Clinical epidemiology in the era of Big Data: New opportunities, familiar challenges

Clinical epidemiology is the study of the outcome of illness, and pharmacoepidemiology is the study of the outcome of treated illness. Nowadays pharmacoepidemiology relies primarily and increasingly on non-experimental studies based on routinely collected data for descriptive, predictive and causal research.  Large data sets reduce the random error and often yield highly statistically significant and precise results. The danger is in succumbing to the fallacy that precise equals valid and statistically significant equals clinically important. The tried-and-tested remedies include transparency, use of valid tools and methods, applying expert knowledge, and bringing randomized trials closer to the real world.