Salameh Jean-Paul, Moher David, McGrath Trevor A, Frank Robert A, Sharifabadi Anahita Dehmoobad, Islam Nabil, Lam Eric, Adamo Robert, Dawit Haben, Kashif Al-Ghita Mohammed, Levis Brooke, Thombs Brett D, Bossuyt Patrick M, McInnes Matthew D F
Department of Radiology, University of Ottawa, Ottawa, Canada.
The Ottawa Hospital Research Institute Clinical Epidemiology Program (Centre for Journalology), Ottawa, Canada.
J Appl Lab Med. 2025 Mar 3;10(2):416-431. doi: 10.1093/jalm/jfae117.
We evaluated reporting of diagnostic test accuracy (DTA) systematic reviews using Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)-DTA and PRISMA-DTA for abstracts.
We searched MEDLINE for recent DTA systematic reviews (September 2023-Mar 2024) to achieve a sample size of 100. Analyses evaluated adherence to PRISMA-DTA (and abstracts), on a per-item basis. Association of reporting with journal, country, impact factor (IF), index-test type, subspecialty area, use of supplemental material, PRISMA citation, word count, and PRISMA adoption was evaluated. Comparison to the baseline evaluation from 2019 was done. Protocol: https://doi.org/10.17605/OSF.IO/P25TE.
Overall adherence (n = 100) was 78% (20.3/26.0 items, SD = 2.0) for PRISMA-DTA and 52% (5.7/11.0 items, SD = 1.6) for abstracts. Infrequently reported items (<33% of studies): eligibility criteria, definitions for data extraction, synthesis of results, and characteristics of the included studies. Infrequently reported items in abstracts were characteristics of the included studies, strengths and limitations, and funding. Reporting completeness for full text was minimally higher in studies in higher IF journals [20.7 vs 19.8 items; 95% confidence interval (95%CI) (0.09; 1.77)], as well as studies that cited PRISMA [21.1 vs 20.1 items; 95%CI (0.04; 1.95)], or used supplemental material (20.7 vs 19.2 items; 95%CI (0.63; 2.35)]. Variability in reporting was not associated with author country, journal, abstract word count limitations, PRISMA adoption, structured abstracts, study design, subspecialty, open-access status, or index test. No association with word counts was observed among full text or abstracts. Compared to the baseline evaluation, reporting was improved for full texts [71% to 78%; 95%CI (1.18; 2.26)] but not for abstracts [50% to 52%; 95%CI (-0.20; 0.60)].
Compared to the baseline evaluation published in 2019, we observed modest improved adherence to PRISMA-DTA and no improvement in PRISMA-DTA for abstracts reporting.
我们使用系统评价与Meta分析的首选报告项目（PRISMA）-诊断试验准确性（DTA）以及PRISMA-DTA摘要版，对诊断试验准确性（DTA）系统评价的报告情况进行了评估。
我们在MEDLINE中检索了近期的DTA系统评价（2023年9月至2024年3月），以获得100个样本量。分析按项目逐一评估对PRISMA-DTA（及摘要）的依从性。评估了报告与期刊、国家、影响因子（IF）、索引测试类型、亚专业领域、补充材料的使用、PRISMA引用、字数以及PRISMA采用情况之间的关联。与2019年的基线评估进行了比较。方案：https://doi.org/10.17605/OSF.IO/P25TE。
PRISMA-DTA的总体依从率（n = 100）为78%（20.3/26.0项，标准差 = 2.0），摘要的依从率为52%（5.7/11.0项，标准差 = 1.6）。报告较少的项目（<33%的研究）：纳入标准、数据提取定义、结果综合以及纳入研究的特征。摘要中报告较少的项目是纳入研究的特征、优势与局限性以及资金情况。IF较高的期刊中的研究[20.7对19.8项；95%置信区间（95%CI）（0.09；1.77）]，以及引用PRISMA的研究[21.1对20.1项；95%CI（0.04；1.95）]，或使用补充材料的研究（20.7对19.2项；95%CI（0.63；2.35）]，全文报告的完整性略高。报告的变异性与作者所在国家、期刊、摘要字数限制、PRISMA采用情况、结构化摘要、研究设计、亚专业、开放获取状态或索引测试无关。全文或摘要中均未观察到与字数的关联。与基线评估相比，全文报告有所改善[71%至78%；95%CI（1.18；2.26）]，但摘要报告未改善[50%至52%；95%CI（-0.20；0.60）]。
与2019年发表的基线评估相比，我们观察到对PRISMA-DTA的依从性有适度提高，而PRISMA-DTA摘要报告方面没有改善。
