Department of Psychiatry, Vagelos College of Physicians and Surgeons, Columbia University, New York, New York.
Mailman School of Public Health, Columbia University, New York, New York.
JAMA Psychiatry. 2020 Sep 1;77(9):967-973. doi: 10.1001/jamapsychiatry.2020.0954.
The last 40 years of JAMA Psychiatry are reviewed as a celebration of its achievements. The focus of this article is on the evolution of big data as reflected in key journal articles. The review begins in 1984 with the introduction of the Epidemiology Catchment Area (ECA) study and Freedman's editorial "Psychiatric Epidemiology Counts." The ECA study (N = 17 000), for the first time in a survey, used clinical diagnosis in 5 urban communities, thus linking research and care to population rates of psychiatric diagnosis. The review then traces the subsequent evolution of big data to 5 overlapping phases, other population surveys in the US and globally, cohort studies, administrative claims, large genetic data sets, and electronic health records. Each of these topics are illustrated in articles in JAMA Psychiatry. The many caveats to these choices, the historical roots before 1984, as well as the controversy around the choice of topics and the term big data are acknowledged. The foundation for big data in psychiatry was built on the development of defined and reliable diagnosis, assessment tools that could be used in large samples, the computational evolution for handling large data sets, hypothesis generated by smaller studies of humans and animals with carefully crafted phenotypes, the welcoming of investigators from all over the world with calls for broader diversity, open access and the sharing of data, and introduction of electronic health records more recently. Future directions as well as the opportunities for the complementary roles of big and little data are described. JAMA Psychiatry will continue to be a rich resource of these publications.
回顾《美国医学会精神病学杂志》过去 40 年的发展历程，以庆祝其取得的成就。本文重点介绍反映在重点期刊文章中的大数据演变。本文从 1984 年的《流行病学抽样区（ECA）研究》和 Freedman 的社论“精神病流行病学有数据支撑”开始回顾。ECA 研究（N=17000）首次在 5 个城市社区的调查中使用临床诊断，从而将研究和护理与人群精神诊断率联系起来。然后，本文追溯了大数据随后的演变，共分为 5 个重叠阶段：美国和全球的其他人群调查、队列研究、行政索赔、大型遗传数据集和电子健康记录。《美国医学会精神病学杂志》中的文章都对这些主题进行了说明。承认了这些选择存在的许多警告、1984 年之前的历史根源以及围绕主题选择和大数据术语的争议。精神病学大数据的基础是建立在明确可靠的诊断、可用于大样本的评估工具、处理大数据集的计算演变、基于精心设计的表型对人类和动物进行的较小研究提出的假说、欢迎来自世界各地的研究人员加入，呼吁更广泛的多样性、开放获取和数据共享以及最近引入电子健康记录之上。本文还描述了未来的发展方向以及大数据和小数据互补作用的机会。《美国医学会精神病学杂志》将继续成为这些出版物的丰富资源。
