Bulging health record databases and burgeoning artificial intelligence are coming together to revolutionize medical research. One research team on its own can’t see tens of thousands of patients with a challenging ailment, or read data from millions of Fitbits. But machine learning can scan such heaps of information for insights—if the patients give their permission amid privacy concerns. The U.S. Food and Drug Administration and IBM Watson Health are starting a two-year study on how to give patients full control over those permissions using blockchain, the same kind of auditing technology that bitcoin uses.
Quick blockchain description: a digital ledger of transactions, synchronized across several computers, so that every one has a complete record of what’s happened. FDA wants to test blockchain as a universal audit trail, tracking what institutions have a patient’s (possibly already anonymized) records and where copies are transferred to—all contingent on the patient approving each transfer. It’s still an honor system that every transaction will get recorded, but IBM envisions software that ties the medical databases and blockchain ledgers together. You can’t transfer a record until the patient gives permission, through an entry in the ledger; and the actual transaction automatically creates its own ledger entry.