study of the links between physical activity patterns and obesity. This could provide new insights into this issue.

Physical activity, broadly defined as all forms of movement including during leisure time, for transport to get to and from places, and/or as part of a person’s work ⁵5. World Health Organization. Physical activity. https://www.who.int/news-room/fact-sheets/detail/physical-activity (accessed on 10/Jun/2024).
https://www.who.int/news-room/fact-sheet... , is considered a modifiable factor associated with many health conditions, including obesity and its associated chronic diseases ⁶6. Cuadri Fernández J, Tornero Quiñones I, Sierra Robles Á, Sáez Padilla JM. Revisión sistemática sobre los estudios de intervención de actividad física para el tratamiento de la obesidad. Retos 2018; 33:261-6.^,77. Jakicic J, Powell K, Campbell W, Dipietro L, Pate RR, Pescatello LS, et al. Physical activity and the prevention of weight gain in adults: a systematic review. Med Sci Sports Exerc 2019; 51:1262-9.. Its health advantages include psychological effects, maintenance of physical fitness, and promotion of healthy behaviors, which result in greater well-being and quality of life ⁸8. Marquez D, Aguiñaga S, Vásquez P, Conroy D, Erickson KI, Hillman C, et al. A systematic review of physical activity and quality of life and well-being. Transl Behav Med 2020; 10:1098-109.. In particular, the favorable effects of physical activity on maintaining a healthy body weight are well established, but most of the evidence has focused on overall physical activity and regular physical activity during leisure time ⁷7. Jakicic J, Powell K, Campbell W, Dipietro L, Pate RR, Pescatello LS, et al. Physical activity and the prevention of weight gain in adults: a systematic review. Med Sci Sports Exerc 2019; 51:1262-9.. Comparatively, fewer studies have focused on other domains of physical activity, such as active transportation (cycling and walking to work) or sedentary behavior (i.e., sitting or lying down activities during waking hours that require low energy expenditure). Although interest on this subject is growing, the role of different domains of physical activity on obesity occurrence needs a better explanation.

In physical activity research, some review studies have explored the impact of time spent in sedentary behavior on several health outcomes, including obesity in adulthood ⁹9. Rezende L, Rodrigues M, Rey JP, Matsudo V, Luiz O. Sedentary behavior and health outcomes: an overview of systematic reviews. PLoS One 2014; 9:e105620.^,1010. Biddle S, Bengoechea E, Pedisic Z, Bennie J, Vergeer I, Wiesner G. Screen time, other sedentary behaviours, and obesity risk in adults: a review of reviews. Curr Obes Rep 2017; 6:134-47.^,1111. Katzmarzyk PT, Powell KE, Jakicic JM, Troiano RP, Piercy K, Tennant B, et al. Sedentary behavior and health: update from the 2018 Physical Activity Guidelines Advisory Committee. Med Sci Sports Exerc 2019; 51:1227-41.. Overall, it is agreed that the evidence needed to understand this complex association is inconclusive. Regarding the link between obesity/adiposity and sedentary behaviors, more consistent associations have been shown for screen time (mainly TV viewing) ⁹9. Rezende L, Rodrigues M, Rey JP, Matsudo V, Luiz O. Sedentary behavior and health outcomes: an overview of systematic reviews. PLoS One 2014; 9:e105620. and car use ¹⁰10. Biddle S, Bengoechea E, Pedisic Z, Bennie J, Vergeer I, Wiesner G. Screen time, other sedentary behaviours, and obesity risk in adults: a review of reviews. Curr Obes Rep 2017; 6:134-47., although in general these factors have been studied independently. The means of transportation is another topic increasingly studied since strong evidence points that active transport behavior (primarily walking and cycling) can result in substantial health benefits ¹²12. Winters M, Buehler R, Götschi T. Policies to promote active travel: evidence from reviews of the literature. Curr Environ Health Rep 2017; 4:278-85.^,1313. Saunders LE, Green JM, Petticrew MP, Steinbach R, Roberts H. What are the health benefits of active travel? A systematic review of trials and cohort studies. PLoS One 2013; 8:e69912.^,1414. Dinu M, Pagliai G, Macchi C, Sofi F. Active commuting and multiple health outcomes: a systematic review and meta-analysis. Sports Med 2019; 49:437-52.^,1515. Andersen LB. Active commuting: an easy and effective way to improve health. Lancet Diabetes Endocrinol 2016; 4:381-2.. However, while some studies report that active transportation appears to be associated with a lower risk of obesity or healthier body weight ¹⁶16. Wu J, Li Q, Feng Y, Bhuyan SS, Tarimo CS, Zeng X, et al. Active commuting and the risk of obesity, hypertension and diabetes: a systematic review and meta-analysis of observational studies. BMJ Glob Health 2021; 6:e005838.^,1717. Flint E, Cummins S. Active commuting and obesity in mid-life: cross-sectional, observational evidence from UK Biobank. Lancet Diabetes Endocrinol 2016; 4:420-35., others note that the evidence is still unclear ¹³13. Saunders LE, Green JM, Petticrew MP, Steinbach R, Roberts H. What are the health benefits of active travel? A systematic review of trials and cohort studies. PLoS One 2013; 8:e69912.^,1818. Wanner M, Götschi T, Martin E, Kahlmeier S, Martin B. Active transport, physical activity, and body weight in adults: a systematic review. Am J Prev Med 2012; 42:493-502.^,1919. Brown V, Moodie M, Mantilla A, Veerman J, Carter R. Active transport and obesity prevention: a transportation sector obesity impact scoping review and assessment for Melbourne, Australia. Prev Med 2017; 96:49-66..

