Classification tree model analysis on related factors of early renal damage in type 2 diabetic patients / 中华肾脏病杂志

ABSTRACT

Objective To analyze the impact factors for early renal damage in type 2 diabetic patients by the classification tree model.Methods A total of 601 patients with type 2 diabetes were enrolled.According to glomerular filtration rates and urine albumin quantification,the patients were divided into type 2 diabetes group (418 cases) and early diabetic renal damage group (183 cases).The clinical data of the patients were recorded to analyze the main influential factors for the microalbuminuria of type 2 diabetic patients using the Exhaustive CHAID classification tree algorithm.Results Six important explanatory variables were screened out by the classification tree model from the 34 candidate variables related to early renal damage,including fibrinogen,history of hypertension,retinopathy,Cys C levels,SBP and peripheral neuropathy.Elevated fibrinogen was the main factor.Conclusion The classification tree model can analyze the major influential factors of early renal damage in type 2 diabetic patients effectively,and it can help develop the prevention and treatment methods.