Mendes et al.

BMC Public Health 2013, 13:988

http://www.biomedcentral.com/1471-2458/13/988

RESEARCH ARTICLE Open Access

Individual and environmental factors associated

for overweight in urban population of Brazil
Larissa L Mendes1*, Helena Nogueira2,4, Cristina Padez3,4, Maria Ferrao3,4 and Gustavo Velasquez-Melendez5

Abstract
Background: Obesity is a significant global public health problem and the main cause of many chronic diseases in
both developed and developing countries. The increase in obesity in different populations worldwide cannot be
explained solely by metabolic and genetic factors; environmental and social factors also have a strong association
with obesity. Thus, it is believed that the current obesity epidemic is the result of a complex combination of
genetic factors and an obesogenic environment .The purpose of this study was to evaluate individual variables and
variables within the built and social environment for their potential association with overweight and obesity in an
urban Brazilian population.
Methods: Cross-sectional study was carried out in a sample of 3404 adults living in the urban area of the city.
Information from the surveillance system for chronic diseases of Brazilian Ministry of Health was used and individual
data was collected by telephone interviews. The database was geocoded using the Brazilian System of Postal Codes
for participant residences. An updated, existing list based on the current addresses of supermarkets and
hypermarkets in the city was used as an indicator variable of the availability and access to food. Georeferenced
information on parks, public squares, places for practicing physical activity and the population density were also
used to create data on the built environment. To characterize the social environment, we used the health
vulnerability index (HVI) and georeferenced data for homicide locations.
Results: The prevalence was 44% for overweight, poisson regression was used to create the final model. The
environment variables that independently associated with overweight were the highest population density, very
high health vulnerability index and the homicide rate adjusted for individuals variables. The results of the current
study illustrate and confirm some important associations between individual and environmental variables and
overweight in a representative sample of adults in the Brazilian urban context.
Conclusions: The social environment variables relating to the socioeconomic deprivation of the neighborhood and
the built environment variables relating to higher walkability were significantly associated with overweight and
obesity in Belo Horizonte.

Background Briefly, weight gain occurs when the energy balance is

Obesity is a significant global public health problem and positive, that is, when the calories consumed are greater
the main cause of many chronic diseases in both devel- than the calories expended. However, the increase in
oped and developing countries [1-3]. Study conducted in obesity in different populations worldwide cannot be
Brazil showed that both male and female overweight explained solely by metabolic and genetic factors; environ-
individuals account for 50% of the population, while the mental and social factors also have a strong association
prevalence of obesity is 12.5% for men and 16.9% for with obesity [5]. Thus, it is believed that the current obe-
women [4]. sity epidemic is the result of a complex combination of
genetic factors and an obesogenic environment [6,7].
The ecological model proposed by Swinburn et al. [7]
* Correspondence: [email protected] that is used to analyze the causes of obesity considers
1
Departamento de Nutrição, Instituto de Ciências Biológicas, Universidade the environment an important determinant of obesity,
Federal de Juiz de Fora, Juiz de Fora 36036-900, Brazil
Full list of author information is available at the end of the article
one that directly affects the dietary behavior and physical

© 2013 Mendes et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.
Mendes et al. BMC Public Health 2013, 13:988 Page 2 of 7
http://www.biomedcentral.com/1471-2458/13/988

