摘要：

【目的】 对我国中文医学期刊近年发表的临床研究论文质量的整体情况进行综合性评价。【方法】 采用滚雪球抽样法在CNKI、万方数据库抽取中文医学期刊发表的临床研究论文649篇,经排除后纳入一次研究论文229篇。综合文献研究及专家访谈结果确定评价指标,对纳入论文的研究伦理、创新性、报告质量、方法学质量4个方面进行审读评判与数据分析。【结果】 229篇纳入论文中,77.3%(177篇)单组样本量低于100;单中心研究占97.38%,阳性结果占99.22%。①伦理:纳入论文中仅3.06%的研究经过注册;16.95%(38篇)报告了伦理审查;无一发布利益冲突相关信息。核心期刊的伦理审查报告情况显著优于非核心期刊(χ²=6.603,P=0.010)。②创新性:经查新,同类研究>50篇的纳入论文占12.23%,同类研究>5篇且单组样本量<100的占34.50%,这两部分论文很可能为虚假研究。核心期刊与非核心期刊的创新性无显著性差异(U=6236.0,P=0.966)。③报告质量:纳入论文的国际清单报告率平均为(37.87±15.80)%;清单报告率<30%的占39.30%;不足一半的占85.15%;报告较全的(>70%)仅占5.68%。④方法学质量:纳入论文中高风险论文占44.10%;中风险论文占34.50%。偏倚风险与报告质量呈显著负相关(r=-0.132,P=0.044)。核心期刊比非核心期刊发表的RCT明显更少(21.12% vs 78.87%),清单报告率(U=4744.000,P<0.001)与偏倚风险评分结果(U=5260.000,P=0.012)显著更优。【结论】 近年中文医学期刊发表了大量重复性的单中心、小样本、纯阳性结果、低质量(报告质量低、偏倚风险高)的临床研究;研究注册和伦理审查率均低。核心期刊相对重视质量控制,但更多的中文医学期刊对临床研究论文质量把关不严。现有的评价体系不能反映中文医学期刊的学术质量现况,亟待建立以学术质量为核心的医学期刊评价体系。

关键词: 中文医学期刊, 学术质量, 临床研究论文, 评价体系

Abstract:

[Purposes] This paper aims to comprehensively evaluate the overall quality of clinical research articles published on Chinese medical journals in recent years. [Methods] Snowball sampling was used to screen 649 clinical research articles from CNKI and Wanfang Data and finally 229 experimental and analytical articles were included. Based on literature research and expert interviews, the evaluation indices were determined and thereby articles were assessed from the aspects of research ethics, innovativeness, reporting quality, and methodology. [Findings] Among the 229 articles, 77.3% (177) had the single-group sample size <100 and 97.38% were single-center studies. In addition, positive results were found in 99.22% of the articles. ① Research ethics: Only 3.06% (7) of the 229 articles were registered on the clinical trial registration platform and 16.95% (38) reported ethical approval. None of them disclosed information about conflicts of interest. More ethical examination reports were displayed in articles on core journals than on non-core journals (χ²=6.603, P=0.010). ② Innovativeness: Each of the 28 articles (12.23% of 229) and 79 articles (34.50% of 229, single-group sample size <100) was found to have a similar topic to >50 previous articles and >5 previous articles, respectively. The above two cases might involve data fabrication or falsification. There was no significant difference in innovativeness between articles on core journals and non-core journals (U=6236.0, P=0.966). ③ Reporting quality: The overall adherence to corresponding international checklists was (37.87±15.80)%. A total of 39.30%, 85.15%, and 5.68% of the articles demonstrated the adherence of <30%, ≤50%, and >70%, separately. ④ Methodology: The proportions of high risk, medium risk, and low risk of bias among the 229 articles were 44.10%, 34.50%, and 21.40%, respectively. Spearman rank correlation analysis showed that the risk of bias was negatively correlated with the reporting quality (r=-0.132, P=0.044). Core journals published significantly fewer RCTs (21.12% vs 78.87%) and articles of higher adherence to international checklists (U=4744.000, P<0.001) and lower risk of bias (U=5260.000, P=0.012) than non-core journals. [Conclusions] In recent years, Chinese medical journals have published lots of repetitive single-center clinical studies with small sample size, totally positive results, low quality (low reporting quality and high risk of bias), and small proportion of registration and ethical review. Core journals attach more importance to the quality of articles as compared with non-core journals. The current evaluation system cannot comprehensively reflect the quality of the Chinese medical journals. It is urgent to establish an evaluation system centering on the quality to improve the overall academic quality of such journals.

Key words: Chinese medical journal, Academic quality, Clinical research article, Evaluation system