Given the above, it is reasonable to assume that the notion of physical activity involves a complex phenomenon that can be represented as a structure or system of interrelated variables, which presents a methodological challenge to its study. The generalized structural equation model (GSEM) emerges as a useful analytic strategy to address this challenge because it allows examining theoretical connections between variables, whether observable or not directly observable, for a comprehensive analysis. The classic models are insufficient to evaluate the effects of measurable and latent variables associated with obesity. Finding latent dimensions, GSEM enables verifying the suitability of theoretical models in the study population, thereby offering a more robust approach to understanding the intricate dynamic between multiple variables ²⁰20. Darbandi M, Najafi F, Pasdar Y, Mostafaei S, Rezaeian S. Factors associated with overweight and obesity in adults using structural equation model: mediation effect of physical activity and dietary pattern. Eat Weight Disord 2020; 25:1561-71.^,2121. Tarka P. An overview of structural equation modeling: its beginnings, historical development, usefulness and controversies in the social sciences. Qual Quant 2018; 52:313-54..

This study aimed to (a) identify latent (unobservable) dimensions representing specific physical activity-related behaviors and estimate their direct and indirect effects on obesity, and (b) explore the spatial distribution of the population scores of adherence to the identified latent variables of physical activity and the obesity prevalence in Colombia. We hypothesized that there are underlying dimensions (not directly observable) representing non unidimensional aspects of the physical activity associated with the presence of obesity.

Methods

Study design and sample

In this cross-sectional study, the population-based dataset corresponding to the last ENSIN was used ²²22. Ministerio de Salud y Protección Social. Encuesta Nacional de Situación Nutricional en Colombia - 2015. Bogotá: Instituto Colombiano de Bienestar Familiar; 2015.. The last edition of this survey was conducted in 2015, following a multistage stratified random sampling design, with urban and regional coverage. Further details of the ENSIN survey can be found elsewhere ²2. Muñoz FL, Pou SA, Diaz MDP. An empirically derived "prudent" dietary pattern is associated with lower obesity occurrence: modeling and mapping from a national nutrition survey. Nutr Res 2023; 109:26-34..

First, a subset of 10,635 people aged 18-64 years with complete information on physical activity and energy intake was extracted from the total number of participants in the 2015 ENSIN. The exclusion criteria included pregnant or breastfeeding women, incomplete anthropometric data, report of physical or mental disabilities, and unreliable energy intake reporting (< 1st percentile or > 99th percentile). After applying the criteria, the final sample size was 9,658.