activity patterns of the individual and, therefore, their interviews containing individual weight and height infor-
energy balance. In this context, the environment is mation were used. We also used data from a standard-
divided into two parts, dimension (micro or macro) and ized questionnaire, delivered by phone interview, which
type (physical, economic, political and sociocultural). collected self-referred information regarding sociodemo-
The physical environment, also defined as the built graphic characteristics, eating patterns, weight, height,
environment, is the most studied dimension. It refers to physical activity and health-related variables. Of the
the physical aspects of an environment built or modified original sample, 101 individuals (3.6%) were excluded
by man and favors unhealthy behaviors within the com- due to a body mass index (BMI) of less than 18.5 kg/m2
munity [8]. Thus, the built environment may limit and an additional 125 individuals were excluded due to
healthy eating habits, either through limiting access to errors in georeferencing. The final sample consisted of
fresh fruits and vegetables [9] or by having a high dens- 3,404 individuals.
ity of fast food restaurants and convenience stores [10];
it may also create barriers that limit physical activity Dependent variable
[11-14]. Moreover, the social environment can also affect BMI was considered the dependent variable in this
healthy eating habits. Communities with higher socio- study and was categorized into two groups, normal
economic deprivation have fewer opportunities to access (18 kg/m2 ≤ BMI < 25 kg/m2) and overweight (BMI ≥ 25
healthy food [15,16] or to practice physical activity, kg/m2), according to the values proposed by the World
either for leisure or for travel by foot or bicycle [17,18]. Health Organization [3]. Data on weight and height were
Thus, obesity-related policies and programs should self-reported. These data have been widely used in epi-
encourage individual changes as well as changes to the demiological studies [20] and have been previously
society and the environment in which people live and validated for Brazilian adults. High correlation coefficients
work [19]. were found when the measurements were compared to
In Brazil, studies evaluating the physical and social the corresponding self-referred values and showed good
environments related to obesity are still scarce, making results when analyzed for sensitivity and specificity [21].
it critical to investigate and identify potential obesogenic
environments within Brazilian cities. The purpose of this Independent variables
study was to evaluate individual variables and variables The independent variables used in the study were based
within the built and social environment for their poten- on a review of the literature and were selected from the
tial association with overweight and obesity in an urban VIGITEL database and the available geocoded data. Indi-
Brazilian population. vidual variables were divided into three categories:
sociodemographic, lifestyle and health. Some variables
Methods relating to diet and physical activity were also validated
Study population in previous studies [22,23].
Participants were selected from the Surveillance of Risk
Factors for Chronic Diseases through Telephone Inter- Characterization of geographic data
view (VIGITEL), organized by the Ministry of Health To verify and analyze the spatial distribution of vari-
and conducted annually in all Brazilian state capitals. ables, the VIGITEL database was geocoded using the
For the present study, samples from the years 2008 and Brazilian System of Postal Codes (CEP) for participant
2009 for the city of Belo Horizonte were used. The city residences [24]. To characterize the built and social
of Belo Horizonte is the capital of the Minas Gerais environment, a geocoded base was developed that incor-
State, is located in the southeastern region, has a total porated the individual data for each participant. An
area of 331 Km2, population of 2.365.151 inhabitants updated, existing list based on the current addresses of
and population density of 7.177 habitants Km2 according supermarkets and hypermarkets in the city was used as
to the population census of the year 2010. an indicator variable of the availability and access to food.
Sampling of participants was conducted in two steps: Georeferenced information on parks, public squares and
(1) random selection of households with landlines and places for practicing physical activity and the population
(2) drawing of respondents 18 years of age or older. The density were also used to create data on the built environ-
VIGITEL system seeks to obtain random population ment. These data were provided by the Computer and
samples of adults living in households with a landline Information Company of Belo Horizonte.
from each Brazilian state capital. Sample weights for the To characterize the social environment, we used the
individuals interviewed were later determined to correct health vulnerability index (HVI), which highlights the
potential biases resulting from the non-universal cover- inequalities in the epidemiological profile of different
age of the telephone network. For the present study, social groups. This measurement combines different
4.000 interviews were considered eligible, of which 3,661 socioeconomic and environmental variables in a single
Mendes et al. BMC Public Health 2013, 13:988 Page 3 of 7
http://www.biomedcentral.com/1471-2458/13/988

