模糊分类的Markov模型在慢性肾脏病分期预后研究中的应用  

Application of Fuzzy-Markov model in hierarchical prognosis of chronic kidney disease

作　　者：刘迅[1,2] 凌莉[3] 彭晖[1] 陈珠江[1] 娄探奇[1] 

机构地区：[1]中山大学附属第三医院肾内科,广东广州510630 [2]华南理工大学生物工程学院医学院,广东广州510640 [3]中山大学统计与流行病学系,广东广州510080

出　　处：《第四军医大学学报》2009年第19期2025-2028,共4页Journal of the Fourth Military Medical University

摘　　要：目的：以模糊分类的Markov模型探讨慢性肾脏病（CKD）患者的CKD各期进展情况.方法：对272例CKD随访病例进行回顾性研究,建立基于模糊分类的CKD1期、CKD2期、CKD3期、CKD4期、CKD5期、死亡/终末期肾病期6个状态的Markov模型,探讨患者的分级预后.结果：平均观察年限2.0a.CKD12期模糊转移的概率为15.1%/a,CKD23期模糊转移的概率为14.0%/a,CKD34期模糊转移的概率为15.7%/a,CKD45期模糊转移的概率为16.4%/a,CKD5期死亡/终末期肾病模糊转移的概率为51.5%/a.以模糊分类的Markov模型预期本队列今后50a处于CKD1期的平均时间为12.5a,处于CKD2期的平均时间为7.2a,处于CKD3期的平均时间为4.5a,处于CKD4期的平均时间为2.0a,处于CKD5期的平均时间为0.9a,预期平均寿命或肾存活期为27.1a.结论：应根据CKD不同分期的预后评估患者病情和指导临床治疗.AIM：To evaluate the hierarchical prognosis of chronic kidney disease（CKD）.METHODS：272 outpatient CKD patients were investigated.Fuzzy-Markov model including six states [CKD1 stage,CKD2 stage,CKD3 stage,CKD4 stage,CKD5 stage as well as death/end-stage renal disease（ESRD）stage]was established.RESULTS：The mean following period was 2.0 years.Fuzzy transition rate from CKD1 stage to CKD2 stage,from CKD2 stage to CKD3 stage,from CKD3 stage to CKD4 stage,from CKD4 stage to CKD5 stage and from CKD5 stage to death/ESRD stage were 15.1%/year,14.0%/year,15.7%/year,16.4%/year and 51.5%/year respectively.The fuzzy mean duration in CKD1 stage,CKD2 stage,CKD3 stage,CKD4 stage and CKD5 stage were 12.5 years,7.2 years,4.5 years,2.0 years and 0.9 years respectively.Fuzzy mean renal survival time or dialysis free period was 27.1 years.CONCLUSION：Evaluation of severity and the treatment of CKD patients should be made according to the prognosis of different stage in CKD patients.

关 键 词：模糊 多状态MARKOV模型 慢性肾脏病 预后 分期 

分 类 号：R195.4[医药卫生—卫生统计学]