Data and instruments

Sociodemographic characteristics

The following sociodemographic variables were selected from the ENSIN in order to characterize the study population and/or adjust the estimates: sex (male, female), age group in years (18-29, 30-49, 50-64), ethnicity (people of African ancestry, Indigenous people, people of other ethnic origins), schooling level (highest level of schooling attained: primary, secondary, or tertiary), and wealth index (levels by quartiles: 1 - low, 2 - middle-low, 3 - middle-high, 4 - high). In particular, the ENSIN administers a structured questionnaire on socio-demographic and economic characteristics to the household head, developed by the Colombian Demographic and Health Survey.

Anthropometric variables

In this study, obesity was defined as a body mass index (BMI) greater than or equal to 30, following the World Health Organization (WHO) criterion. The WHO states that BMI, estimated as weight (kg)/height² (m²), is a surrogate marker of fatness ²³23. World Health Organization. Obesity and overweight. https://www.who.int/news-room/facteets/detail/obesity-and-overweight (accessed on 13/Jun/2024).
https://www.who.int/news-room/facteets/d... and can therefore be considered a suitable proxy for obesity.

The ENSIN used anthropometric measurements of height and weight carried out by trained personnel. Subjects were asked to be barefoot and wear light clothing. A stadiometer (1mm precision, Shor Productions LLC, https://weighandmeasure.com/shorrboards) and a digital scale (model 874, 100g precision, SECA; https://www.seca.com) were used.

Behavioral factors

Data on physical activity and dietary intake from the ENSIN was considered in the modeling phase of this study. Specifically, to construct the physical activity dimensions (latent construct), the study considered time (minutes/week) spent using different modes of transport (passive transport in a motor vehicle; cycling or walking) and using screens (computer, smartphone, or other digital devices for playing video games). Other behavioral variables derived from the ENSIN data were leisure time (minutes/week) dedicated to moderate or vigorous physical activity (from the results of the International Physical Activity Questionnaire - IPAQ), and daily energy intake (kcal/day, estimated from two nonconsecutive 24-hour recalls applied by the ENSIN) ²2. Muñoz FL, Pou SA, Diaz MDP. An empirically derived "prudent" dietary pattern is associated with lower obesity occurrence: modeling and mapping from a national nutrition survey. Nutr Res 2023; 109:26-34..

For physical activity, the ENSIN employs the IPAQ (long format) ²⁴24. Medina C, Barquera S, Janssen I. Validity and reliability of the International Physical Activity Questionnaire among adults in Mexico. Rev Panam Salud Pública 2013; 34:21-8. and a structured questionnaire on sedentary behaviors. Dietary data was obtained from the 24-hour diet recall technique. Further details are available in previous publications ²2. Muñoz FL, Pou SA, Diaz MDP. An empirically derived "prudent" dietary pattern is associated with lower obesity occurrence: modeling and mapping from a national nutrition survey. Nutr Res 2023; 109:26-34..

Statistical analysis

To describe the sociodemographic characteristics of the sampled individuals, frequency tables were constructed and summary measures were estimated. The chi-square test and the Mann-Whitney U test were used for comparative purposes.

In the modeling phase of the study, based on the available data from the ENSIN survey, two physical activity dimensions were theoretically proposed, considering their potential impacts on the energy balance: (i) sedentary behaviors mediated by screens (leisure time that the person spends on computer and smartphone or playing video games, measured in minutes/week), and (ii) the modes of transport in daily life (time spent on motor vehicle, cycling, and walking as means of transport, in minutes/week).

A GSEM was proposed, combining exposure and measurement models to define a disease model, as graphically illustrated in Figure 1. As shown, measurement models were proposed to estimate latent (unobserved, represented by ellipses) variables from their constituent indicators (measurable variables with non-Gaussian distribution, represented by rectangles). In turn, the exposure model estimates the direct effects of some covariates on the expected value of obesity occurrence. Finally, the disease model combines the two aforementioned models integrating both direct and indirect effects of the covariates and the latent variables (representing specific physical activity-related behaviors, for example) on the outcome.

The modified Poisson regression model with a robust variance estimator ²⁵25. Cummings P. Methods for estimating adjusted risk ratios. Stata J 2009; 9:175-96.^,2626. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004; 159:702-6. that adequately estimates the prevalence ratios (PR) was selected in this study, in comparison with the logistic regression model. When the outcome is binary, the exponential coefficients estimated by Poisson regression are risk ratios instead of incidence-rate ratios. Then, Y ₁ was considered the outcome variable, representing the obesity occurrence (1/0, yes/no), in which π(x) = Pr (Y ₁ = 1∣x).

