Page 53 - FCW, August 2019

CaseStudy
Automating the review of Medicare claims
A new AI-based tool is improving accuracy and freeing CMS employees from rote tasks
BY STEPHANIE KANOWITZ
Employees and contractors at the Centers for Medicare and Medic- aid Services spend thousands of hours every year reviewing thousands of medical records to ensure the accuracy of payments under Medicare Advan- tage plans. An automated intake tool could change that.
Using emerging technologies such as robotic process automation, optical character recognition, machine learn- ing and artificial intelligence, KPMG’s Intake Process Automation (Intake PA) Tool ingests records as they are submitted and identifies potential problems according to set parameters, submission rules and coding guidance. Specifically, RPA orchestrates the steps of the intake process, OCR digitizes
the scanned document and then AI and machine learning are applied to understand the document and extract the details necessary to validate the information.
An easier way to verify unstructured data
The tool will save CMS time and money, said Payam Mousavi, KPMG’s lead director for government intelli- gent automation and technical lead for the CMS project. The tool processes a record in about a minute, where- as human reviewers average about 65 minutes per record. The agency received 40,000 records in 2018 and expects to receive 50,000 this year.
To submit records, health care pro-
viders scan them in, resulting in vary- ing formats and lengths. “It could go anywhere from a few pages to over 1,000 pages,” Mousavi said. “Because they’re scanned in, there’s sometimes handwriting in them.”
In the past, the CMS staff reviewed all that unstructured data, making sure the scans of records were rotated prop- erly and going page by page to check that they had the right date of service, provider and beneficiary.
Now, OCR handles the digitization, but to make sure that names, birth dates and other identifiers within one record match, KPMG created microser- vices that use natural language pro- cessing and machine learning to dis- tinguish between a date of birth and
August 2019 FCW.COM 49