四分类logistic回归模型对肺癌分组织类型的多因素研究

A Multiple Factors Study of Histologic Lung Cancer in Miners with Polychotomous logistic Model

logistic回归模型已经广泛用于医学研究中,但多分类logistic回归模型应用较为少见。本文建立了四分类logistic回归模型,分析了云南锡矿公司126例病例-对照资料,提示了鳞癌与氡暴露有一定关系,其相对危险度为1.7671, P<0.01;鳞癌发病年龄较其他组织类型肺癌高;吸烟对腺癌的危险性较其他类型低。

The logistic regression is a probability model depended on logistic function. This model has been generaly applied in the medical research. however it is very few report that polychoto- mows logistic regression was used. In this paper, the following investigation was presented, 1. A BASIC compiler for polychotomous logi- stic model has been performed. This program is able to select and pick out automatically the risk factors with computer conveneintly and time saving. 2. Data including 126 cases and controls of lung cancer in Yunnan Tin Company (YTC) in China were analyzed, a quaternary logistic model was established and a satifactory result of the studing cause of lung cancer was obtained. 3. The relation of lung cancer with radon and smoking was invesgated and following con- clution of epidemiologic research was showed. (1)epidermoid cancer, relates closely with Radon exporsure relative risk is 1.767,P<0.01. (2)The age suffered from epidermoid is older than from other histologic lung cancers.(3) Smoking risk leading to adenocarcinomar is less serious than to other cancers.