Doctors, researchers, and companies have made big promises about how machine learning and AI may someday change medicine. But a particularly American issue may be holding doctors and data scientists here back: electronic medical records.
Machine learning algorithms only work if they have data — lots and lots of data. The conclusions they draw about what might happen to a particular person will generally only work if an algorithm has been trained with a bunch of records of people who have some similar characteristics. And especially in the traditional “doctor’s office,” companies that want to work on AI and machine learning must somehow pull in information from a plethora of EMRs — a burden not often seen in other countries.
In the U.S., patient data are stored across more than 700 electronic medical records systems. About 92 percent of hospitals use one of the top five or six EMR providers; but just 60 percent of health care providers’ offices can say the same.
About the Author Reprints
To submit a correction request, please visit our Contact Us page.
STAT encourages you to share your voice. We welcome your commentary, criticism, and expertise on our subscriber-only platform, STAT+ Connect