综合类医学期刊已发表文献统计学报告质量及其影响因素分析  

Analysis of the quality of statistical reporting and its influencing factors in publications from Chinese general medical journals

作　　者：朱珈仪[1] 刘元元[2] 姜恬 ZHU Jia-yi;LIU Yuan-yuan;JIANG Tian(Editorial Department,Key Laboratory of Birth Defects and Related Diseases of Women and Children,Ministry of Education,West China Second University Hospital,Sichuan University,Chengdu,Sichuan 610041,China;不详)

机构地区：[1]四川大学华西第二医院编辑部、出生缺陷与相关妇儿疾病教育部重点实验室,四川成都610041 [2]四川大学华西公共卫生学院/四川大学华西第四医院流行病与卫生统计学系 [3]四川大学期刊社 [4]四川大学学报(医学版)编辑部

出　　处：《现代预防医学》2024年第16期2918-2924,2942,共8页Modern Preventive Medicine

基　　金：中国高校科技期刊研究会2021年医学期刊专项基金(CUJS-YX-2021-2-6)。

摘　　要：目的对综合类医学期刊已发表文献中常见统计学报告质量进行调查,并分析其影响因素。方法随机抽取综合医药卫生类2种Medline期刊和6种中文核心期刊,调查其2010年和2020年发表的所有原始研究文献共计480篇。基于SAMPL指南、ICMJE建议、CONSORT、STROBE及相关规范,采用自制调查表对文献的统计学报告质量进行评价,并分析其影响因素。结果被调查的综合类医学期刊文献统计学报告质量的总分值为40.0±8.6,统计设计分值为43.2±6.8,统计描述分值为67.0±24.3,t检验分值为23.6±17.5,方差分析分值为26.8±17.2,χ^(2)检验分值为24.2±13.7,秩和检验分值为45.5±18.4,相关分析分值为41.9±13.2,回归分析分值为40.5±14.8,生存分析分值为45.6±14.9。单因素分析显示,2020年发表的文献统计学报告质量总分值高于2010年(35.8±7.4 vs 44.3±7.5,t=-12.510,P<0.001);多水平模型结果显示,期刊收录体系、研究领域和出版年份是影响总分的主要因素,与Medline收录期刊相比,中文核心期刊总分值较低。与人为研究对象的文献相比,以其他为研究对象的总分值较低。与2010年文献相比,2020年文献的总分值较高。结论国内综合类医学期刊文献常见统计学方法报告质量在近十年有一定提高,但仍然有较大改进空间,建议加强相关报告规范推广力度。Objective To investigate the quality of statistical reporting in publications from general medical journals and analyze the influencing factors.Methods A random sample of 480 original research articles published in 2010 and 2020 was selected from two Medline journals and six Chinese core journals in the field of medicine and health.The quality of statistical reporting was evaluated using a self-developed checklist based on the SAMPL guidelines,ICMJE recommendations,CONSORT,STROBE,and other relevant standards,followed by an analysis of influencing factors.Results The overall quality score for statistical reporting in the surveyed general medical journals was 40.0±8.6.The scores for different statistical aspects were as follows:statistical design(43.2±6.8),statistical description(67.0±24.3),t-test(23.6±17.5),ANOVA(26.8±17.2),chi-square test(24.2±13.7),rank-sum test(45.5±18.4),correlation analysis(41.9±13.2),regression analysis(40.5±14.8),and survival analysis(45.6±14.9).Univariate analysis indicated that the overall quality score for statistical reporting in articles published in 2020 was higher than that in 2010(35.8±7.4 vs.44.3±7.5,t=-12.510,P<0.001).Multilevel modeling showed that the indexing system,research field,and publication year were the main factors influencing the overall score.Compared to Medline-indexed journals,the overall scores were lower for Chinese core journals.Articles focusing on non-human subjects had lower scores compared to those involving human subjects.Additionally,the overall scores for articles published in 2020 were higher than those in 2010.Conclusion TThe quality of statistical reporting in general medical journals in China has improved over the past decade,but there remains substantial room for further improvement.It is recommended to enhance the promotion of relevant reporting standards.

关 键 词：统计报告 中文医学文献 SAMPL指南 

分 类 号：R195.1[医药卫生—卫生统计学]