Q2CRBench-3
收藏资源简介:
Q2CRBench-3数据集是一个用于临床推荐开发的基准数据集,基于三种不同疾病的临床指南开发记录构建。该数据集包括三种疾病的数据集,分别为2021年美国风湿病学会(ACR)类风湿性关节炎(RA)指南、2020年欧洲神经学会(EAN)痴呆症指南和2024年改善全球肾脏病预后(KDIGO)慢性肾脏病(CKD)指南。每个数据集包括五个相互关联的阶段:问题分解、文献检索、研究选择、证据评估和推荐制定。数据集包含近26200条文献记录,其中99.49%在筛选过程中被排除。Q2CRBench-3是第一个全面捕捉临床推荐制定完整端到端过程的综合基准数据集,旨在促进未来方法学的发展和领域内研究。
Q2CRBench-3 is a benchmark dataset for clinical recommendation development, constructed based on clinical guideline development records of three distinct diseases. This dataset encompasses three sub-datasets corresponding to three clinical guidelines: the 2021 American College of Rheumatology (ACR) rheumatoid arthritis (RA) guideline, the 2020 European Academy of Neurology (EAN) dementia guideline, and the 2024 Kidney Disease: Improving Global Outcomes (KDIGO) chronic kidney disease (CKD) guideline. Each sub-dataset covers five interrelated stages: problem decomposition, literature search, study selection, evidence appraisal, and recommendation formulation. The dataset contains nearly 26,200 literature records, with 99.49% of them being excluded during the screening process. Q2CRBench-3 is the first comprehensive benchmark dataset that fully captures the entire end-to-end process of clinical recommendation development, aiming to facilitate the advancement of methodological research and relevant studies in this field.

- 1From Questions to Clinical Recommendations: Large Language Models Driving Evidence-Based Clinical Decision Making浙江大学医学院生物医学工程与仪器科学学院; 教育部EMR与智能专家系统工程研究中心, 中国科学院北京协和医学院北京协和医院风湿免疫科和临床免疫科; 科技部国家临床研究中心皮肤病与免疫病; 国家复杂重大疾病重点实验室; 教育部风湿免疫学及临床免疫学重点实验室, 中国科学院北京协和医学院; 浙江实验室科学数据中心 · 2025年



