麻醉与数据库:以小儿心脏病为范例
本文以小儿先天性心脏病为例,阐述大型数据集在临床医学中的应用,包括行政数据与临床数据的分类、概率统计匹配、协作质量改进及风险调整等原则,并指出美国及国外心脏病患者诊疗结局存在显著差异,区域化与协作学习可减少差异。
要点
- 大型数据集分行政数据与临床数据,可概率匹配整合用于评估诊疗模式、结局与费用。
- 小儿先心病已有超过十几个数据集或注册登记,聚焦不同诊疗要素。
- 使用数据需遵循联邦指南保护健康信息,研究需风险调整与伦理监管。
核心结论
- 大型数据集有助于高 acuity 低容量疾病如先天性心脏病的多中心数据收集与质量改进。
研究设计与样本
- 本文为综述,未报告具体检索范围与样本量。
主要发现
- 美国及国外心脏病患者诊疗在死亡率、并发症和费用上持续存在广泛差异。
- 基于中心容量和结局的区域化诊疗受到大量讨论,但缺乏区域化时协作学习可减少差异并改善所有中心诊疗。
- 已通过概率统计匹配将行政数据与临床数据事后结合,以获取诊疗模式、结局和费用的有价值信息。
关键数值与推荐
- 大型数据管理的基本原则是使用共同语言和定义(术语)以确保有效性。
- 源自这些信息源的研究必须适当平衡,确保分析中考虑术前和手术因素的风险调整。
- 结局测量通常属于医院质量改进项目,但更详细的分析和研究需要机构审查委员会和行政监管。
临床怎么用
- 麻醉科医生可参与协作学习,利用数据集测量并改善小儿心脏麻醉诊疗结局。
- 心脏麻醉提供者已与胸外科医师学会先天性心脏外科医生合作,纳入麻醉要素以协助质量改进。
局限性
- 数据可能包含受保护健康信息,使用时须遵守联邦指南。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
Large data sets have now become ubiquitous in clinical medicine; they are particularly useful in high-acuity, low-volume conditions such as congenital heart disease where data must be collected from many centers. These data fall into 2 categories: administrative data arising from hospital admissions and charges and clinical data relating to specific diseases or procedures. In congenital cardiac diseases, there are now over a dozen of these data sets or registries focusing on various elements of patient care. Using probabilistic statistic matching, it is possible to marry administrative and clinical data post hoc using common elements to determine valuable information about care patterns, outcomes, and costs. These data sets can also be used in a collaborative fashion between institutions to drive quality improvement (QI). Because these data may include protected health information (PHI), care must be taken to adhere to federal guidelines on their use. A fundamental principle of large data management is the use of a common language and definition (nomenclature) to be effective. In addition, research derived from these information sources must be appropriately balanced to ensure that risk adjustments for preoperative and surgical factors are taken into consideration during the analysis. Care of patients with cardiac disease both in the United States and abroad consistently shows wide variability in mortality, morbidity, and costs, and there has been a tremendous amount of discussion about the benefits of regionalization of care based on center volume and outcome measurements. In the absence of regionalization, collaborative learning techniques have consistently been shown to minimize this variability and improve care at all centers, but before changes can be made it is necessary to accurately measure accurately current patient outcomes. Outcomes measurement generally falls under hospital-based QI initiatives, but more detailed analysis and research require Institutional Review Board and administrative oversight. Cardiac anesthesia providers for these patients have partnered with the Society of Thoracic Surgeons Congenital Heart surgeons to include anesthesia elements to help in this process.
引用
- Vener DF, Pasquali SK, Mossad EB, et al. Anesthesia and Databases: Pediatric Cardiac Disease as a Role Model. Anesth Analg. 2017. PMID: 28099323. doi:10.1213/ANE.0000000000001448
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。