AI and Analytics in Healthcare  

Video produced by Steve Nathans-Kelly

Analytics has been used in healthcare for decades. It has played a central role in helping companies to improve the healthcare of patients for decades as well as improve the healthcare system, especially here in the U.S., by trying to understand how it works, and how great outcomes are delivered—but that can be really hard.

At Data Summit Connect 2021, Optum SVP Sanji Fernando discussed how rule-based analytics, statistical reasoning, machine learning, and augmented intelligence deliver outcomes in the healthcare system.

Analytics in healthcare has strived to understand how costs are distributed, how care is distributed, how it might vary across this very heterogeneous system—and fixing this is not an easy task. Fernando said. For decades healthcare has used rule-based analysis to try to get at common units of analysis.

But rules can be constraining. In an effort to try to get better at this, new methods have been introduced—statistical reasoning in unstructured data, where we tried to best fit to the data that's available to us, has been an incredibly important part of healthcare and, and, and something that has been used to understand trends and how to serve populations better. 

"Now in the last few years with the renaissance of artificial intelligence, primarily driven by machine learning, we have new tools to allow us to almost, to some extent, but not completely, disregard that heterogeneity," said Fernando.

But, he added, if we can not worry about the features or the structure that we pass into a method, can we still classify things? That's at the heart of some of these new breakthroughs, like deep learning, a type of machine learning. And with the application of these new methods, we're now on the cusp of applying the best methods with what people need to make great decisions, whether they be a provider or a payer or a consumer.

The question is how to take the best methods, apply them to a problem, provide people with information that synthesizes very complex information, and do that safely and appropriately—and also do that in a way that helps people make better decisions and augments their intelligence. "That's what we're really excited about and the journey we're on with this type of innovation."