Application of two-level ordinal logistic regression models in determining factors affecting the economic burden of type 2 diabetes in Iran

Hadipour, Maryam; Jafaraghaei, Razieh; Yadgarfar, Qasem; Faizi, Avat; Abolhasani, Farid

Application of two-level ordinal logistic regression models in determining factors affecting the economic burden of type 2 diabetes in Iran

Document Type : Original Article

Authors

Maryam Hadipour

Razieh Jafaraghaei

Qasem Yadgarfar

Avat Faizi

Farid Abolhasani

University

Abstract

In recent years, multilevel regression models have been developed significantly in various sciences, including medicine, psychology, economics, and other sciences. These models are used for data with a hierarchical structure where each lower level is nested in higher levels. To model this type of data with discrete response variables, generalized regression models are used. In this article, the two-level ordinal logistic regression model is first introduced and different methods for estimating the model parameters are described. Then, the application of this model is discussed using data from the Diabetes Cost Estimation Study in Iran, which was collected by the Endocrinology and Metabolism Center and Tehran University of Medical Sciences in 2006, in determining the individual and environmental factors affecting the economic burden of type 2 diabetes on diabetic patients.

Keywords

Economic burden of type 2 diabetes

two-level ordinal logistic regression

marginal and predictive pseudo-likelihood methods

Gaussian and Gaussian-likelihood methods