Company: Optum Serve
Company background: At Optum Serve, we’re dedicated to improving health across the nation. As the federal-facing business segment of UnitedHealth Group, we leverage our connection to both United HealthCare and Optum to deliver solutions that meet the broad spectrum of healthcare needs throughout the federal government. In bringing together our unmatched data with deep insights from our commercial businesses, we help solve challenges facing the government today—and uncover smarter solutions for tomorrow. Through close partnerships, we design tailored offerings that help customers work towards our shared goal: better health outcomes nationwide.
Our solutions include:
• Community care services ensure our nation’s Veterans have access to the right care, at the right time and in the right setting.
• Health services, designed to deliver tools needed to meet specific populations’ health care needs such as Military Service members, Veterans, and those that keep our country running and preparedness planning and emergency response for natural disasters and public health crises
• Health technology and consulting solutions, which provide policy research and design to help create and manage the best strategies to maximize agency services, as well as improve data and technology systems and increase efficiency in managing health across wide populations
Contact: Johnny Brevik
Title: Chief Technology Officer, Optum Serve
Country: United States of America
Award Category: Innovative Problem Solver
Tell us about the business problem you were trying to solve?
Optum Serve supports the Department of Veterans Affairs (VA) by performing medical disability examinations in support of VA's adjudication process for disability benefit claims for Veterans. To prepare for an appointment, the provider is required to review the Veteran’s clinical and service records—a set of documents which averages 1,500 pages. Only a subset of the documents is relevant to the Veteran's claimed condition, so for providers, it can feel like searching for a needle in a haystack. In partnership with SAS’s AI Modeling technology, we automated this review, improving efficiency and reducing the upfront time needed from providers allowing them to focus that time on interacting with Veterans.
How did you use SAS to solve that business problem? What products did you use and how did you use them?
Optum Serve utilized a strategic partnership with SAS Institute to build a fully-automated solution to analyze the Veteran's records and extract relevant insights. We used an agile approach to iterate on solutions and formed a provider advisory group to conduct usability testing and gain real user feedback improving results. Several Artificial Intelligence (AI) technologies were combined to power the solution, including advanced optical character recognition, machine learning, computer vision, and natural language processing. Fifty analytic data models were tailored to produce a claim-specific summary for providers, along with structured bookmarks to help the provider review the document efficiently and effectively.
What were the results or outcomes?
The solution is fully operational today, leveraging innovative SAS modeling to process more than three million pages per day. Previously, Optum Serve's clinically trained workforce scrolled through the documents, scanning for evidence, then adding bookmarks manually. As the solution was built, we benchmarked its output against that of the manual process. The final AI-based solution significantly outperformed the manual solution, demonstrating that it extracts more accurate data, developing higher quality insights, leading to a higher quality exam. It also enabled business scalability and faster turnaround for exams by decreasing the time a provider must spend on each claim improving effectiveness.
Why is this approach innovative?
Several Artificial Intelligence (AI) technologies were combined to power the solution, including advanced optical character recognition, machine learning, computer vision, and natural language processing. Fifty analytic data models were tailored to produce a claim-specific summary for providers, along with structured bookmarks to help the provider review the document efficiently. The final AI-based solution significantly outperformed the manual solution, demonstrating that it extracts more accurate data, developing higher quality insights, leading to a higher quality exam. It also enabled business scalability and faster turnaround for exams by decreasing the time a provider must spend on each claim improving effectiveness.
