中文发表乳腺癌随机对照试验报告质量评价

Quality Assessment of Randomized Controlled Trial Report in the Field of Breast Cancer Published in Chinese

目的采用CONSORT声明2010版对乳腺癌领域发表的中文随机对照试验(RCT)文献报告质量进行客观评价,分析其存在的问题,为今后开展高质量的研究提供参考。方法检索中国生物医学文献数据库、万方数据资源、中国知网和维普数据库,检索策略为:(乳腺癌OR乳腺肿瘤OR乳腺恶性肿瘤)AND(随机),时间限定为建库至2013年9月。依据CONSORT声明2010版设计资料提取表格,由2名研究员对纳入的20篇文献进行预提取,交叉核对提取结果,针对有分歧的条目达成统一之后,2名评价人员独立评价纳入文献。根据各文献对CONSORT声明2010版各条目的符合程度将其判定为"完整报告"和"不完整报告或未报告"。结果共纳入968篇RCT,完整报告率较高(≥70%)的条目主有2a(765篇,79.03%)、4a(690篇,71.28%)、5(895篇,92.46%)、13a(836篇,86.36%)、21(690篇,71.28%)、22(692篇,71.49%)。多个条目完整报告率较差(≤10%),如1a(10篇,1.03%)、3b(15篇,1.55%)、6b(5篇,0.52%)、7b(2篇,0.21%)、8b(79篇,8.16%)、9(20篇,2.07%)、10(6篇,0.62%)、11a(11篇,1.14%)、12b(2篇,0.21%)、14b(3篇,0.31%)、17a(19篇,1.96%)、17b(70篇,7.23%)、18(9篇,0.93%)、24(0篇)、25(55篇,5.68%)。2010年及之后,条目2a、4b、12a完整报告率高于2010年之前,条目19完整报告率低于2010年之前,差异有统计学意义(P<0.05);其他条目完整报告率比较,差异无统计学意义(P>0.05)。结论目前发表的中文乳腺癌相关RCT文献整体报告质量不高,主要表现为随机分配序列的产生方法、分配隐藏以及盲法实施过程等方面报告缺陷。

Objective To objectively assess the quality and potential problems of randomized controlled trials（ RCTs）published in China in the field of breast cancer using 2010 CONSORT statement,in order to provide a reference for high- quality research in the future. Methods Retrieval was made in Chinese Biomedical Literature Database,Wanfang Database,China National Knowledge Infrastructure and VIP database from establishment to September 2013,with a searching strategy of（ breast cancer OR breast tumor OR breast malignant tumor） AND（ random）. Data extraction table was designed based on 2010 CONSORT statement. Preliminary data extraction was made from 20 pieces of literature included by 2 researchers,and cross-check was made on the extracted data. After the 2 researchers reached consensus on items with divergence, they made assessment on the included literatures independently. Based on the accordance degree of the included literatures with each item of2010 CONSORT statement, the literatures were assigned into ＂ complete reports ＂ and ＂ incomplete report or no report ＂.Results A total of 968 RCTs were included,the items with high rate of complete reports（ ≥70%） were 2a（ 765 pieces,79. 03%）,4a（ 690 pieces, 71. 28%）, 5（ 895 pieces, 92. 46%）, 13a（ 836 pieces, 86. 36%）, 21（ 690 pieces,71. 28%）,22（ 692 pieces,71. 49%）. There were a number of items with low complete report（ ≤10%）,such as 1a（ 10 pieces,1. 03%）, 3b（ 15 pieces, 1. 55%）, 6b（ 5 pieces, 0. 52%）, 7b（ 2 pieces, 0. 21%）, 8b（ 79 pieces,8. 16%）,9（ 20 pieces,2. 07%）,10（ 6 pieces,0. 62%）,11a（ 11 pieces,1. 14%）,12b（ 2 pieces,0. 21%）,14b（ 3 pieces,0. 31%）,17a（ 19 pieces,1. 96%）,17b（ 70 pieces,7. 23%）,18（ 9 pieces,0. 93%）,24（ 0 pieces） and25（ 55 pieces,5. 68%）. From 2010,the complete report rates of item 2a,4b and 12 a were higher than those before 2010（ P 〈0. 05）, and the complete report rate of item 19 was lower than that before 2010（ P 〈0. 05）; other items were not significantly different in complete report rate（ P 〉0. 05）. Conclusion The present reports of RCTs published in Chinese in the field of breast cancer is not in a good quality. The potential problems lie in the methods used to generate the random allocation sequence,allocation concealment mechanism and the process of blind method.