Big data have changed the way we generate, manage, analyze and leverage data in any industries. There is no exception in clinical medicine where large volume of data is generated from electronic healthcare records, wearable devices and insurance companies ( ). This has greatly changed the way we perform clinical studies. Instead of performing data entry and curation manually, the information technology has significantly improved the efficacy of data management. With such a large volume of data, many clinical questions can be addressed by using big data analytics ( - ). Three steps are typically involved in the big data analytics ( ). The first step is the formulation of clinical questions ( ), which can be categorized into three types: (I) epidemiological question on prevalence and incidence and risk factors; (II) effectiveness and/or safety of an intervention; and (III) predictive analytics. The second step is the design of a study, which transforms the clinical question into a study design. For example, the prevalence of catheter-related blood stream infection (CRBSI) as well as its risk factors can be addressed with retrospective or prospective cohort study. A case-control study design can be used to identify risk factors. The effectiveness can be addressed by a randomized controlled trial or an observational study. The third step involves the statistical analysis and/or modelling by using data collected under a certain design.
展开▼
