Association between obesity and socioeconomic factors and lifestyle

Background/Aim. The prevalence of overweight and obesity is increasing at an alarming rate and it is a manifestation of the epidemics of a sedentary lifestyle and excessive energy intake. The aim of this study was to determine the prevalence of overweight and obesity in the population of the Province of Vojvodina, Serbia, and to examine the association between obesity and socioeconomic and lifestyle factors. Methods. A cross-sectional study conducted in the Province of Vojvodina in 2006 involved 3 854 participants aged 20 years and over (1 831 men and 2 023 women). The study was a countinuation of the baseline study conducted in 2000 (n = 2 840, 1 255 men and 1 585 women). The main outcome measures were overweight and obesity (Body Mass Index – BMI ≥ 25 kg/m2), sociodemographic factors, including nutrition habits – having breakfast everyday and television watching frequency. Results. The prevalence of overweight and obesity in both sexes in 2006 was 57.4% (35.7% were overweight and 21.7% obese). The prevalence of overweight was higher in men (41.1%) than in women (30.9%) (p < 0.001) while obesity was higher in women (23.1%) as compared to men (20.2%) (p = 0.035). For both sexes, overweight rates were highest at the age 60–69 (men 44.8% and women 39.1%) while obesity rates were peaked to men aged 50–59 (25.1%) and women aged 60–69 years (37.8%). Increasing ageing, males, rural population, single examinees, lower educational level, improved income, examinees that never/sometimes have breakfast and frequently watch TV were associated with obesity. Conclusions. The population of Vojvodina, with 23.1% obese women and 20.2% obese men is one of severely affected European populations. High prevalence of obesity requires urgent public health action. Healthy lifestyle, balanced nutrition with low energy intake and increased physical activity have to be promoted within a prevention strategy and obesity management.


Introduction
Excess weight is a significant health problem worldwide.Globally, there are more than 1 billion overweight adults, at least 300 million obese 1 .Recent estimations have showed that the prevalence of overweight and obesity increases at alarming rates, both in developed and developing countries 2,3 .
Overweight and obesity are major risk for serious dietrelated chronic diseases, including diabetes type 2, cardiovascular diseases, cerebro-vascular diseases, some cancers, psychosocial disturbances, osteoarthritis and increased risk ranged from premature death to serious chronic conditions that reduce the overall quality of life 4,5 .The risk of associated disease increases with body mass index (BMI).There is also an evidence that obesity decreases longevity 6,7 .
Excess body weight poses one of the most serious public health challenges of the 21st century for the World Health Organization (WHO) -European Region.It is responsible for more than 1 million deaths and 12 million disabilityadjusted life-years (DALYs) in the Region every year 4 .The body mass index is the most commonly used indicator of overweight and obesity as it provides useful estimations of weight -for -height at the population level.In developed countries, obesity is high and has a tendency of growing.In the United States the prevalence rate of obesity is 20% in men and 25% in women, in Canada 35% for men and 27% for women and in Europe ranges from 10-20% among men and 10-25% among women [8][9][10] .
The prevalence of overweight and obesity in Province of Vojvodina (northern part of the Republic of Serbia with 2 031 992 inhabitants) in the 70-ties of the 20 th century was 51.6% 11 .A study carried out in 2000 reported 58.5% adults with overweight and obesity, 23% of them were obese 12 .
This article provides the most recent prevalence estimates of overweight and obesity based on the representative sample of the Province of Vojvodina in 2006 and examines the trend in overweight and obesity compared with 2000.This article discusses adults aged 20 years and over classified as overweight or obese according to their BMI, based on measured body height and body weight 13,14 .Impact of education level, gender and age, income, type of settlements, nutrition habits -everyday having breakfast and watching television was explored.