synthetic indicator. The HVI can be understood as the were fitted to determine the association of overweight
result of combining both socioeconomic and demographic with individual and environmental variables.
dimensions. Census tracts and their residents are classified
into groups according to social vulnerability. In addition, Ethical issues
georeferenced data for homicide locations were used to The present study was approved by the National Com-
map homicide rates by census tracts. These data were mittee for Ethics in Human Research, part of the Ministry
provided by the Integrated Information Center for Social of Health, and by the Ethics Committee of the Federal
Defense of the Military Police of Minas Gerais. The me- University of Minas Gerais, according to protocol number
dian household income per census tract was obtained 552/08.
from the databases of the Brazilian Institute of Geography
and Statistics was used as a social environment variable. Results
The final sample consisted of 3.404 individuals of whom
49.9% were male and 50.1% were female. The average
Data analysis
age of the participants was 39.7 years, 44,0% of indivi-
Spatial data analysis
duals were overweight (31.60% with BMI ≥ 25 kg/m2;
A spatial analysis using the Geographic Information
12.40% with BMI ≥ 30 kg/m2). Table 1 shows the socio-
System (GIS) was first performed to stratify the geocoded
demographic characteristics of the sample. When com-
data into layers of information, to isolate the spatial rela-
paring individuals with normal BMI (BMI < 25 kg/m2)
tionships and to develop data maps. The spatial informa-
to overweight individuals (BMI ≥ 25 kg/m2), significant
tion was then crosschecked and translated into variables
differences were observed in sex, age distribution, mari-
to be analyzed. These procedures were performed using
tal status and education.
the MapInfo program version 10.5. The information for
Table 2 presents the raw association of variables re-
the exact location (latitude and longitude) based on the
lated to the built and social environment with over-
census tract was used to analyze the spatial distribution of
weight. For the social environment variables, a lower
cases. At this level, it is appropriate to consider the impact
prevalence of overweight and obesity was found in
of social relationships, characteristics of the built environ-
the fourth quartile of population density while a
ment and services and amenities on the health of indivi-
higher prevalence was found in the higher HVI category
duals [25,26].
(p < 0.05).
We used an adjusted Poisson regression model in a
Statistical analysis single level, as shown in Table 3.
The survey module in the STATA 12 program was used In Model 1, the environment variables that inde-
for the analysis of individual and environmental data. pendently associated with overweight were the highest
Various aspects of the complex sampling design are population density (PR = 0.86) and the very high HVI
considered in the analysis as post-stratification weights (PR = 1.32) adjusted for sex and habit of watching TV
and are assigned to the interviewed participants to allow every day of the week. In Model 2, the environment
for the partial correction of bias arising from the non- variables that independently associated with overweight
universal coverage of the phone network [27]. Preva- the very high HVI (PR = 1.37) and the homicide rate
lence ratios (PRs) were used as a measure of association (RP= 1.45) adjusted for sex, age and the commute to
in the bivariate analysis; the ratios were calculated using work by food and bicycle. We tested the interactions
Poisson regression analysis with robust variance because between the variables in the final model, but they were
this was a cross-sectional design in which the searched not significant.
outcome was very frequent [28]. Statistical differences
were evaluated according to the likelihood ratio (LR). Discussion and conclusions
Initially, the sample description was presented according The results of the current study illustrate and confirm
to the participants’ BMI in addition to the bivariate some important associations between individual and
associations with the variables relating to the built and environmental variables and overweight in a representa-
social environments. The bivariate analysis was initially tive sample of adults in the Brazilian urban context. The
performed to identify which explanatory variables had relationship between the environment (built and social)
the greatest influence on variations in overweight. Since and obesity has long been studied in developed countries
the respondents were not sampled by a multi-stage [7,29,30], but in Brazil, this type of research is relatively
sampling method (i.e., they were not geographically recent [31,32].
clustered), we report the results estimated by single level In this study, living in neighborhoods with a high popu-
analysis (we confirmed that using multilevel modelling did lation density was inversely associated with a prevalence of
not alter our findings). Two multiple regression models overweight. According to Frank et al. [33], neighborhoods
Mendes et al. BMC Public Health 2013, 13:988 Page 4 of 7
http://www.biomedcentral.com/1471-2458/13/988

Table 1 Sociodemographic characteristics for all participants by BMI

Variables Total BMI < 25 kg/m2 BMI ≥ 25 kg/m2
% 95 CI% % 95 CI% % 95 CI%
Sex*
Males 50.10 47.53–52.74 46.90 43.14–50.64 53.10 49.40–56.90
Females 49.90 47.26–52.47 41.10 37.50–44.80 58.90 55.20–62.50
Age (years)*
18–24 19.40 16.80–22.50 22.50 14.50–33.20 77.50 66.8–85.50
25–34 23.50 21.30–26.00 41.20 35.40–47.30 58.80 52.70–64.60
35–44 22.80 21.00–24.70 48.10 43.90–52.40 51.90 47.60–52.10
45–54 15.70 14.20–17.00 53.20 48.70–57.60 46.80 42.40–51.30
55–64 10.00 9.00–11.20 60.70 55.50–65.60 39.30 34.40–44.50
≥65 8.60 7.60–9.50 53.00 47.70–58.20 47.00 41.80–53.00
Skin Color
White 34.30 32.15–36.53 42.40 39.17–45 57.60 54.40–60.80
Non White 65.70 63.47–67.85 44.90 41.30–48.50 55.10 51.50–58.70
Marital Status*
Single 43.00 40.20–45.70 34.40 29.60–39.70 65.60 60.40–70.40
Married 47.60 45.10–50.20 50.50 47.60–53.40 49.50 46.60–52.40
Others(seperated/divorced/widowed) 9.40 8.30–10.70 54.60 48.40–60.60 45.40 39.40–51.60
Education (years)*
≤4 17.00 14.60–19.76 50.80 42.20–59.40 49.20 40.60–57.80
5–8 30.80 28.15–33.60 46.50 40.90–52.20 53.50 47.80–59.10
9–11 31.60 29.58–33.66 41.60 38.60–44.70 58.40 55.30–61.40
≥12 20.60 19.07–22.20 37.50 34.20–41.00 62.50 58.90–65.90
Notes: aPercentages weighted to adjust the sociodemographic distribution of the VIGITEL sample to the distribution of the population 18 years and older from the
city according to the 2000 Census. *p < 0.05 (Chi-square). N = 3.405.