Since π(x) is a positive function, the logarithm link function is a natural choice for modeling the expected value, that is: log(π(x)) = α + L ₁ β ₁ + L ₂ β ₂ + β ₂ X ₁ + ··· + β ₈ X ₆ . Figure 1 shows this expression, in which the set of endogenous variables contains computer and smartphone use, television viewing, playing video games, passive (car) transportation, cycling, and walking as modes of transport indicators. The new dimensions or latent variables were expressed as L ₁ and L ₂ (exogenous variables). The covariates sex, age, energy intake, socioeconomic score, and leisure physical activity were also included.

Goodness-of-fit tests were applied to identify the probability function with the best fit for each variable distribution. Kaiser-Meyer-Olkin (KMO) was used to estimate sample size adequacy and the sphericity test was used to assess the correlation between the variables. Based on the factorial analysis, two measurement models were estimated to identify the factorial loads and the adequacy of the variables of the physical activity dimensions that are hypothesized to be associated with obesity.

The structural model was estimated by maximum likelihood with 95% confidence intervals (95%CI). Considering the Akaike information criterion (AIC) and the Bayesian information criterion (BIC), the most parsimonious model, with significant path coefficients and theoretical significance, was selected. The expansion factors supplied by the database were used to adjust the statistical analyses. Stata, version 14.0 (https://www.stata.com), and RStudio, version 2022.02.3 (https://rstudio.com/), softwares were used.

Then, a score coefficient was estimated for each latent dimension identified, using the Bayesian empirical prediction method. This score represents the degree of adherence that each subject has to each latent variable. Therefore, the tertiles of the distribution of the scores define three categories: low, moderate, and high levels of adherence.

Mapping

At the population level, the prevalence of obesity (% of people with a BMI greater than or equal to 30) was estimated in the 32 departments of Colombia and its capital, and its spatial distribution (in quartiles) was mapped. The QGIS software, version 3.26.3 (https://qgis.org/en/site/), was employed to create a map illustrating the adherence to the latent dimensions that showed a significant association with obesity.

Ethical aspects

The Ethics Committee of Profamilia, under Resolution n. 8,430/1993 of the Colombian Ministry of Health, approved the ENSIN, as it complied with the guidelines established in the Declaration of Helsinki. The database used in this work is in the public domain.

Results

Table 1 describes the sample of 9,658 participants (56.2% females, mean age 38.5±13.1 years). Obesity prevalence was 19.1% (95%CI: 17.8; 20.4), with a greater prevalence in females (23.4%) compared to males (13.7%) (p < 0.001). About 75% of the subjects had completed high school and a third belonged to the lowest wealth quartile. Most participants (90.9%) reported no ethnic affiliation. When we carried out a descriptive analysis stratified by the occurrence of obesity, most people with obesity were females (68.6%), the age group with the highest prevalence was 30-49 years (49.9%), and the vast majority (74.2%) had a high school education. All the sociodemographic variables were significantly associated with having obesity (yes/no).

In addition, we found that 50% of the sample spent 140 minutes/week traveling by passive means of transport (motorized vehicles) and about 70 minutes on foot. Regarding sedentary behaviors, watching television is the activity to which people dedicate the most time (median 420 minutes/week), followed by using a computer or smartphone (45 minutes/week). In a comparative analysis, people with obesity spend more time watching television (63 minutes/week) than people without obesity (p < 0.001), and this latter group spends more time using computers and smartphones (p < 0.001). We also found a significant difference in terms of weekly minutes dedicated to leisure physical activity (p = 0.001), with a higher maximum value in people without obesity.

To estimate the GSEM, goodness-of-fit tests first indicated that the Weibull probability function is adequate for the asymmetry observed in weekly time spent on transportation and use of screens. For the measurement model, Bartlett’s sphericity test (p < 0.001) confirmed that the variables are correlated, as the similarity matrix was not an identity matrix. The overall KMO test value was 0.526.

