With a surge in claims denials overwhelming many hospitals and health systems, it’s more important than ever for providers to understand what options are available to them in their denials management strategy, and how their denials workflows can be improved to file appeals as quickly and accurately as possible. After all, according to recent data from the American Hospital Association, 50% of all hospitals and health systems have more than $100 million in accounts receivables for claims that are older than six months, while 95% report increases in staff time spent seeking prior authorization approval.
Addressing these challenges requires innovative solutions. Artificial intelligence (AI) has emerged as the most significant development by far—representing a sea change in how Revenue Cycle Management (RCM) is performed. By leveraging AI-powered denials workflows, healthcare providers can optimize RCM efficiency, accelerate revenue collection, and reduce operational costs.
“We are seeing a massive increase in the number of denials and the speed with which the denials are being sent back by the payers,” says Aspirion Chief Strategy Officer Jim Bohnsack. “I actually believe that they’re not really applying different rules but that they’re applying them faster and more consistently with the use of technology. So, they’re outpacing the provider community in terms of their investment in technology.”
How AI Fits into the Healthcare Landscape
Large Language Models (LLMs), such as ChatGPT, are advanced machine learning (ML) algorithms designed to interpret and generate complex text from simple prompts. These models, trained on extensive internet data, can be tailored for specific domains such as healthcare. These models are increasingly being applied across various RCM functions, offering opportunities for optimization and efficiency improvements.
RCM functions offer a wide array of opportunities for AI application throughout the revenue cycle:
Front-End
- Scheduling Optimization
- Prior Authorization Automation
- Patient Chatbots
Mid-Cycle
- Clinical Ambient Listening
- CDI Clinical Documentation Improvement (CDI) Queries
- Coding Automation
Back-End
- Patient Pay Modeling
- Denial Modeling
- Appeal Automation
Aspirion Use Case
One of the areas where AI is making a significant impact is in the drafting of appeal letters for clinical denials, a process that traditionally requires multiple inputs completed by high-cost, specialized resources. Aspirion has developed refined processes for applying AI to optimize denials management. For instance, here are some of the AI-enhanced workflows for clinical appeals:
- Sub-step: Clinical denials requiring clinical validation for appeal generation
- Inputs: Clinical documentation, clinical guidelines, payer policies, managed care agreements
- Output: Detailed appeal letter with evidence from clinical documentation
- User(s): Clinicians, certified coders, lawyers
Aspirion Chief AI Officer Spencer Allee says it’s important to focus on where specifically a company will apply ML technology. “Throughout my career in machine learning and AI, there has often been this kind of myth or desire to—I call it ‘sprinkle machine learning on top of it and hope that it works.’ But that doesn’t really work. It’s all about where in the flow of the business process today do you actually intervene and interject, and what are the inputs and the outputs to that?”
Allee said he started with clinical appeals specifically because there is so much information and text bound up in that process, as well as being time consuming and complicated. It was the exact type of use case where solutions powered by large language models can make a real difference
Machine Learning Platforms in Denials Management
An innovative machine learning platform, such as Aspirion’s DocIQ, can be a game-changer in RCM—and especially in denials management. Here are some capabilities of advanced machine learning platforms:
- AI-generated summaries based on denial type with citations
- Fully digitized and searchable document for easy navigation
- Automated extraction of document taxonomy and bookmarks
AI is enabling a significant improvement in team effectiveness by accelerating throughput and increasing quality simultaneously. Achieving both requires thoughtful AI deployment and workflow design. Aspirion uses a comprehensive measurement framework to track the impact of AI-powered denials workflows:
- Model Metrics; Accuracy/LLM Fidelity: ~95%
- User (Operator) Behavior; Clinical Review Step Efficiency: 20-30% Improvement
- Operational KPIs; Example: Increased Throughput Capacity
- Client Benefit; Example: Improved Time-to-cash
Beyond the Hype: Realizing the Impact of AI
Capitalizing on AI-powered denials workflows is no longer an option, but a necessity for healthcare organizations striving to optimize their revenue cycles. The question isn’t whether to adopt AI in RCM, but how to do it effectively. Aspirion’s approach demonstrates the tangible benefits of AI in appealing and preventing denials, optimizing efficiency, reducing costs, and increasing collections.
Bohnsack says providers contemplating the use of AI need to ask themselves the right questions, such as:
- Do we have the capital to invest in a journey like this?
- Do we have the capability, with a plethora of resources at our fingertips?
- Do we have the capacity to work on revenue cycle management optimization?
- Do we have the commitment, knowing it’s not just a silver bullet with magic results, but rather a total journey?
“But do we have a choice?” Bohnsack asks. “Payers are investing in this at a much faster clip than we are as providers.”
Want to learn more? Tune in to our Level Up Revenue Cycle on-demand webinar, “AI Results and the Provider Perspective.” .
