BIG DATA AND MEDICINES. The Economist article seems to assume that the only interesting source of information about the effectiveness of drugs is a clinical trial. I will take Tamiflu as a hypothetical. The article points out that the British government has spent about $660 million dollars since 2006 stockpiling Tamiflu to protect against a future influenza pandemic. It rightly laments that at the time the decision was made, 60% of the trial data on Tamiflu had not been published. The Economist notes that analyzing that clinical trial data which was not published in 2006 now raises questions about the effectiveness of Tamiflu.
However, Tamiflu has been widely used since 2006. What does the data relating to all those uses show about the effectiveness of Tamiflu? There has been a lot of experience with Tamiflu, and in principle there should be a lot of data about what happened to people who took the drug—albeit that the data would not be the product of a randomized controlled experiment.
We now have the computing power to analyze huge amounts of data. Each time somebody takes a drug can be considered as a mini-experiment. Subject to privacy constraints and cost considerations, we should be collecting information systematically about what the aggregate of these experiments show about the success and safety of these drugs.