Tehran University of Medical Sciences Journal of Biostatistics and Epidemiology 2383-4196 4 3 2018 12 05 119 123 EN Ahmad Basir Biostatistics Department, Faculty of Medical Sciences, Tarbiat Modares University, Tehran, Iran Mohammad Gholami Fesharaki Assistant Professor of Biostatistics, Biostatistics Department, Faculty of Medical Sciences, Tarbiat Modares University, Tehran, Iran Mohsen Rowzati Worksite Follow-Up Unit, Occupational Health Center, Mobarakeh Steel Company, Isfahan, Iran 2018 07 07 Background & Aim: One of the basic assumptions in simple linear regression models is the statistical independence of observations. Sometimes this assumption is not true for study subject and consequently the use of general regression models may not be appropriate. In this case, one of the leading methods is the use of multilevel models. The present study utilizesmultivariate logistic regression model using a multilevel  model to exhibit the chance of having elbow, wrist and knee disorders over the past year based on elbow, wrist and disorders during the past week. Methods & Materials: This study is a cross-sectional study that was carried out from April 2015 to May 2016 in Mobarakeh Steel Company, Isfahan. The study population includes 300 male employees of Mobarakeh Steel Company, with a mean age of 41.40±8.17 years and an average working experience of 16.0±7.66 years. Data were analyzed using SPSS (version 24) and MLwiN software. Results: Based on this study, results obtained from single variable and multivariable regression were different. Conclusion: Based on this study, it can be suggested that multivariable regression cause a better and more accurate deduction compared to single variable method. https://jbe.tums.ac.ir/index.php/jbe/article/view/204 https://jbe.tums.ac.ir/index.php/jbe/article/download/204/166