Figure 2 illustrates the two latent variables, one called “Screens use” that represents the minutes per week in front of screens, including the use of computer and smartphone (as reference), television viewing (β = 3.9, p < 0.001), and use of video games (β = -20.3, p < 0.001) variables. The other, called “Active transportation”, is made up of the weekly minutes spent in transportation by motor vehicles (as reference), cycling (β = -5.4, p < 0.001), and walking (β = 0.8, p < 0.001).

As shown in Table 2, “Active transportation” was established as a protective factor for obesity (PR = 0.86, 95%CI: 0.7; 0.9), with the use of the bicycle being the most contributing indicator, detrimental to the use of passive transportation. The variable “Screens use” showed no significant effect on obesity.

The exogenous variables were directly associated with obesity: age (PR = 1.02, 95%CI: 1.02; 1.03), sex (female: PR = 1.95, 95%CI: 1.8; 2.2), ethnicity (Afro-descendant: PR = 1.36, 95%CI: 1.18; 1.58, reference: no ethnic origin) and wealth index level (middle-low: PR = 1.26, 95%CI: 1.1; 1.4, reference: low). Conversely, the variable leisure-time physical activity was directly related to obesity as a protective factor (PR = 0.999, 95%CI: 0.998; 0.999) and indirectly through the dimension “Active transportation” (PR = -0.0006, 95%CI: -0.0068; -0.00059) and “Screens use” (PR = -0.0002, 95%CI: -0.00019; -0.00015).

Figure 3a shows the spatial distribution of obesity prevalence in 32 departments and in Bogotá. The departments with higher prevalence are concentrated in the southwest of the country and in coastal departments. In total, 57% of all departments have a prevalence above the national average (19.1%).

Regarding adherence to the “Active transportation” dimension (Figure 3b), we found it to be mainly concentrated in the most densely populated and industrialized departments, which are located mainly in the center of the country. As already shown, this area presents low prevalence levels, below the national averages. This suggests possible coincidences of effects, in which the departments with greater “Active transportation” also have a lower burden of obesity prevalence.

Discussion

The results of this research are pioneering in exploring the direct and indirect effects of weekly time spent in screens use and active transportation, as well as their association with obesity in adults in Colombia. Our latent dimensions, identified via a GSEM estimation, were named “Screens use” and “Active transportation”, according to their conformation indicators, in an attempt to define two interpretable dimensions. Both dimensions were related to the leisure time physical activity, but only the “Active transportation” dimension had a protective effect on obesity occurrence. In addition, it showed higher levels of adherence in the departments with better socioeconomic conditions and less prevalence of obesity.

