ROBIS工具评价影像学诊断性试验系统评价质量

ROBIS Evaluation of Quality Assessed by Iconographic Diagnostic Test System

目的采用ROBIS工具评价影像学诊断性试验系统评价的质量。资料与方法以"诊断性试验、系统评价、Meta分析、diagnos~*test、diagnos~*trial、systematic review、Meta-analysis"为检索词,检索2014年1月1日—2016年12月31日中国生物医学文献数据库、中国知网、万方数据库、Pub Med、Embase以及Cochrane图书馆中收录的相关文献。由2名研究者独立筛选文献和提取信息后,使用ROBIS工具对所纳入文献进行质量评价。结果本研究共纳入文献219篇,其中中文文献93篇、英文文献126篇。亚组分析结果显示,中、英文文献质量差异有统计学意义(P=0.018);中国地区与其他地区文献质量差异有统计学意义(P<0.001);发表年代间质量差异无统计学意义(P=0.34)。ROBIS评价结果显示,低风险文献仅15篇(6.85%),其中中文文献2篇、英文文献13篇。结论影像学诊断性试验2014—2016年公开发表的中、英文诊断性试验系统评价的质量整体较低,其中中文文献及国内学者发表的文献质量与国际文献尚存在差距。该领域系统评价的制订者在今后的研究中应当提前撰写计划书并完善报告,充分获取证据,针对文献筛选、信息提取以及评价原始研究偏倚等方面尽量减小偏倚,并考虑到结果的稳定性,在此基础上进一步提高该领域系统评价的质量。

Purpose To evaluate the quality assessed by iconographic diagnostic test system using ROBIS. Materials and Methods ＂Diagnostic tests, system evaluation, Meta analysis, diagnoses^＊test, diagnoses^＊trial, systematic review, meta-analysis＂ were used as search term to retrieve relevant literatures recorded in CBM, CNKI, Wanfang Data, Pub Med, Embase, and Cochrane Library from January. 1, 2014 to December. 31, 2016. After two independent researchers screened the literature and extracted the information, the quality of included literatures was evaluated using ROBIS. Results Totally 219 articles were included in the study, including 93 Chinese literatures and 126 English literatures. The results of subgroup analysis showed that the quality difference between Chinese and English documents was statistically significant（P=0.018）; the quality difference between the Chinese literatures and those of other regions was statistically significant（P0.001）; there was no statistical significance in quality difference regarding the year of publication（P=0.34）. The ROBIS evaluation results showed that there were only 15 literatures（6.85%） with low-risk, including 2 Chinese literatures and 13 English literatures. Conclusion The quality evaluated by Chinese and English diagnostic test system, which is publicly published in iconographic diagnostic test from 2014 to 2016, is generally low, and the quality of literatures published by domestic scholars and in Chinese is still behind international levels. The maker of system evaluation in this field should prepare their research plans in advance and fully report them in future research, adequately obtain evidence, minimize bias in respects of document screening, information extraction, and evaluation of original research bias, and take into account of results stability, based on which the quality of system evaluation in this field should be further improved.