Methods
The Ministry of Health of the Republic of Serbia carried out "The 2006 National Survey for the population of Serbia" (without data on Kosovo and Metohija).The study was conducted as a continuation of the baseline study 2000 in according with its methodology.Target population was an age group of 20 and above who were members of household (persons from particular institutions were not included -old people houses, social institutions, prisons, psychiatric institutions).Sample frame encompassess all households within the census for 2002 in Serbia.
A stratified two-stage sample of the population was used.In Serbia, six geographic areas were identified as main stratum in the sample -Vojvodina, Belgrade, Central, West, Eastern and South East Serbia.The further stratum classification was to the urban and rural areas.Two-stage sampling was conducted so that in the first stage census circles selected by the Propability Proportional Sampling were distinguished.In the second stage households were selected (10 households and 3 reserves from the list) by the Simple Random Sampling Without Replacement.
A sample was chosen as to reliable statistical evaluation for all characters whose frequency of appearance was above 5%.For the total population, relative error of 5% of evaluated proportions was anticipated.Concerning requirements for precise evaluation and the level of the obtained evaluation relialibility, the number of examinees that enable sufficient size of sample according to stratums was selected.The poll included 6 156 households in Serbia (14 522 adult persons aged 20 and above).The survey conducted in the Province of Vojvodina covered 1 810 households (1 060 in town and 750 in other settlements) when 3 854 adults were examined.
Measurement of body mass was performed using a decimal scale in kilograms, after taking off shoes and excess clothing.The weight was measured to the nearest 100 g.Height was measured to the nearest 0.1 cm, without shoes, using a specially mounted metal centimeter ruler.Body mass index was calculated as weight in kilograms, divided by height in m 2 .Those with a BMI 25.0-29.9kg/m 2 were classified as overweight, while those with BMI ≥ 30.0 kg/m 2 were classified as obese 13,14 .Obesity was divided into three categories, with successive values representing escalating health risk 2 .People in class I (BMI 30.0 to 34.9 kg/m 2 ) had a high risk of developing health problems.For those in Class II (BMI 35.0 to 39.9 kg/m 2 ), the risk was very high, and in class III (BMI 40 kg/m 2 or more), extremely high 15 .
Data on age and sex, education, marital status, types of settlements, household income, nutrition habits -regular having breakfast and watching TV were obtained by questionnaire.
Educational level was categorized as primary (elementary complete and incomplete school), secondary (secondary completed) and postsecondary (any college or university), marital status as being married or single while types of settlements as town and other settlements.Household income according to the wealth index was divided into five levelsthe poorest (first level), poorer (second level), middle class (third level), richer (fourth level) and the richest (fifth level).Waching TV and regular having breakfast were estimated for the previous week.
The obtained data were statistically processed using the SPSS version 14.0.Statistical analysis was done using descriptive and inferential statistics.The results are given as mean ± SD and proportion.Differences in frequency were tested by χ 2 test.Differences between two sample means were tested by applying Student's t-test (for both p values < 0.05 were considered to be statistically significant).All the reported p values were two-tailed.In construction of a model to connect obesity and potential risk factors, multivariate logistic regression was used, where we analyzed the effect of an independent variable if it is controlled for the other factors.Importance of each variable included in the model was verified by examination of the Wald statistics.Any variable whose univariable test (with dependent variable obesity) had p < 0.10 was a candidate for the multivariable model along with all variables of known clinical importance: sex, age, type of settlements, married status, education level, income, regular intake of breakfast, lunch, supper and snacks, physical activity in spare time, frequency of physical activity, watching TV (full model).Variables were selected by backward elimination (probability for removal was 0.25).Model fit was analyzed by using a classification table (cutoff 0.50), for calculation sensitivity and specificity.

Results
In 2006 a research carried out in Vojvodina estimated that more than one third of population aged 20 and more were overweight, while more than one fifth of population was obese.
The prevalence rate for overweight was 35.7% (95% CI: 34.1-37.2) and 21.7% for obesity (95% CI: 20.4-23.0).Rates of overweight and obesity vary depending on age and sex.The prevalence of overweight in men (41.1%) was higher than in women (30.9%) (p < 0.001) while the prevalence of obesity was 23.1% in women and 20.2% in men (p = 0.035).The prevalence of overweight and obesity defined by BMI showed an increase up to the age 69, and thereafter the prevalence decreased.For both sexes, overweight rates were lowest at the age 20-29 (29.6% -men and 11.5% -women) and highest at the age 60-69 (44.8% men and 39.1% women) while obesity rates were lowest at age 20-29 years (9.7% men and 7.7% women) and peaked to men aged 50-59 (25.1%) and women aged 60-69 years (37.8%)(Table 1, Figure 1).Statistically important decrease of the prevalence of overweight occurred in 2006 compared with 2000 among women aged 20-29 (p = 0.027) and obesity in women aged 50-59 (p = 0.001) and 70 and over (p = 0.040).
The association between obesity (including overweight BMI ≥ 25 kg/m 2 ) and potential risk factors is shown in Table 3.
The results of logistic regression, which was used to evaluate factors that were associated with odds of being overweight or obese recognized by indication that increasing ageing influences increased incidence of overweight and obesity.Males had 50.9% greater odds for obesity regarding to females.Rural examinees had 29.6% greater odds for obesity regarding those from urban settlements.Marital status had a significant association with obesity, so single examinees (non-married, divorced, widowhood) had 26.2% less odds for obesity regarding to married examinees.Increased educational level influenced decreased odds for obesity, so examinees with secondary school had 7.6% less odds for obesity regarding those with lower educational level (without education, incomplete or finished elementary school), i.e. 32.1% less odds for obesity had examinees with finished college or university.Improved income was followed by the increased odds for obesity.The richest examinees had nearly two times greater odds for obesity regarding the poorest population.Examinees that never/sometimes having breakfast had 10.0% greater odds for obesity regarding examinees have regular breakfasts.More frequent watching TV increased odds for obesity, so examinees who frequently watched TV had 76.0% greater odds for obesity as compared to those who did not.Sensitivity for the Logistic regression model (Table 3) for prediction obesity was 81.8% and specificity 45.8%.Overall rate of correct classification was 66.5%.

