Using quadratic inference functions to determine the factors associated with obesity: findings from the STEPS Survey in Iran Fereshteh Asgari MD, MPH a , Akbar Biglarian PhD b , Behjat Seifi PhD c , Andisheh Bakhshi PhD Student d , Hamid Heidarian Miri PhD Student e , Enayatollah Bakhshi PhD b, * a Center for Diseases Control, Ministry of Health and Medical Education, Tehran, Iran b Department of Biostatistics, University of Social Welfare and Rehabilitation Sciences, Tehran, Iran c Department of Physiology, Medicine School, Tehran University of Medical Sciences, Tehran, Iran d Department of Mathematics and Statistics, University of Strathclyde, Glasgow, UK e Department of Epidemiology, Shiraz University of Medical Sciences, Shiraz, Iran article info Article history: Received 12 May 2013 Accepted 6 July 2013 Keywords: GEE BMI Obesity abstract Purpose: The last study that had a representative sample of the Iranian population and examined the association between sociodemographic and lifestyle factors with obesity used data from the 1999 to 2000 National Health Survey in Iran. Concern about the increased prevalence of obesity has heightened interest in the association between varieties of these factors with obesity among Iranian households. Methods: Data (20,917 adults) were taken from the 2009 STEPwise approach to Surveillance survey in Iran that was conducted under the supervision of World Health Organization. Using quadratic inference function, we evaluated the relation of age, sex, physical activity, serving of fruit and vegetable, employment status, smoking, and place of residence with obesity. Results: Using 20e30 years as the reference group, the obesity odds ratios (ORs) were 2.48, 3.79, 3.52, and 3.15 for ages 30e40, 40e50, 50e60, and 60þ years, respectively. Obesity OR for female was 2.63, compared with male. Obesity ORs for smokers and ex-smokers were 0.64 and 1.18, respectively, com- pared with nonsmokers. Among Iranian adult, obesity ORs for the moderate and vigorous physical activity were 0.99 and 0.90, respectively, compared with low level. The estimated odds of obesity were 46% higher for rural adults. Obesity ORs for public and private employed were 0.79 and 0.76, respectively, compared with unemployed. Conclusions: Our results may provide better insights of the factors associated with obesity and can be used as a basis to reinforce health programs to prevent obesity in Iran. Ó 2013 Elsevier Inc. All rights reserved. Introduction It is known that there is a global rising obesity epidemic. One- third of U.S. adults were obese in 2009e2010 [1]. In 2010, almost 43 million children (35 million in developing countries and 8 million in developed countries) were estimated to be overweight or obese [2]. Being obese is not only associated with health dis- orders, such as cardiovascular diseases [3], diabetes [4], asthma [5], arthritis [6], chronic pain [7], Alzheimer disease [8], and functional impairment [9], but also with sociodemographic status and lifestyle behaviors [10]. One potential predictor could be age [11e 13]. Studies have also found difference in the prevalence of obesity between the genders [12]. Few studies have reported differences in the body mass index (BMI)-mortality association for men and women [14,15]. Another predictor is socioeconomic status, which has shown an inverse relation with obesity [16]. Nevertheless, these variables have predictive values more in women than in men [17]. Although the association of obesity with smoking has been ana- lyzed in many studies, the findings are not consistent [18e20]. The last study that had a representative sample of the Iranian population and examined the association between sociodemo- graphic and lifestyle factors with obesity, used data from the 1999 to 2000 National Health Survey in Iran. Concern about the increased prevalence of obesity has heightened interest in the association between varieties of these factors with obesity among Iranian households. In this article, we aimed to explore the association between age, sex, physical activity (PA), employment status, * Corresponding author. Department of Biostatistics, University of Social Welfare and Rehabilitation Sciences, Koudakyar Street, Tehran, Iran. Tel./fax: +98 21 22180146. E-mail address: bakhshi@razi.tums.ac.ir (E. Bakhshi). Contents lists available at SciVerse ScienceDirect Annals of Epidemiology journal homepage: www.annalsofepidemiology.org 1047-2797/$ e see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.annepidem.2013.07.006 Annals of Epidemiology 23 (2013) 534e538