CONTEXT: Surely the next big thing in RWR – “Synthetic Data”. “Yes” you say “and don’t call me Shirley”. Airplane jokes aside, it makes a huge amount of sense to let the silicon pathways run big numbers to see how a particular scenario plays out. Most data gurus will give you back the “garbage in / garbage out” mantra, but as our algorithms get better and the real life databases more accessible, synthetic research will get more reliable. As a quick and easy way to test a hypothesis that you can then replicate in-vivo with much more confidence, this looks like the future people. Maybe a little while longer for the FDA to accept as part of a submission package, but after the weirdness of the last 12 months, I wouldn’t bet against it.
IMPACT: High
READ TIME: 1 min
Quality Level Mean [1 – 10]: 9
1. “In a study published recently in the Journal of the American Medical Informatics Association: Open, researchers at the Institute for Informatics showed that synthetic data accurately mimicked the results of clinical studies that had been performed using the real patient datasets.”
2. “But what if some research could be conducted using synthetic datasets that mimic real patient populations but don’t carry the risk of disclosing protected health information?”
3. “The institute has shown that software, called MDClone, can accurately produce synthetic data based on real patient data in electronic health records.”
4. ““We’ve realized the power of synthetic data to accelerate the process of asking and answering questions involving real patient data,” said senior author Philip R.O. Payne, PhD, the Janet and Bernard Becker Professor and director of Washington University’s Institute for Informatics.”
5. “If researchers want to run queries on synthetic data, get some preliminary results or generate some hypotheses before requesting access to real data, that would be a good use of this platform.”