Discussion
National health survey of Serbia 2006 indicated that 54.5% of adult population of the Republic of Serbia was overweight, subdivided as 18.3% obese and 36.2%preobese.The highest percentage of obese people was found in population of Vojvodina (21.7%) and eastern Serbia (21.4%) 16 .
The surveys carried out in the Province of Vojvodina (2000 and 2006 year) indicated that a higher proportion of men were in the overweight category than women, while proportion of women were more present in the obese category.Compared to 2000 decrease of the prevalence of overweight occurred in 2006 among women aged 20-29 (p = 0.027) while obesity decreased in women aged 50-59 (p = 0.001) and 70 and over (p = 0.040).In 2000, higher prevalence of obesity was in women aged 50 years and over.
In 2006, the prevalence of obesity was lower in young women than in young men, except the age of 30-39 when obesity prevalence was higher in men than in women (p = 0.001) while in the age 60 and over (p < 0.001) prevalence of obesity was higher in women.The high prevalence of overweight and obesity in Vojvodina is still lower than the reported rate for 2003-2004 from the United States (66.3%), but is very similar to the reported data from an Australian study in 1999-2000 AusDiab (59.8%) and Canada (51.3%) [17][18][19] .Increasing levels of adult overweight and obesity can be found through Europe, although there are variations in prevalence.In parts of Europe the combination of overweight and obesity in men exceeds even 67% prevalence found in the United States most recent measured survey.Finland, Germany, Greece, Cyprus, the Czech Republic, Slovakia and Malta all have overweight rates which surpass that of the USA.Discussing on obesity alone, at least nine European countries have male obesity rates above 20%, including Greece and Cyprus reaching 27%.For women obesity levels vary from 10 to 26% in the Czech Republic.In at least seven countries one in five women is obese 20 .Population of Vojvodina, with 23.1% obese women and 20.2% obese men is one of severely affected European populations.
The results of our research indicate that there is a strong correlation between sex and age pattern of BMI.The prevalence of overweight and obesity are lower in women than in men and are increasing with ageing.Similar results were also found in some other research 18 .
We found a consistent relationship between lower educational level and overweight and obesity.These findings are supported by other studies 21 .Strong association was found between BMI and marital status.We confirmed results from other studies 22 .Marriage was associated with a weight gain while unmarried status, divorce or widowhood with a weight loss.One of the examined factors was type of settlements.People from villages had a higher chance for overweight and obesity than those from towns.On the other hand, there was no difference in prevalence of overweight and obesity in rural and urban areas of ten European countries 23 .Increase in income was followed with a chance for overweight and obesity.The richest people showed two times higher odds for overweight and obesity compared with the poorest people.Similar results were obtained in other studies, too 24 .People who never/sometimes had breakfast had higher odds for overweight and obesity compared with those who had regular breakfast.This was found in many studies [25][26][27] .Regular television watching showed a significant correlation with overweight and obesity.Similar results are found in some other surveys 18 .

Conclusion
The existing situation concerning a high prevalence of obesity in the Province of Vojvodina requires urgent public health action.It is essential to develop preventive public health strategies which would affect the whole society.Healthy lifestyle, combined with balanced nutrition with low energy intake and increased physical activity have be promoted, but it is not a responsibility of individuals only.Health services, community, food industry, mass media etc, have a crucial role in modification of environment in a way that it will be a less supportive for increase of body weight.Strategy of prevention and management of obesity should be in accordance with the existing public health policy and program for prevention of non-communicable diseases.

Women Men Fig. 1 -Mean Body Mass Index (BMI) by sex and age groups (bars = 95% Cis)
There were no statistically significant differences in overweight of adult population in the Province of Vojvodina in 2000 (35.2%) and in 2006 (35.7%), as well as obesity (23.2% in 2000 to 21.7% in 2006).An average BMI of adults was 26.66 ± 5.19 kg/m 2 in 2000 and 26.45 ± 4.96 kg/m 2 in 2006.Among obese people BMI varied greatly.In 2006, 15.6% of population had a BMI class I, 4.6% were class II, while 1.5% class III.There were no statistically significant differences in those three categories in 2000 and 2006 (Table 2).