【摘要】 为了探讨及验证Pitman指数作为最适配比观察指标的理论在病例对照研究中的实际应用,我们利用在浙江省嘉善县进行的大肠癌人群序贯筛检数据库作实例验证,对检出的11例大肠癌,在62 667例筛检人群中,随机抽取1:1至1:50的配比作病例对照研究。经过统计学处理,结果显示:随着R值的增大,具有显著性意义的因子检出率及与1:50配比研究结果的一致率随之增高;离差系数变化趋势越来越小;稳定性也相对提高,但当R值大于4时则增加值就较为缓慢。因此笔者认为:结合配比研究的灵敏度及稳定性,当在研究经费并不宽裕,且仅限于对某病作探索性研究时,最适配比以1:4或1:5左右为宜。更多还原

【Abstract】 We applied the data base of a screening program for colorectal cancer in the general population of Jiashan county, Zhejiang province to verify the theory of Pitman Efficiency in 1:R matched case-control study. Forty one detected cases of colorectal cancer were matched with normal controls as 1:R (R from 1 to 50) who were randomly selected from the screened population in cluding 62 667 subjects. Results showed that the number of risk factors with statistical significance and its coincidence rate with results of 1:50 match were increased along with the increasing of R value; meanwhile the trends of standard deviation with 1:R over 1:50 weakened and the stability of results increased accordingly. The change of these values, however, reached their plateau when R>4. The authors suggest that when taking consideration of sensitivity and stability of matched study, 1:4 of 1:5 ratio might be the most suitable case-control match, particularly when the resources are limited or the study has a exploratory nature.更多还原