Clinical Datasets: Open Access Electronic Health Records Datasets

This chapter details open access to electronic health records (EHR) datasets that are available through a variety of sources. We discuss why EHR datasets are not routinely made open access due to patient privacy concerns and laws. We also discuss the most widely used open-access EHR dataset, namely MIMIC, which is made available through physionet and provides open-access diagnostic code information for diagnoses, laboratory values, medications, clinical notes, and even imaging. In addition, we describe less well-used datasets from Norway, the United Arab Emirates, and Brazil along with resources on how to identify future EHR datasets that may be released. We describe the United Nations repository of health information (humdata). We also provide an overview of a popular synthetic EHR dataset that is widely available and made to be as “realistic” as possible. After reading this chapter, you should be able to confidently answer the following questions: