It’s becoming abundantly clear that without technology, health systems will simply be unable to meet growing patient demand. AI is already showing real promise by taking on repetitive and time consuming tasks like analyzing medical records or handling medical paperwork. This is of great value to frontline clinicians, but it’s just as vital in other day-to-day health provider operations.
Take revenue cycle teams. They are grappling with a growing volume of claims, and a claims process that is getting more complicated, more expensive and more laborious. This is coming at a time when federal support for Covid is coming to an end, operating costs are spiraling and many senior, experienced revenue cycle team members are exiting the market. Reimbursement has never been more critical or more complex.
Large health systems are already beginning to embrace AI to automate clinical document improvement (CDI) and denials management, saving valuable staff time and helping to more effectively ensure reimbursement for the work they do.
Boosting revenue cycle teams
Modern AI approaches apply a highly specialized kind of AI called Clinical Natural Language Processing (CNLP), a linguistic approach that can understand billions of different word and phrase combinations relating to hundreds of thousands of detailed clinical concepts. This enables CNLP-based tools to understand the clinical free text, or “narrative”, which makes up a large proportion of the electronic medical record, and is the only place where the detail – essential not just for diagnostic and screening purposes, but also reimbursement – is recorded.
As healthcare gets more complex, so does the work needed to ensure that patient care is properly documented and fully reimbursed. Revenue cycle staff find themselves having to manually review the dense, unstructured clinical narrative, which can run to hundreds of pages per visit. Enter CNLP-led AI technologies which can rapidly read and understand the entire medical record, quickly locating the supporting clinical evidence needed to strengthen hospitals’ clinical documentation, utilization review and appeals management workflows. These tools can process millions of documents in hours instead of days.
The arrival of generative AI
Generative AI potentially takes AI’s capabilities further still by introducing the ability to create clinical narrative. While the technology has driven concern, as it matures, generative AI is likely to have a positive impact in healthcare when applied to well-defined, specific problems such as CDI.
For example, Microsoft’s recent offering – which combines its transcription solution and ChatGPT – not only automates the transcription of, for example, dictated discharge summaries, visit summaries and radiology reports, but also via ChatGPT offers suggested improvements or flags important missing information. This is likely to be very popular with busy physicians, saving them precious time and helping to ensure medical records contain the most salient information. This in turn could help deliver AI-enabled CDI at source.
Prioritizing user experience
Regardless of how effective or ground-breaking the AI is, delivering technology-enabled change at scale in a complex operating environment like healthcare is never easy. In the recent past, front-line and back-office staff have been exposed to multiple generations of complex and hard-to-implement tools that have often added to workloads. This has, understandably, created significant resistance to new technologies.
An empathetic approach has to be taken for technology-enabled transformation to be successful. The technology must address a clearly defined problem and should significantly benefit its frontline users. One of the most exciting aspects about the latest generation of AI-enabled tools is its potential to do just that.