队列研究医学论文中常见统计学问题

Analysis of common statistical problems in medical papers on cohort study

目的了解目前我国医学期刊中采用队列研究设计的文章中统计学方法的应用情况,发现问题并提出相应对策,以提高此类文章的数据处理及撰写水平。方法检索2014~2015年中国知网、万方数据库收录的有关队列研究的论文,对其进行系统的统计学方法分析。结果论文中普遍存在的统计学问题包括χ2检验和Logistic回归的误用、研究对象描述不清楚、结局事件及其判断标准描述不全面等。结论论文作者应充分认识流行病与医学统计学方法在科学研究中的重要性,并具备一定的相关理论知识;期刊编辑部应加强论文的流行病与医学统计方法学的审查工作。

Objective To understand the application of statistical methods in papers adopting cohort study design in medical journals,uncover some important issues,and put forward the corresponding countermeasures,furthermore,to improve the data processing and writing level of such papers.Methods Papers published on cohort study from 2014 to2015 were retrieved in China National Knowledge Infrastructure and Wanfang Database, and were analyzed systematically by statistical methods.Results Common statistical problems found in these papers included the followings： misuses of chi-square test and Logistic regression, no clear description of study subjects, and no comprehensive description of outcomes and its judgment standard.Conclusion The papers′ authors should fully realize the significance of epidemiology and health statistics,and master the relevant theoretical knowledge;the editorial departments should strengthen epidemiological and statistical examination on medical papers.