with high population density often exhibit greater mixed that communities with a high health vulnerability index
land use and greater street connectivity; consequently, they (HVI) presented the highest prevalence of overweight.
promote higher levels of foot traffic (walkability), which The relationship between higher rates of overweight and
may be related to lower BMI. obesity in populations with poor socioeconomic condi-
Identical results were found in a study conducted in tions is well established in developed and many develop-
New York. People who lived in census tracts with higher ing countries [35,36]. Furthermore, studies show that
population density had significantly lower BMIs than individuals living in socially disadvantaged neighbor-
those who lived in areas with low population density. hoods, which tend to be characterized by insecure and
This study also found an inverse association between dangerous environments, present a higher risk of having
BMI and other variables of the built environment such unhealthy behaviors [37-40] with respect to diet [41,42]
as mixed land use, the number of bus stops and subway or physical activity [43-46].
stations and the number of street intersections [25]. A longitudinal study conducted in England found that
Another study using data from the Behavioral Risk women living in economically disadvantaged neighbor-
Factor Surveillance System (BRFSS) in Massachusetts hoods had higher BMI and greater weight gain compared
found that after adjusting by individual variables, popu- to women in more affluent neighborhoods. The study also
lation density was inversely associated with obesity. This found that over a period of ten years, women from disad-
study suggests that the highest population density was vantaged neighborhoods gained one kilogram more than
associated with greater walkability, indicating that cars women from neighborhoods that were more affluent. No
were used less frequently, thus increasing physical acti- associations were found for men between socioeconomic
vity [34]. status of the neighborhood and BMI [47].
When exploring the contributions of the social envir- Identical results were found in other studies. One
onment to the prevalence of overweight, it was found study conducted in the United States revealed that
Mendes et al. BMC Public Health 2013, 13:988 Page 5 of 7
http://www.biomedcentral.com/1471-2458/13/988

