, 7 rue St.Hippolyte, 69008, Lyon, France.
Eur J Epidemiol. 2018 Mar;33(3):245-257. doi: 10.1007/s10654-018-0385-9. Epub 2018 Apr 5.
Big Data and precision medicine, two major contemporary challenges for epidemiology, are critically examined from two different angles. In Part 1 Big Data collected for research purposes (Big research Data) and Big Data used for research although collected for other primary purposes (Big secondary Data) are discussed in the light of the fundamental common requirement of data validity, prevailing over "bigness". Precision medicine is treated developing the key point that high relative risks are as a rule required to make a variable or combination of variables suitable for prediction of disease occurrence, outcome or response to treatment; the commercial proliferation of allegedly predictive tests of unknown or poor validity is commented. Part 2 proposes a "wise epidemiology" approach to: (a) choosing in a context imprinted by Big Data and precision medicine-epidemiological research projects actually relevant to population health, (b) training epidemiologists,
大数据和精准医学是流行病学面临的两大当代挑战，本文从两个不同角度对其进行了批判性的审视。在第一部分中，本文根据数据有效性这一基本要求，讨论了出于研究目的而收集的大数据（Big research Data）和虽为其他主要目的而收集但可用于研究的大数据（Big secondary Data），这种要求超越了“大数据”本身。本文还探讨了精准医学，指出通常需要较高的相对风险才能使一个或多个变量适用于疾病发生、结果或治疗反应的预测；并对那些所谓的预测性检测的商业扩散进行了评论，这些检测的有效性未知或较差。第二部分提出了一种“明智的流行病学”方法，用于：(a) 在大数据和精准医学流行病学研究项目背景下，选择真正与人群健康相关的项目；(b) 培训流行病学家，