Active transportation

Although the literature has evaluated the impact of different forms of transportation on health, especially referring to injuries and pollution from motor vehicle emissions, the impact of means of transportation on BMI-related morbidity has not been explored in depth ²⁷27. Nazelle A, Nieuwenhuijsen M, Antó J, Brauer M, Briggs D, Braun C, et al. Improving health through policies that promote active travel: a review of evidence to support integrated health impact assessment. Environ Int 2011; 37:766-77.. Several systematic reviews indicate that evidence linking transport and obesity remains unclear ¹³13. Saunders LE, Green JM, Petticrew MP, Steinbach R, Roberts H. What are the health benefits of active travel? A systematic review of trials and cohort studies. PLoS One 2013; 8:e69912.^,1818. Wanner M, Götschi T, Martin E, Kahlmeier S, Martin B. Active transport, physical activity, and body weight in adults: a systematic review. Am J Prev Med 2012; 42:493-502.^,1919. Brown V, Moodie M, Mantilla A, Veerman J, Carter R. Active transport and obesity prevention: a transportation sector obesity impact scoping review and assessment for Melbourne, Australia. Prev Med 2017; 96:49-66.; however, interventions promoting active transportation can provide benefits to prevent obesity. Cycling and walking in daily life are encouraged as part of strategies for obesity control and health promotion since active transportation shows health benefits, reduces air and noise pollution, and improves social and urban capital ¹²12. Winters M, Buehler R, Götschi T. Policies to promote active travel: evidence from reviews of the literature. Curr Environ Health Rep 2017; 4:278-85.^,2828. Filigrana P, Levy J, Gauthier J, Batterman S, Adar S. Health benefits from cleaner vehicles and increased active transportation in Seattle, Washington. J Expo Sci Environ Epidemiol 2022; 32:538-44.. In particular, our findings showed that the active transportation dimension, mainly represented by cycling as a means of transport, presents a protective effect against obesity. In line with this, some studies report that active commuting or transportation was associated with a lower risk of obesity and other non-communicable diseases among adults ¹⁶16. Wu J, Li Q, Feng Y, Bhuyan SS, Tarimo CS, Zeng X, et al. Active commuting and the risk of obesity, hypertension and diabetes: a systematic review and meta-analysis of observational studies. BMJ Glob Health 2021; 6:e005838., as well as healthier body weight and composition in midlife ¹⁷17. Flint E, Cummins S. Active commuting and obesity in mid-life: cross-sectional, observational evidence from UK Biobank. Lancet Diabetes Endocrinol 2016; 4:420-35.. Similarly, the multi-country Latin American Health and Nutrition Study (ELANS, acronym in Spanish) found that time spent in active transportation was significantly associated with lower body mass index ²⁹29. Habinger J, Chávez J, Matsudo S, Kovalskys I, Gómez G, Rigotti A, et al. Active transportation and obesity indicators in adults from Latin America: ELANS Multi-Country Study. Int J Environ Res Public Health 2020; 17:6974.. These favorable effects may be related to the contribution of active transport behaviors to overall daily physical activity ³⁰30. Celis-Morales CA, Lyall DM, Welsh P, Anderson J, Steell L, Guo Y, et al. Association between active commuting and incident cardiovascular disease, cancer, and mortality: prospective cohort study. BMJ 2017; 357:j1456., increasing daily energy expenditure. However, the association between overall or leisure-time physical activity and other domains of physical activity such as active transportation remains not fully understood ¹³13. Saunders LE, Green JM, Petticrew MP, Steinbach R, Roberts H. What are the health benefits of active travel? A systematic review of trials and cohort studies. PLoS One 2013; 8:e69912.^,1818. Wanner M, Götschi T, Martin E, Kahlmeier S, Martin B. Active transport, physical activity, and body weight in adults: a systematic review. Am J Prev Med 2012; 42:493-502.^,3131. Menai M, Charreire H, Feuillet T, Salze P, Weber C, Enaux C, et al. Walking and cycling for commuting, leisure and errands: relations with individual characteristics and leisure-time physical activity in a cross-sectional survey (the ACTI-Cités project). Int J Behav Nutr Phys Act 2015; 12:150.. As a contribution to this field, using a rigorous estimation procedure based on GSEM, we found a direct protective effect on the occurrence of obesity in people who practice physical activity in their leisure time, which in turn was related to active transportation. This fact is consistent with other studies, which state that associations with obesity in the expected direction have been found in people who use active means of transport and practice regular physical activity ¹⁸18. Wanner M, Götschi T, Martin E, Kahlmeier S, Martin B. Active transport, physical activity, and body weight in adults: a systematic review. Am J Prev Med 2012; 42:493-502.^,3232. Smith L, Stubbs B, Hu L, Veronese N, Vancampfort D, Williams G, et al. Is active transport and leisure-time physical activity associated with inflammatory markers in US adults? A cross-sectional analyses from NHANES. J Phys Act Health 2019; 16:540-6.. Moreover, we highlight that our latent variable called active transportation was mainly characterized by cycling in contrast to motor vehicle commuting. Our results corroborate previous studies suggesting that car use may be associated with a higher risk of obesity ¹⁰10. Biddle S, Bengoechea E, Pedisic Z, Bennie J, Vergeer I, Wiesner G. Screen time, other sedentary behaviours, and obesity risk in adults: a review of reviews. Curr Obes Rep 2017; 6:134-47.^,3333. McCormack GR, Virk JS. Driving towards obesity: a systematized literature review on the association between motor vehicle travel time and distance and weight status in adults. Prev Med 2014; 66:49-55.. Taken together, these findings support the recommendation to choose or promote active transportation instead of passive transportation whenever possible.