Table 2 Prevalence ratios and 95% confidence intervals adults who lived in neighborhoods with low socioeco-
for overweight/obesity and variables for the built and nomic conditions presented mean BMI values that were
social environment in the census tracts higher than the BMI values of adults living in neighbor-
Environment variables % PR PR 95% CI hoods with better socioeconomic conditions [16]. An-
Built environment other study conducted in the Netherlands found that the
Presence of supermarkets odds ratio for the occurrence of overweight significantly
No 44.10 1.00 -
increases with deprivation in the neighborhood [46].
A number of findings suggest that the socioeconomic
Yes 42.60 0.97 0.78–1.20
environment of the neighborhood is related to the preva-
Presence of fruit and lence of overweight and obesity through mediators related
vegetables markets
to the built environment, such as access to healthy [48,49]
No 44.50 1.00 -
and unhealthy food stores [15,50]. Thus, economically
Yes 34.70 0.78 0.57–1.06 favored neighborhoods tend to have a higher density and
Parks/public squares/places a greater number of supermarkets and grocery stores that
for practicing physical activity are associated with a greater consumption of healthy
No 44.00 1.00 - foods and lower rates of overweight and obesity [51].
Yes 43.90 0.99 0.72–1.37 In the urban context of Brazil, only one study, conduc-
Population density ted in São Paulo, showed that there was a higher density
1st 49.40 1.00 -
of markets specializing in fruits and vegetables in the city's
wealthiest areas. However, no associations between the
2nd /3rd/ 4th quartil 42.30 0.85 0.74–0.98
availability of healthy foods in the neighborhood and over-
Social environment weight were observed [31]. Similarly, the present study
HVI showed no associations with overweight when the contri-
Low 43.70 1.00 - butions of the built environment based on the availability
Medium 43.80 1.01 0.86–1.16 of healthy foods in the neighborhood were investigated.
High 42.80 0.98 0.82–1.16
However, the food environment (food availability and
access) alone does not explain the association between
Very High* 58.00 1.33 1.04–1.69
neighborhood deprivation and obesity. Other aspects of
Homicide rate (100s)* 46.10 1.19 0.85–1.65 the built environment related to physical inactivity and/
Household income (real) or physical activity may contribute this association. Sev-
1st quartile 44.60 1.00 - eral studies have found that economically disadvantaged
2nd /3rd/4th quartile 43.80 0.98 0.86–1.11 neighborhoods are characterized by offering fewer op-
Notes: HVI: Health Vulnerability Index. *p ≤ 0.05. portunities for physically activity [52-54].
Also in relation to the social environment, an asso-
ciation was found between the highest homicide rate in
Table 3 Final Poisson regression model with
the census tract and the highest prevalence of over-
overweighthobesity as a response variable
weight. Previous findings suggest that higher rates of
Variables Model 1* Model 2 **
obesity are found in neighborhoods characterized by so-
PR PR 95% CI PR 95% CI cial disorder [42,55]. This association may be mediated
Built environment by restrictions on physical activity as lower rates of walk-
Population density - - ing and other physical activities or outdoors due to fear
1st 1 - [44,56,57]. A study with adults found that the increased
2nd /3rd/ 4th quartil 0.86 0.75–0.98 perception of safety for the crime was significantly asso-
ciated with lower BMI values. The authors suggest that
Social environment
the use of objective measures of security, such as homi-
HVI
cide rates and crimes in the neighborhood, may be more
Low 1 - 1 - related to BMI than measurements based on the per-
Medium 1.02 0.88–1.17 1.07 0.92–1.24 ception of individuals [58]. Another study found similar
Hight 0.99 0.85–1.16 1.02 0.86–1.22 results for the perception of security, since individuals
Very hight 1.32 1.04–1.67 1.37 1.05–1.80 who perceived their neighborhoods as unsafe had higher
BMI than those who perceived the neighborhood as safe.
Homicide rate (1000s) - - 1.45 1.02–2.05
These associations remained after adjustment for indivi-
*Model 1 adjusted for sex.
**Models 2 adjusted for sex, age and commute to work by food or bicycle.
dual, sociodemographic variables and the socioeconomic
Notes: HVI: Health Vulnerability Index. status of the neighborhood [40].
Mendes et al. BMC Public Health 2013, 13:988 Page 6 of 7
http://www.biomedcentral.com/1471-2458/13/988

