非线性半相依回归模型在生长曲线研究中的应用  

Seemingly Unrelated Nonlinear Regression Model with Its Application in Growth Curve Analysis

作　　者：赵俊康[1] 梁洪川[1] 王彤[1] 

机构地区：[1]山西医科大学公共卫生学院,030001

出　　处：《中国卫生统计》2012年第3期348-350,共3页Chinese Journal of Health Statistics

基　　金：国家自然科学基金(81072385);全国统计科研计划重点项目(2009LZ033)

摘　　要：目的探讨非线性半相依回归模型在医学研究领域中的应用。方法以毒理学试验中对四氯化碳CCl4肝细胞毒性的研究为例,应用非线性回归模型分析CCl4肝细胞毒性剂量-反应关系及其与时间的关系。结果 CCl4浓度及时间对乳脱酸氢酶渗出率的影响有统计学意义,且CCl4浓度与时间之间存在交互作用。结论当生长曲线资料中变量之间的关系不满足线性或近似线性关系时,采用非线性半相依回归模型的参数估计方法可以更好的解决这类问题。Objective To explore the application of non-line- ar seemingly unrelated regression model in the field of medical research. Methods Taked the CC14 carbon with tetrachloride hepatotoxicity＇s study in toxicology tests for example and use non-linear seemingly unrelat- ed regression model to analysis the dose-response relationship of CC14 car- bon with tetrachloride hepatotoxicity and time. Results The concentra- tion of CC14 and time had influences on the rate of lactate dehydrogenase leakage and there existed the interaction of them. Conclusion The rela- tionship between the variables not satisfied with linear or nearly linear rela- tionship, using nonlinear seemingly unrelated regression model can be a bet- ter solution to such problems.

关 键 词：非线性半相依回归模型 生长曲线 重复测量设计 

分 类 号：R195[医药卫生—卫生统计学] O212.1[医药卫生—卫生事业管理]