Artificial intelligence (AI) in health care can have a high return on investment — not by taking the place of humans, but by enabling them to work at their highest level of certification. It can allow doctors to focus on patients, auditors to focus on assessing finances, and medical coders to focus on accurately reflecting the services provided.
Natural language processing (NLP) is a form of AI used in many industries to distill data out of electronic documentation (as my colleagues have already written about here and here). In health care, that data is most commonly distilled from medical records.
While electronic health records (EHRs) provide a rich repository for a person’s medical history, these files contain a variety of clinical notes that don’t fit neatly into a database. They also contain results in varied forms from multiple care providers — all without any overarching filing or naming conventions to present the information uniformly.
It’s like when you need to clean your desk in a hurry, so you grab a manila folder and fill it with all the loose papers relating to a project you’re leading. Then, when you need to find a specific recommendation you remember reading, you spend a lot of time looking through everything in the folder only to find the information was scribbled on the back of an unrelated document. NLP can reduce that wasted time and energy by structuring data into something more usable.
As a result, employees in health information management (HIM) can start working more efficiently and help contribute more to the bottom line. Since computers are great at multi-tasking and don’t get tired, they can refine the workload for HIM professionals, as well as auditors and expert reviewers.
Clinically aware NLP, built specifically for health care, goes much further than organizing data. It can help improve it. While other NLP may understand medical terminology, clinically aware NLP understands context. It finds gaps in documentation and provides feedback to clinicians and administrators to make sure records are clear and consistent, and final conclusions are stated.
Clinical documentation improvement (CDI) is a great example. While most provider organizations have some type of CDI program in place, inaccurate and incomplete clinical documentation still causes providers to lose earned revenue. Using technology with clinically intelligent NLP can transform their CDI programs. For example, the CDI software is able to review 100 percent of a health system’s records to help uncover cases with the greatest likelihood of improvement across a patient’s hospital stay. Further, providers can ensure accurate coding on payer claims by linking this CDI technology with advanced computer-assisted coding (CAC) platforms that also leverage NLP to identify meaning and context in clinical data.
The result? These highly skilled people spend the greatest portion of their days on their core functions, not on scouring countless records for very specific pieces of information that may or may not be there. That’s how AI provides a return on investment — by efficiently serving up the data that all health care professionals need to make decisions and take action.
About the Author
Mark Morsch, vice president of technology for Optum, is an innovation leader with a focus on developing products using natural language processing and other forms of artificial intelligence to help transform our health care system. Guided by the goal to deliver the right information at the right time to the right audience. Inspired to build world-class development teams that can drive innovation and surpass client expectations through creativity, engagement and hard work.