Table 1.

Sources of two different types of observational data

Research-generated data	Non-research generated data
Epidemiological studies	EHRs/EMRs
Data from cohort, cross-sectional, and case–control studies.	EHRs are digital records of patients’ medical data. Data stored in EHRs are structured (i.e. tabular data) and unstructured (e.g. free-text in clinical notes or image reports) [7].
Patient registries	National registry
A patient registry is an organised collection of uniform data to evaluate a pre-specified outcome(s) for a population with a specific disease, condition, or exposure [8].	A national registry collects uniform demographics and/or health related data on all its respective country nationals [9].
Biobanks	Health insurance claims databases
A biobank collects biological samples and in-depth health information on a specific group of people [10].	A health insurance claims database collects data entered on bills (claims) by hospitals, nursing homes, etc. [11].

In this paper no distinction is made between electronic health records and electronic medical records. This table was adapted from a lecture given by Miguel Hernán [12]

Abbreviations: EHRs Electronic health records, EMRs Electronic medical records