Screens use

The “Screens use” latent variable showed no significant effect on obesity. Although international organizations report that a sedentary lifestyle and physical inactivity are risk factors associated with all-cause mortality ³⁴34. Lavie C, Ozemek C, Carbone S, Katzmarzyk PT, Blair SN. Sedentary behavior, exercise, and cardiovascular health. Circ Res 2019; 124:799-815., many of which are obesity-related mortality causes, our results agree with reviews reporting inconclusive evidence in adults ⁹9. Rezende L, Rodrigues M, Rey JP, Matsudo V, Luiz O. Sedentary behavior and health outcomes: an overview of systematic reviews. PLoS One 2014; 9:e105620.^,1010. Biddle S, Bengoechea E, Pedisic Z, Bennie J, Vergeer I, Wiesner G. Screen time, other sedentary behaviours, and obesity risk in adults: a review of reviews. Curr Obes Rep 2017; 6:134-47.. However, the most consistent associations refer to screen time (mainly TV viewing) and markers of adiposity ¹⁰10. Biddle S, Bengoechea E, Pedisic Z, Bennie J, Vergeer I, Wiesner G. Screen time, other sedentary behaviours, and obesity risk in adults: a review of reviews. Curr Obes Rep 2017; 6:134-47., similar to the results of our study, in which people with obesity reported spending more screen time watching television than people without obesity. To clarify the obesogenic effect of screen time, which appears to be stronger in children and adolescents than in adults, much more information is needed on the influence of potential confounder or mediator factors, including unhealthy dietary behaviors ⁹9. Rezende L, Rodrigues M, Rey JP, Matsudo V, Luiz O. Sedentary behavior and health outcomes: an overview of systematic reviews. PLoS One 2014; 9:e105620.. Our study advances in this direction by presenting a theoretical model of obesity that incorporates, in the estimation process, selected adjustment variables (including total energy intake, leisure time physical activity, among others) and the potential covariation between the latent dimensions of screens use and active transportation. However, other relevant aspects not investigated in our study require further research. Some of the gaps identified in the literature in this field include the effects of frequent breaks from sitting, the role of unhealthy dietary patterns when using screens, and the consideration of different types of sedentary occupations and age groups ⁹9. Rezende L, Rodrigues M, Rey JP, Matsudo V, Luiz O. Sedentary behavior and health outcomes: an overview of systematic reviews. PLoS One 2014; 9:e105620.^,1010. Biddle S, Bengoechea E, Pedisic Z, Bennie J, Vergeer I, Wiesner G. Screen time, other sedentary behaviours, and obesity risk in adults: a review of reviews. Curr Obes Rep 2017; 6:134-47..

Social inequalities and spatial distribution

In our study, the spatial distribution of obesity also suggests that, in departments of the southwest and coastal areas, where the population is largely Afro-Colombian or has Indigenous origins, the prevalence reaches 20%. In contrast, in the other Mediterranean departments, which are more densely populated and economically developed, the prevalence of obesity was lower, even when including industrialized capital cities with transportation problems ³⁵35. Prada E, Camargo D, Férmino R. Participation and physical activity in recreovia of Bucaramanga, Colombia. J Phys Act Health 2021; 18:1277-85.^,3636. Ospina J, López V, Botero V, Duque J. A database to analyze cycling routes in Medellin, Colombia. Data Brief 2020; 32:106162.^,3737. Ramírez R, Beltrán C, Correa J, Vivas A, Prieto D, Martínez J, et al. Factors associated with active commuting to school by bicycle from Bogotá, Colombia: The FUPRECOL study. Ital J Pediatr 2016; 42:97.^,3838. Gómez LF, Mosquera J, Gómez OL, Moreno J, Pinzon JD, Jacoby E, et al. Social conditions and urban environment associated with participation in the Ciclovia program among adults from Cali, Colombia. Cad Saúde Pública 2015; 31 Suppl:S257-66.^,3939. Arellana J, Alvarez V, Oviedo D, Guzman LA. Walk this way: pedestrian accessibility and equity in Barranquilla and Soledad, Colombia. Research in Transportation Economics 2021; 86:101024.. In line with this, some studies have documented that ethnic minorities, especially Afro-descendants, are more likely to have a high prevalence of obesity ⁴⁰40. Lee A, Cardel M, Feingold K, Chrousos G. Social and environmental factors influencing obesity. Endotext 2019; 12 oct. https://www.endotext.org/chapter/factors-influencing-obesity/social-and-environmental-factors-influencing-obesity/.
https://www.endotext.org/chapter/factors... . For the Latin America and the Caribbean region, as well as Colombia in particular, it has been highlighted that Indigenous and Afro-descendant peoples are ethnic-racial groups that are systematically exposed to multiple forms of material and social deprivation, trapped in a context of poverty and vulnerability ⁴¹41. Comisión Económica para América Latina y el Caribe. La matriz de la desigualdad social en América Latina. Santo Domingo: Naciones Unidas; 2016.^,4242. Viáfara CA, Palacios G, Banguera A. Ethnic-racial inequity in health insurance in Colombia: a cross-sectional study. Rev Panam Salud Pública 2021; 45:1.. Moreover, the historical lack of statistical visibility of these communities has prevented adequate identification and recognition of the magnitude and various manifestations of the poverty they experience, including health impacts. The inequitable access to health, education, and employment services translates into substantial disparities in the quality of life and physical and mental health. In a cross-sectional study based on data from the 2019 Colombian National Quality of Life Survey, the authors concluded that ethnic-racial status is a structural component of inequity in access to health services and heightens the disadvantages of people with low socioeconomic status ⁴²42. Viáfara CA, Palacios G, Banguera A. Ethnic-racial inequity in health insurance in Colombia: a cross-sectional study. Rev Panam Salud Pública 2021; 45:1.. This complex scenario could explain, in part, the unequal distribution of obesity across the Colombian territory.

