PROSPERO平台注册诊断性试验系统评价/Meta分析的基本特征及合作情况分析

Cooperation and basic characteristics of systematic reviews/meta-analyses on diagnostic test registered in PROSPERO platform

目的评价PROSPERO平台注册诊断性试验系统评价/Meta分析的基本特征及合作情况,为研究人员注册和开展诊断性试验系统评价/Meta分析提供借鉴。方法计算机检索PROSPERO注册平台,搜集注册的诊断性试验系统评价/Meta分析,检索时限截至2017年11月12日。由2名研究员独立筛选文献、提取资料并采用STATA 13.0制作森林图,采用BICOMS(书目信息共现分析系统)分析软件对国家、作者和单位等信息进行抽取并生成共现矩阵。采用Ucinet 6.0软件中NetDraw功能绘制国家、作者和作者单位合作社会网络图。结果 PROSPERO平台共注册了240个诊断性试验系统评价/Meta分析,其在作者(1050个)、机构(360个)和国家(34个)分布方面均不平衡。诊断性试验系统评价/Meta分析涉及21个系统疾病,最多的疾病为肿瘤(52个,22%);组织活检是使用最多的金标准,影像学诊断是使用最多的诊断试验。检索使用频率最高的外文数据库为PubMed、EMbase、The Cochrane Library,使用频率最高的中文数据库为CNKI、WanFang Data和CBM。105个诊断性试验系统评价/Meta分析(43.8%)报道了检索策略,159个(66.3%)使用QUADAS-2评价纳入研究质量,105个(43.8%)报告了基金资助情况。结论 PROSPERO平台注册诊断性试验系统评价/Meta分析的数量相对较少,作者、机构和国家间存在一定的联系,但对诊断性试验系统评价/Meta分析的效应指标和统计方法仍描述不够。

Objectives To analyze the cooperation and basic characteristics on systematic reviews/meta-analyses (SRs/MAs) in diagnostic test registered in PROSPERO platform, in order to provide reference for registering and preparing the SRs/MAs on diagnostic test. Methods We searched PROSPERO platform from inception to November 12th, 2017, to identify SRs/MAs on diagnostic test. Two reviewers independently screened records and extracted data by the inclusion and exclusion criteria. The forest figure was prepared by STATA 13.0 software. The figure of network of authors, organizations and countries were prepared by NetDraw software. Results A total of 240 SRs/MAs in diagnostic test were included. There was a serious imbalance in the quantity of the distribution of authors(1050), organizations(360) and countries(34). These SRs/MAs in diagnostic test were related to 21 systematic diseases, with cancer ranking the first(52,22%). Tissue biopsy was the most used gold standard, and imaging diagnosis was the most used diagnostic test.The most commonly retrieved English databases were PubMed, EMbase The Cochrane Library, while the Chinese databases were CNKI, WanFang Data and CBM. The search strategy were reported in 105(43.8%) SRs/MAs on diagnostic test. The QUADAS-2 was used to assess the quality of included studies in 159(66.3%) SRs/MAs in diagnostic test, and 105(43.8%) SRs/MAs were supported by funding. Conclusions The absolute quantity of SRs/MAs in diagnostic test is still minimal. There are relationships among different authors, organizations and countries. The researcher should focus on the effective sizes and statistical methods in future.