Software with 20/20 vision streamlines claims processing

Handwritten notes. Fields left blank. Data saved as an image. Advances in what software can “see” mean health care is moving toward a day when these will no longer trigger tedious, manual rework.

Clearing the way to focus on members

Optum is working with our state partners to improve optical character recognition (OCR) tools. We’re applying their automation capabilities to health care challenges, including maintenance and operations of Medicaid enterprise systems. Together, we’re creating new efficiencies that will allow Medicaid agencies to focus on serving their populations.

Claims processing is just one area where OCR can make a difference. It solves problems related to information that’s missing or submitted in a nonstandard way. For example, a dollar amount entered by hand or saved as an image file.

“Typically, traditional programs can’t decipher images and handwritten data. In these situations, most of the cases fail and claims may be inappropriately denied,” says Amit Kumar, associate director, Data Analytics, State Government Solutions at Optum.

Combining OCR, ML and NLP technologies

Images and handwritten notes do not stump OCR programs. They can recognize the information, pull it from a document and auto-populate it into standard fields.

By combining OCR tools with natural language processing (NLP) and machine learning (ML), systems can:

• Spot a blank field in a submission and automatically alert physicians or facilities to the omission.
• Enhance document quality to allow for better character recognition in a form. This provides better categorizing of the data.
• Support intake and routing processes, case creation, decision-making and letter generation.

Optum is combining OCR, ML and NLP technologies in our claims processing solution.

Our OCR technology ranks in the high 90th percentile in accuracy for digital content. For handwritten content, the program can completely interpret 60 out of 100 cases. And the system gets better with every new piece of data provided, thanks to machine learning.

“We're thrilled to offer a solution to this process that helps states significantly benefit their community,” says Kumar. The innovative technology protects beneficiaries in that they won’t receive an inappropriate denial because of a single error, omission or nonstandard submission.

It also reduces the risks of delays. Automation capabilities keep documents moving through the system even when there is an omission or error. The process isn’t suspended until a state employee can take the time to review the files manually.

It assists with a better claims submission experience since any issue won’t necessarily mean the entire claim must be reworked and resubmitted.

Automation solutions allow states to do more with less, do it faster and with greater accuracy. That means they can focus on delivering the services their beneficiaries rely on.