The present study does have certain limitations. First, References

despite the use of a causal model, the cross-sectional 1. Popkin BM, Doak CM: The obesity epidemic is a worldwide phenomenon.
Nutr Rev 1998, 56:106–114.
design does not allow us to draw such conclusions. A 2. Ogden CL, Carroll MD, McDowell MA, Flegal KM: Obesity among adults in
second limitation is inherent in the proposed methodology the United States-no statistically significant chance since 2003–2004.
because it uses health surveillance systems. Although it is NCHS Data Brief 2007, 1:1–8.
3. World Health Organization: Obesity: preventing and managing the global
a practical methodology, it depends on self-referred mea- epidemic. Report of a WHO consultation, Volume 894. Geneva, Switzerland:
surements reported by phone, which may include po- WHO Tech Rep Ser; 2000:1–253.
tential biases depending on the variable being assessed. 4. INSTITUTO BRASILEIRO DE GEOGRAFIA E ESTATÍSTICA: Pesquisa de
orçamentos familiares 2008–2009: análise da disponibilidade domiciliar de
Third, we used secondary data from governmental and alimentos e do estado nutricional no Brasil. Rio de Janeiro: IBGE; 2010:38.
commercial sources to describe the environmental charac- 5. Lahti-Koski M, Vartiainen E, Männistö S, Pietinen P: Age, education and
teristics (built and social), but these may also be subject to occupation as determinants of trends in body mass index in Finland
from 1982 to 1997. Int J Obes Relat Metab Disord 2000, 24:1669–1676.
inaccuracies. Other important limitation of this study is 6. Loos RJ, Bouchard C: Obesity — is it a genetic disorder? J Intern Med 2003,
the fact of the walkability was not measured and this is an 254:401–425.
important aspect of the obesogenic environments. Finally, 7. Swinburn B, Egger G, Raza F: Dissecting obesogenic environments: the
development and application of a framework for identifying and
we recognize that there is a limitation with respect to prioritizing environmental interventions for obesity. Prev Med 1999,
sample selection, as only people living in households with 29:563–570.
a landline participated in this study. However, sample 8. Booth KM, Pinkston MM, Poston WS: Obesity and the built environment.
J Am Diet Assoc 2005, 105(5):110–117.
weights were taken into consideration when analyzing the
9. Morland K, Diez Roux AV, Wing S: Supermarkets, other food stores, and
data to adjust for the sociodemographic composition of obesity: the atherosclerosis risk in communities study. Prev Med 2006,
the population. 30:333–339.
In conclusion, the social environment variables relating 10. Timperio A, Ball K, Roberts R, Campbell K, Andrianopoulos N, Crawford D:
Children's fruit and vegetable intake: associations with the
to the socioeconomic deprivation of the neighborhood neighbourhood food environment. Prev Med 2008, 46:331–335.
and the built environment variables relating to higher 11. Feng J, Glass TA, Curriero FC, Stewart WF, Schwartz BS: The built
walkability were significantly associated with overweight environment and obesity: a systematic review of the epidemiologic
evidence. Health Place 2010, 16:175–190.
in Belo Horizonte, Brazil. These initial findings require 12. McCormack G, Giles-Corti B, Lange A, Smith T, Martin K, Pikora TJ:
further investigation to account for the realities of other An update of recent evidence of the relationship between objective and
Brazilian cities. The inclusion of other cities in future self-report measures of the physical environment and physical activity
behaviours. J Sci Med Sport 2004, 7:81–92.
research will facilitate the understanding of the role of 13. Zenk SN, Schulz AJ, Hollis-Neely T, Campbell RT, Holmes N, Watkins G:
