olegdudko ©123RF.com

By Brett Fraser, Vice President of Artificial Intelligence, NCI Information Systems, Inc. and Michael Ezzo, Director of Software Engineering at AdvanceMed, an NCI Company

Artificial intelligence (AI) and machine learning are rapidly transforming Healthcare, medical research and health services. These technologies also have the power to help Government agencies improve their services while reducing costs.

Healthcare and related health services are among our biggest Government expenditures. With deficits growing, agencies like the Centers for Medicare and Medicaid Services (CMS), the National Institutes of Health (NIH), and the Department of Veterans Affairs (VA) are under tremendous pressure to reduce budgets—without reducing care or services. In fact, they must also meet a growing demand as the number of people qualifying for Government-provided health services swells.

AI can bridge the gap, streamlining processes and reducing personnel needs while improving the quality of care for patients. Here are four ways AI can help Government agencies right now.

1. Detecting Errors and Fraud in Medical Claims  

CMS is the largest single Healthcare payer in the U.S., covering 75 million people through Medicaid and nearly 60 million through Medicare. In 2017, combined payments for medical services received by Medicare and Medicaid recipients totaled more than $600 billion. As many as 10% of these claims may contain billing errors or fraudulent charges, costing CMS tens of billions of dollars in overpayments each year.

The volume of claims—and the complexity of medical billing—makes it impossible for humans to thoroughly review each claim and detect potential billing problems. But this is exactly the kind of problem that AI excels at solving. For example, NCI’s SmartDoc intelligent automation system processes millions of pages of medical records, classifies documents by type, and extracts key information for human review. This vastly improves productivity and accuracy in fraud or error investigations. AI systems can also be trained using machine learning to detect and flag signs of possible errors or potential fraud and abuse. Combined, these applications could save tens of billions of dollars annually.

2. Improving Patient Care and Customer Service

Agencies such as the VA, that provide direct patient care, can tap into the power of AI to reduce the cost of services, improve response times and provide better care for the people they serve. Chatbots are already widely used in the business sector for rapid response and screening inquiries via websites. This same technology can help patients connect with the resources or people they need, reducing workforce need and improving patient experience.

Sophisticated, AI-based chatbots can address a range of patient problems and inquiries, ranging from billing questions to appointment requests, to medication renewals. Simple issues—for example, setting an appointment time or pointing a patient to educational resources about their condition—could be resolved completely. More complicated issues, such as billing disputes, would be escalated to a human for further review and response. An intelligent customer service bot would be able to anticipate and gather the information needed to resolve the issue (such as medical files, billing records and basic patient information), providing a seamless transition for the patient between the chatbot and a human representative.

3. Evaluating the Efficacy and Value of Treatments and Services  

What happens after a drug, device or procedure gets put into practice in the market? Real-world data—including metrics such as adverse events or reactions, hospital readmissions and health outcomes—provides important post-market insights into drug and device efficacy and safety. AI can comb through millions of data points from Electronic Health Records (EHRs) and hospital quality data to find patterns that aren’t apparent to the human eye.

Instead of waiting years to discover that a new therapy is creating adverse outcomes, machine learning can quickly identify problems and predict which groups of patients may be at risk. This post-market data would support FDA and other Government agencies making better decisions when recommending recalls and withdrawals or issuing safety warnings. Agencies such as CMS also have the opportunity to  better evaluate the costs and benefits of various treatments to support value-based purchasing and coverage decisions.

Machine learning allows providers of direct patient care to proactively identify patients at risk of readmission or adverse events and recommend interventions. Agencies involved in accreditation or hospital quality monitoring can better identify institutions where patient outcomes are better or worse than average and find patterns that help them understand what is driving these outcomes.

4. Remote Patient Monitoring and Mobile Health

Telemedicine is in its infancy, but AI can bring it into the mainstream, helping agencies like the VA, which supports a large and widely distributed population, better serve their constituents.

New home diagnostic and monitoring tools enable doctors and patients to work together remotely to gather the data doctors need to provide accurate diagnosis, effective care and long-term monitoring of health conditions. AI reduces communication barriers, providing real-time translation for doctors and patients who do not share a common language.

With AI, doctors can quickly analyze patient data for faster and more effective diagnosis. AI-based diagnostics look for patterns in patient symptoms and test results, compare those patterns to millions of other patient profiles to identify risk factors for disease and suggest possible diagnoses that may otherwise be missed. This enables more effective care when doctors do not have the luxury of extended in-person time with a patient.

Preparing for the Future of AI-Based Health Services  

Over the next five years, we can expect to see AI and machine learning become more integrated into every aspect of medicine, from drug and device development to long-term patient monitoring. Machine learning and advanced algorithms will help the industry reduce development timelines while improving risk analysis, bringing effective treatments to market quickly and safely. They also have the potential to help Government agencies make better regulatory decisions, streamline services, and better use taxpayer money.

Supporting these emerging technologies will require changes in regulation, research and development protocols, provider behaviors, and agency processes. But the technologies themselves are here today, and the time to understand how to incorporate them into our Healthcare systems and agencies is now. A healthier, more efficient future enabled by AI is just around the corner.

About the Authors

Brett Fraser, Vice President of Artificial Intelligence, is expanding NCI’s AI offerings to its government and commercial clients. Brett develops and drives NCI’s AI technology market strategy to help clients realize the benefit and value of AI solutions. Prior to joining NCI, Brett was the director of automation for Booz Allen Hamilton where he launched their RPA service into the federal market. He successfully developed automation programs for the National Science Foundation, the Veterans Administration and managed pilot efforts for the FDA, Navy and Immigration Services and Booz Allen.  Brett was the director of sales engineering for the federal program at IPsoft, where he launched several Department of Defense and federal pilots focused on Amelia, the artificial intelligence virtual agent.  While global head of video services at Cisco Systems, Brett pioneered the use of IT RPA to fully automate more than 90 percent of Cisco’s internal IT processes.

Michael Ezzo is the Director for Software Engineering for AdvanceMed, part of NCI’s Agile and Analytics (AAS) Business Unit. He is directly responsible for the design, development, and implementation of software solutions to maximize business value for NCI federal customers.  Prior to joining AdvanceMed in 2009, Mr. Ezzo worked for CSC since 1992 in a variety of leadership positions, his last assignment as Program Manager for CSC’s Global SAP Supplier Relationship Management (SRM) System.  Prior to CSC, Mr. Ezzo served as a commissioned officer in the U.S. Army.  



Please enter your comment!
Please enter your name here