Imai Shinobu
School of Pharmacy, Tokyo University of Pharmacy and Life Sciences.
Yakugaku Zasshi. 2021;141(2):169-174. doi: 10.1248/yakushi.20-00196-2.
Medical big data, also referred to as 'real-world data' (RWD) is defined as "data related to patient health status and/or health care delivery collected routinely from a variety of sources". This includes data from disease and drug registries, electronic health records, claims and billing data and census data collected from clinicians, hospitals, and payers. Observational studies using RWD collected during general clinical practice are considered complementary to randomized control trials. However, since this design does not allow the random assignment of patients, causal inference analyses are required. Researchers should study the protocol properly before considering the combination of study design, the characteristics of data source, calculation of the appropriate sample size and the validity of outcomes. Data definition using data code should also be considered. Furthermore, the reliability of the source studies must be considered and discussed when the article is written. This review aims to outline the methods for performing reliable observational studies using RWD.
医学大数据，也被称为“真实世界数据”（RWD），被定义为“从各种来源常规收集的与患者健康状况和/或医疗保健服务相关的数据”。这包括来自疾病和药物登记处、电子健康记录、理赔和计费数据以及从临床医生、医院和付款人收集的人口普查数据。在一般临床实践中使用收集到的RWD进行的观察性研究被认为是随机对照试验的补充。然而，由于这种设计不允许对患者进行随机分配，因此需要进行因果推断分析。研究人员在考虑研究设计的组合、数据源的特征、适当样本量的计算以及结果的有效性之前，应该正确研究方案。还应考虑使用数据代码进行数据定义。此外，在撰写文章时，必须考虑并讨论源研究的可靠性。本综述旨在概述使用RWD进行可靠观察性研究的方法。
