Vee Healthtek needed a tool for managing and navigating the tumultuous payer services market for a client. They needed to identify a way to support multiple disciplines in one, consistent, uniform tool. For that, Vee Healthtek created an internal advisory team to find the answer, culling the best ideas to develop a system called the Coding Assistance Platform.
The goal of CAP was to provide a single processing application that could work in multiple disciplines, including claims data capture, member enrollment, utilization management, mail room services, and more. The system would be module-based, allowing client-specific workflows to be created.
While the system has multiple capabilities, its biggest attribute is its ability in data capture on claims. The system can support medical, dental, and workers comp claims, as well as both structured and unstructured documents.
For structured documents that come on standardized forms, image pre-processing and manual ID are implemented. The pre-processing step is an automated form identification process. Multiple types of forms such as HCFA’s, UB’s, and dental forms can be intermixed in a single batch. The system will automatically identify the form type and, while utilizing multiple OCR engines, systematically extract the data keyed in the field. This data is then matched against rules and database extracts to validate the data. For example, date fields are checked to make sure they follow the MM/DD/YYYY format. Future dates are flagged for review by the system as well as dates prior to a date of incident. ID fields are compared to eligibility databases to match address and other demographic fields. City, state and zip codes are compared with a Postal database for validity. All these checks result in high-quality data being captured.
For unstructured documents (i.e., free flowing text not on a standard form), a manual ID process is implemented. A user views the document and assigns a form type which then determines the fields presented to the operator in the keying stage.
Finally, any claims that were manually keyed are sent to a verify stage. Here a second operator rekeys the file. The system then compares the two entries to ensure they match. If they don’t, the second operator must rekey it again. It is checked again and is accepted if matches. This allows that any manually keyed field is entered successfully at least two times before being accepted.
Vee Healthtek has been actively implementing Machine Learning (“ML”) and Natural Language Processing ("NLP") into the CAP environment. These technologies, combined with Robotic Processing Automation (“RPA”), continue to improve the process by eliminating the manual processing steps. Automation of these tasks results in faster processing times and improved quality. Robust dashboards present important data to clients in an easy to read view. Vee Healthtek utilizes cutting-edge technologies to provide their customers with fast turnaround and high-quality data. We are committed to our client’s success and strive to consistently adapt to and overcome the challenges that lie ahead.
