Machine learning (ML), AI, and other advanced technology tools used for detecting fraud at the Department of Health and Human Services’ (HHS) Center for Medicare & Medicaid Services (CMS) are likely to play an increasingly important role in the future of the agency, a CMS official said April 6 at General Dynamics Information Technology’s Emerge 2021 conference on digital modernization.
Raymond Wedgeworth, Director of the Data Analytics and Systems Group for the Center for Program Integrity within CMS, said early evidence suggests that the agency’s ML models produce less false positives and are valuable to their partners doing investigations into fraud.
“We’re always looking for ways to innovate to deliver better results, in this case, because we want to find fraud more accurately and faster … we want to be able to predict and prioritize our leads based upon the greatest likelihood of outcome,” said Wedgeworth. He added, “we also want to continue to automate a lot of our manual processes.”
Wedgeworth said that for close to 10 years now, CMS and has been using advanced analytics to curb fraud and waste. More recently, they agency has moved into more predictive models and is seeking to start working on fraud that is not yet well known.
“What we’re hoping, though, is to expand [learning models] even greater in terms of looking at identifying fraud that’s not known, so as you’re talking about with data bias, right now when we have a lot of the actions that we’ve taken in the past … what you did in the past is then being used to reinforce the new machine learning models,” Wedgeworth explained. “So, you can end up in an infinite loop and just keep going back to the same thing, so we’re not identifying potentially new and emerging fraud.”
Wedgeworth said that the second phase of ML for the agency will be to search through data and find patterns that a human might be seeing. This could identify things they haven’t seen before, and also identify those patterns earlier in the process, he said.