T raditionally, human subjects research in the biomedical field engages healthy and sick people as research participants in order to test certain hypotheses about health and disease according to a well-defined study design. The study designs, such as randomized controlled trials and cohort studies, are carefully reviewed by institutional review boards (IRBs)-also known as ethical review committees (ERCs) in some countries. IRBs are committed to protecting the rights and welfare of human subjects recruited to participate in biomedical or behavioral research (including social science research). 1 This approach has long been the standard for biomedical research. Recently, however, biomedical research has begun to pursue opportunities afforded by big data. Big data research relies on large-scale databases, multiplication of data sources, advanced storage capacity, and novel computational tools that allow for high-velocity data analytics. 2 In the biomedical domain, big data trends are enabled by and allow for advances in areas such as whole genome sequencing, brain imaging, mobile health, and digital phenotyping. 3 Today, a large portion of health-related research relies on big data. Big data also enables researchers to draw health insights from data sources that are not strictly medical-data from wearable trackers, social media, and Internet searches, for example. 4 Big data research opens new prospects to accelerate health-related research and potentially elicit breakthroughs that will benefit patients. 5 Big data has been observed to shift the way biomedical researchers design and carry out their studies. 6 This research departs from the traditional research model because it is largely exploratory rather than hypothesis driven. Health-related big data research is based on the acquisition of large amounts of data from multiple and often heterogeneous sources, which are subsequently combined and mined using powerful data analytics tools. This reverse-engineered approach to health-related research allows researchers to extract features and valuable insights from large datasets, without being able to anticipate exactly what the data analysis will find.