environments (built and social) in the current obesity Fruit and vegetable intake in African Americans: income and store
epidemic, thus allowing for the development of useful characteristics. Am J Prev Med 2005, 29:1–9.
14. Saelens BE, Sallis JF, Frank LD: Environmental correlates of walking and
tools able to generate effective strategies for the preven- cycling: Findings from the transportation, urban design, and planning
tion of adult obesity in the Brazilian context. literatures. Ann Behav Med 2003, 25:80–91.
15. Inagami S, Cohen DA, Finch BK, Asch SM: You are where you shop: grocery
Competing interests store locations, weight, and neighborhoods. Ann Behav Med 2006, 3:10–17.
The authors declare that they have no competing interests. 16. Wang MC, Kim S, Gonzalez AA, MacLeod KE, Winkleby MA: Socioeconomic
and food-related physical characteristics of the neighborhood
environment are associated with body mass index. J Epidemiol
Authors’ contributions Community Health 2007, 6:491–498.
GVM and LM assisted with the study coordination, data management, 17. Morland KB, Evenson KR: Obesity prevalence and the local food
carried out the statistical analysis and drafted the manuscript. CP, HN and MF environment. Health Place 2009, 15:491–495.
helped with the statistical analysis and helped to draft and revise the 18. Bodor JN, Rice JC, Farley TA, Swalm CM, Rose D: The association between
manuscript. All authors read and approved the final manuscript. obesity and urban food environments. J Urban Health 2010, 87:771–781.
19. World Health Organization: Diet, nutrition and the prevention of chronic
Acknowledgments diseases. Technical report series 916. Geneva: Joint FAO/WHO Expert
This study was funded by Brazilian Ministry of Health (number 134/209). G. Consultation; 2003.
Velásquez- Meléndez is researcher from the National Council for Scientific 20. Gorber SC, Tremblay M, Moher D, Gorber BA: Comparison of direct vs.
and Technological Development (CNPQ) and the Fundação de Amparo à self-report measures for assessing height, weight and body mass index:
Pesquisa do Estado de Minas Gerais (FAPEMIG). a systematic review. Obes Rev 2007, 8:307–326.
21. Lucca A, Moura EC: Validity and reliability of self-reported weight, height
Author details and body mass index from telephone interviews. Cad Saude Publica 2010,
1 26:110–122.
Departamento de Nutrição, Instituto de Ciências Biológicas, Universidade
Federal de Juiz de Fora, Juiz de Fora 36036-900, Brazil. 2Department of 22. Monteiro CA, Florindo AA, Claro RM, Moura EC: Validity of indicators of
Geography, University of Coimbra, Largo da Porta Férrea, Coimbra 3004-530, physical activity and sedentariness obtained by telephone survey.
Portugal. 3Department of Life Sciences, University of Coimbra, Coimbra Rev Saude Publica 2008, 2:575–581.
3004-530, Portugal. 4Research Centre for Anthropology and Health, University 23. Monteiro CA, Moura EC, Jaime PC, Claro RM: Validity of food and beverage
of Coimbra, Coimbra 3004-530, Portugal. 5Departamento de Enfermagem intake obtained by telephone survey. Rev Saude Publica 2008, 4:582–589.
Materno-infantil em Saúde Pública, Escola de Enfermagem, Universidade 24. Cayo MR, Talbot TO: Positional error in automated geocoding of
Federal de Minas Gerais, Belo Horizonte 30130-100, Brazil. residential addresses. Int J Health Geogr 2003, 19:1–12.
25. Rundle A, Diez Roux AV, Free LM, Miller D, Neckerman KM, Weiss CC:
Received: 10 July 2012 Accepted: 23 September 2013 The urban built environment and obesity in New York City: a multilevel
Published: 21 October 2013 analysis. Am J Health Promot 2007, 21(Suppl 4):326–334.
Mendes et al. BMC Public Health 2013, 13:988 Page 7 of 7
http://www.biomedcentral.com/1471-2458/13/988

