Three questions with Jason Burke: Crucial next steps in the evolution of CDI

I sat down with the 3M Health Information Systems VP of revenue cycle solutions Jason Burke to talk about the evolution of clinical documentation integrity (CDI), including trends and ideas for continued success in the CDI space. 

From your perspective as a health care revenue cycle expert, where is the world of CDI going? What do you see as the most crucial next steps in its evolution? 

Clinical documentation integrity (CDI) will continue to move toward automation. One crucial next step is broader adoption of front-end documentation technology that helps physicians capture a complete and accurate patient story while they’re with the patient. What we’re seeing is an accelerated shift from a mid-revenue cycle CDI approach (with follow-up queries) to a front-end, point-of-care approach powered by artificial intelligence (AI) technology. The benefits are clear: When physicians engage upfront through real time nudges and recommendations, they learn to close documentation gaps early, minimize retrospective queries and avoid rework downstream. The goal is fewer nudges and queries because physicians learn from AI-generated nudges to capture the right information from the start. 

A closely related step forward is standardizing how documentation is captured. By this I mean how the physician describes the diagnoses, procedures and orders—all of the discrete clinical data that AI models use to determine what feedback physicians need. Many health care organizations now use templates to standardize how their physicians document cases. The more standardized it is, the easier it is for machine learning and deep learning models to produce meaningful output—more accurate recommendations, inquiries or nudges, delivered at the right time, in a compliant way.  

Another crucial factor in the evolution of CDI is increased trust in technology. Physicians are more technologically savvy and willing to accept input from AI software and act on it than they used to be. But you need to demonstrate success quickly to get physician buy-in.  

At 3M, we’ve learned through many years of experience working with health systems, that an all-at-once roll-out approach doesn’t work with natural language understanding (NLU) models. You need a deliberate, tactical approach focused on top priority service lines and/or clinical conditions. This differs between organizations. We often recommend starting with targeted services lines such as internal medicine, neurology, pulmonology and cardiology. Some of the top real time nudges we recommend include anemia, encephalopathy, respiratory failure and heart failure. The point is you need to demonstrate success out of the gate. Otherwise, you lose the confidence of both physicians and CDI leaders.  

Electronic health record (EHR) companies are adding more CDI functionality to their solutions. How should CDI leaders respond to this trend?  

EHR vendors want to create an integrated experience, so their users don’t have to leave the EHR workflow to launch separate systems. For years, we have worked closely with our EHR partners to integrate our coding and CDI workflows to automate routine work, prioritize cases with the most opportunity and expand to all payer and quality reviews. These specific 3M workflows can be launched from within the EHR so that data can be shared seamlessly between the systems.  

The most crucial aspect of this collaboration with EHRs is the clinical content expertise 3M provides. We have spent more than 40 years building and refining the clinical content needed to ensure best practice outcomes. The depth of our clinical content leads and guides users through coding and CDI workflows. The clinical data helps users identify and close gaps in the documentation. Sophisticated prioritization models help health systems focus on high priority cases based on clinical criteria.  

Yes, EHR vendors are adding more CDI functionality, but they still want to provide access to 3M’s expert clinical data and automation software—powered by our NLU and deep learning models. Approximately 2,800 U.S. health care facilities send data through 3M’s AI-driven NLU engine daily. 

CDI leaders recognize that this combination of clinical content expertise with a streamlined technology workflow is the best solution.  

What questions should CDI professionals ask themselves about the range of technology and services they will need to improve and sustain success?  

Health systems need to find the right balance of technology and services. CDI leaders look to both of those things to help them drive consistent outcomes. Documentation needs to drive consistent coding and reimbursement, and accurately represent quality of care.  

To accomplish this efficiently and sustain success over time requires a combination of technology and services—including access to clinical intelligence, automated workflows and expert advisors who can help connect the dots from care delivery to coding to final billing. If there’s a gap in any one of those areas, the codes won’t correctly represent the care given. Expert consultants can also help with data management, measuring and reporting on performance and recommending new strategies to match health systems’ top priorities.  

Some health systems take a best of breed approach to technology, choosing the best software for each process or workflow—coding, CDI, denials management, etc. Others choose an EHR, a billing company and a partner like 3M that can standardize clinical content and provide a consistent NLU model throughout the revenue cycle process.   

What we’re seeing from our experience with clients is that it’s a significant advantage to use the same AI model across systems. Otherwise, you’ll have conflicting outcomes. For example, it can become chaotic if you have many different vendors in your revenue cycle process—each with their own NLU technology—interpreting the documentation and adding insights in different ways. The recommendations and systems simply won’t be aligned.  

Our strategy at 3M is to provide clients with a consistent NLU model to process clinical, quality and financial information in the same way. From initial patient care where the documentation starts, and even registration, down to coding correctly, we’re using the same approach, the same clinical content, the same NLU models, which gives health systems consistency that drives improved outcomes.  

Jason Burke is the vice president of revenue cycle solutions at 3M Health Information Systems.  

Steve Cantwell is a senior marketing communications specialist at 3M Health Information Systems.