The next public health pronouncement or clinical treatment might emerge from synthetic data: rows of totally invented people that contain no real data but reflects the actual characteristics of a population such as race, gender, and medical conditions.
Synthetic data plus large language models (LLMs), which lie at the base of current generative AI, provide “the world’s most powerful research assistant” in the words of Josh Rubel, chief commercial officer for MDClone. This video contains a wide-ranging discussion between Rubel and interviewer John Lynn about the current applications and future possibilities of synthetic data, LLMs, and generative AI.
MDClone offers a health data management platform to health care institutions for quality improvement, performance improvement, academic research, and third-party integration. A typical application described by Rubel is hypothesis testing. For instance, will phone calls to patient for visit reminders work better in the morning or in the afternoon?
AI might choose the best chart to illustrate a concept, or even choose among available treatments. In all these cases, well-trained AI can work much faster and more cheaply than a human consultant.
Research on human patients, Rubel says, usually compares them along 10 to 20 dimensions. MDClone can take a data set on real patients and generate a synthetic data set of fake people that contains no personal health information (PHI) but still reveals that limited set of dimensions. Thus, you can use the synthetic data to compare characteristics that matter to you.
Rubel calls this synthetic data a “precision cohort” and says it “tells the same story the real patient data would have told.” The FDA, VA, and others are using synthetic data.
Watch the video for a fascinating conversation on how LLMs work and where they may prove useful in years to come.
Learn more about MDClone: https://www.mdclone.com/
Listen and subscribe to the Healthcare IT Today Interviews Podcast to hear all the latest insights from experts in healthcare IT.
And for an exclusive look at our top stories, subscribe to our newsletter.
Tell us what you think. Contact us here or on Twitter at @hcitoday. And if you’re interested in advertising with us, check out our various advertising packages and request our Media Kit.
