2019 Government Innovation Awards
Automation speeds review of medical records
To limit the amount of federal funds going to fraudsters, the government requires agencies to review their disbursements to identify and recover improper payments.
At the Centers for Medicare and Medicaid Services, employees would conduct an annual, manual audit of more than 50,000 unstructured medical records, often reviewing hundreds of pages of documents to verify that diagnosis codes were supported by medical documentation. With the volume of records expected to increase each year, CMS needed a better option.
The agency created the Intake Process Automation tool, which digitizes records by using advanced optical character recognition and applying natural language processing to understand the language and context of the data. The tool extracts information such as patient name, date of birth, provider type, date of service and other fields requiring confirmation. IPA validates extracted data fields and identifies potential issues based on predefined business rules and coding guidance. It then updates the CMS system with results from the reviews. If the tool doesn't have high confidence in the decision it made, the record will go to a person for verification.
By automating the initial evaluation of medical records, CMS has been able to vastly speed processing of the reviews, increase accuracy and allow employees to spend their time analyzing exceptions instead of verifying existing records. CMS also saves money and delivers faster feedback to insurers and providers on the validity of medical records submitted to support payments.
Connect with the GCN staff on Twitter @GCNtech.