In terms of public policies, Colombia presents a government plan of social development since 2014 that includes the use of bicycles. This policy promotes transportation in non-motorized forms, incorporation of road interconnection projects, construction of bikeways, and improvement of the regulatory framework for traffic on streets, highways, and bikeways. Although progress has been made in this field, given our results, public policies that encourage the use of active transportation must still be developed in remote areas where people of Indigenous and African origins mainly live (southwest and coastal areas).

Limitations and strengths of the study

Recall bias may be one of the possible limitations in identifying physical activity dimensions, although the ENSIN survey excluded the population most susceptible to this type of bias, identified using a cognitive function test. In addition, naming latent dimensions may be subjective, limiting replication and comparability across studies. Moreover, we recognize the limitations of using BMI alone to classify nutritional status, as additional measures are important in the diagnosis of obesity. However, BMI was found to be the most useful measure of overweight and obesity at the population level ⁴³43. Pou SA, Diaz MDP, Velázquez GA, Aballay LR. Sociodemographic disparities and contextual factors in obesity: updated evidence from a National Survey of Risk Factors for Chronic Diseases. Public Health Nutr 2021; 25:3377-89.. Despite these limitations, this work shows important strengths, including the use of a nationally representative probabilistic sample and standardized protocols. Moreover, the use of GSEM for data analysis allowed us to construct a theoretical model that can improve our understanding of the associations between the multiple factors involved in the development of obesity. For interpretation purposes, the Poisson regression model with a robust variance estimator, which was selected for this study, has been reported as one of the most consistent and efficient for estimating parameters of primary interest.

Conclusion

Our results suggest that active transportation, mainly cycling (as opposed to non-active transportation), is a dimension of physical activity-related behaviors that may influence the burden of obesity, an association not found for screens use. Moreover, the spatial distribution of active transportation behaviors and obesity in Colombia seems to be shaped by social factors such as socioeconomic barriers, ethnic-racial inequalities, and limited economic development. These findings highlight the need for primary prevention interventions and public policies that promote social equity across regions, particularly in terms of access to infrastructure that encourages active transportation among the most socioeconomically disadvantaged populations. Finally, we emphasize that daily transportation choices are important factors to consider in public health recommendations, especially considering that cycling and walking to work have been highlighted as effective ways to incorporate regular physical activity into a sedentary lifestyle ¹⁴. This is particularly important given the context of declining levels of physical activity worldwide.

Acknowledgments

This work was supported by the Secretary of Science and Technology (SeCyT) of the Córdoba National University (UNC) (project grant n. 212/2020). F. L. Muñoz received a fellowship from the SeCyT/UNC (Res. 1,559/2019). The authors are grateful to the Colombian Family Welfare Institute (ICBF) and Colombian Ministry of Security and Social Protection for allowing access to the database (ENSIN 2015) and Gabriela Díaz Cortez for doing the language quality check.

References

Publication Dates

History