26. Tessler Lindau S, Whitaker RC, McDade TW, Ludwig J, Sanbonmatsu L, 49. Bodor JN, Rose D, Farley TA, Swalm C, Scott SK: Neighbourhood fruit and
Gennetian L, Adam E, Duncan GJ, Katz LF, Kessler RC, Kling JR: vegetable availability and consumption: the role of small food stores in
Neighborhoods, Obesity, and Diabetes — A Randomized Social an urban environment. Public Health Nutr 2007, 11:413–420.
Experiment. N Engl J Med 2011, 365:1509–1519. 50. Gibson D: The neighborhood food environment and adult weight status:
27. Ministério da Saúde: Secretaria de Vigilância em Saúde. Secretaria de Gestão estimates from longitudinal. Am J Public Health 2011, 101:71–78.
Estratégica e Participativa. Vigitel Brasil, 2009. Vigilância de fatores de risco e 51. Pearce J, Hiscock R, Blakely T, Witten K: The contextual effects of
proteção para doenças crônicas por inquérito telefônico. Estimativas sobre neighbourhood access to supermarkets and convenience stores on
frequência e distribuição sócio-demográfica de fatores de risco e proteção para individual fruit and vegetable consumption. J Epidemiol Community
doenças crônicas nas capitais dos 26 Estados brasileiros e no Distrito Federal Health 2008, 62:198–2008.
em 2009. Brasília: Ministério da Saúde; 2010. 52. Li F, Harmer P, Cardinal BJ: Built environment, adiposity and physical
28. Barros AJ, Hirakata VN: Alternatives for logistic regression in cross- activity in adults aged 50–75. Am J Prev Med 2008, 35:38–46.
sectional studies: an empirical comparison of models that directly 53. Santos R, Silva P, Santos P, Ribeiro JC, Mota J: Physical activity and
estimate the prevalence ratio. BMC Med Res Methodol 2003, 3:1–13. perceived environmental attributes in a sample of Portuguese adults:
29. Reidpath DD, Burns C, Garrard J, Mahoney M, Townsend M: An ecological results from the Azorean Physical Activity and Health study. Prev Med
study of the relationship between social and environmental 2008, 47:83–88.
determinants of obesity. Health Place 2002, 8:141–145. 54. Rundle A, Neckerman KM, Freeman L, Lovasi GS, Purciel M, Quinn J,
30. Ewing R, Schmid T, Killingsworth R, Zlot A, Raudenbush S: Relationship Richards C, Sircar N, Weiss C: Neighborhood food environment and
between urban sprawl, physical activity, obesity and morbidity. Am J walkability predict obesity in New York City. Environ Health Perspect 2009,
Health Promot 2003, 18:47–57. 117:442–447.
31. Jaime PC, Duran AC, Sarti FM, Lock K: Investigating environmental 55. Cohen DA: Obesity and the built environment: changes in environmental
determinants of diet, physical activity, and overweight among cues cause energy imbalances. Int J Obes 2008, 32(Suppl 7):137–142.
adults in Sao Paulo, Brazil. J Urban Health, New York 2011, 56. Foster S, Giles-Corti B: The built environment, neighborhood crime and
88:567–581. constrained physical activity: an exploration of inconsistent findings.
32. Sichieri R, Moura EC: A multilevel analysis of variations in body mass index Prev Med 2008, 47:241–251.
among adults, Brazil, 2006. Rev Saude Publica 2009, 43(Suppl 2):90–97. 57. Gomez JE, Johnson BA, Selva M, Sallis JF: Violent crime and outdoor
33. Frank LD, Andresen MA, Schmid TL: Obesity relationships with community physical activity among innercity youth. Prev Med 2004, 39:876–881.
design, physical activity, and time spent in cars. Am J Prev Med 2004, 58. Christian H, Giles-Corti B, Knuiman M, Timperio A, Foster S: The influence of
27:87–96. the built environment, social environment and health behaviors on
34. Lopez RP: Neighborhood risk factors for obesity. Obesity 2007, 15:2111–2119. body mass index: results from RESIDE. Prev Med 2011, 53:57–60.
35. Monteiro CA, Moura EC, Conde WL, Popkin B: Socioeconomic status and
obesity in adult populations of developing countries: a review. Bull World doi:10.1186/1471-2458-13-988
Health Organ 2004, 82:940–946. Cite this article as: Mendes et al.: Individual and environmental factors
36. Monteiro CA, Conde WL, Popkin BM: Income-specific trends in obesity in associated for overweight in urban population of Brazil. BMC Public
Brazil: 1975–2003. Am J Public Health 2007, 97:1808–1812. Health 2013 13:988.
37. Van Lenthe FJ, Mackenbach JP: Neighbourhood deprivation and
overweight: the GLOBE study. Int J Obes Relat Metab Disord 2002,
26:234–240.
38. McNeill LH, Kreuter MW, Subramanian SV: Social environment and
physical activity: a review of concepts and evidence. Soc Sci Med 2006,
63:1011–1022.
39. Stafford M, Cummins S, Ellaway A, Sacker A, Wiggins RD, Macintyre S:
Pathways to obesity: identifying local, modifiable determinants of
physical activity and diet. Soc Sci Med 2007, 65:1882–1897.
40. Fish JS, Ettner S, Ang A, Brown AF: Association of Perceived
Neighborhood Safety on Body Mass Index. Am J Public Health 2010,
11:2296–2303.
41. Keita A, Casazza K, Thomas O, Fernandez J: Neighborhood perceptions
affect dietary behaviors and diet quality. J Nutr Educ Behav 2011,
43:244–250.
42. Burdette AM, Hill TD: An examination of processes linking perceived
neighborhood disorder and obesity. Soc Sci Med 2008, 67:38–46.
43. Bennett GG, McNeill LH, Wolin KY, Duncan DT, Puleo E, Emmans K: Safe to
walk? Neighborhood safety and physical activity among public housing
residents. PLoS Med 2007, 4:1599–1607.
44. Harrison RA, Gemmell I, Heller RF: The population effect of crime and
neighbourhood on physical activity: an analysis of 15.461 adults.
J Epidemiol Community Health 2007, 61:34–39.
45. Molnar B, Gortmaker S, Bull F, Buka S: Unsafe to play? neighborhod Submit your next manuscript to BioMed Central
disorder and lack os safety predict reduced physical activity among and take full advantage of:
urban children and adolescents. Am J Health Promot 2004, 18:378–386.
46. Kamphuis C, Mackenbach J, Giskes K, Huisman M, Brug J, Van Lenthe F:
• Convenient online submission
Why do poor people perceive poor neighbourhoods? The role of
objective neighbourhood features and psychosocial factors. Health Place • Thorough peer review
2010, 16:744–754. • No space constraints or color figure charges
47. Stafford M, Brunner EJ, Head J, Ross NA: Deprivation and the development
• Immediate publication on acceptance
of obesity a multilevel, longitudinal study in England. Am J Prev Med
2010, 39:130–139. • Inclusion in PubMed, CAS, Scopus and Google Scholar
48. Rose D, Richards R: Food store access and household fruit and vegetable • Research which is freely available for redistribution
use among participants in the US Food Stamp Program. Public Health
Nutr 2004, 7:1081–1088.
Submit your manuscript at
www.biomedcentral.com/submit