---

In recent years, multilevel regression models were intensely developed in many fields like medicine, psychology economic and the others. Such models are applicable for hierarchical data that micro levels are nested in macros. For modeling these data, when response is not normality distributed, we use generalized multilevel regression models. In this paper, at first, multilevel ordinal logistic regression models and some estimation methods are explained. So their applications are investigated in the effect of patient’s environment on economic burden of diabetes type 2.