【摘要】 本文根据175例宫颈癌病例配对资料进行宫颈癌发病因素的研究。为排除混杂因素,先后进行了单因素分析,二因素和三因素分层分析及其Logistic多因素回归分析。结果:宫颈糜烂、不注意月经期卫生、本人结婚次数≥3次三个因素与宫颈的联系强度其差异均有显著意义。条件Logistic回归分析排除了配偶性混乱、未避孕、包皮过长、配偶结婚≥3次等正混杂效应;还排除了本人性混乱的负混杂效应,确定为差异有显著意义。更多还原

【Abstract】 In recent Years,logistic regression model has been used to analyse risk epidemic factors. A casecontrol study of one hundred and seventy five patients with cervical cancer by conditional logistic regression model was performed. To exclude other confusing factors,the data were analysed by single. factor,regulated factors and conitional logistic regression model. It was found that the correlative intensities of three factors, namely, cervical erosion, poor hygiene during menstruation,and three or more marriages with cervical cancer were significantly different by analyses of single factor and conditional logistic regression model. Factors known as positvely con- founding effects,such as husband being sexual promiscuous, no contraceptive measure, redundant pre- puce, huasband having three or more marriages were ruled out by conditional logostic regression mod- el. And sexual promiscuity of the patient herself, known as negatively confounding effect, wsas ruled out as well.更多还原