From CAC to autonomous, AI heralds medical coding's future
AI is pushing medical coding to its next era, in which autonomous coding can scale across all specialties, automate up to 90% of cases and drive revenue integrity.
The future of medical coding is here, at least according to two technology experts who are leveraging AI to transform coding workflows from not only computer-assisted but also autonomous.
"We see AI autonomous coding as the next evolution because of its ability to interpret more data -- discrete and non-discrete -- to get to higher levels of quality and outcomes, as well as faster turnaround times," explained one of those experts, Andrew Ray, chief innovation officer at Ensemble Health Partners.
Applying AI to autonomous coding can bring about unprecedented scale and automation, with the other expert, Jason Burke, vice president of revenue cycle solutions at Solventum, estimating up to 90% automation for outpatient coding and 70% automation for inpatient coding within the next two years. The latter being a notoriously tricky area for effective, scalable autonomous coding.
"The innovation is really around automating all coding -- inpatient, outpatient and professional -- to minimize the human interaction," Burke said.
Together, Ray and Burke's companies have partnered to usher in this next era of medical coding, clinical documentation and revenue integrity. They aim to create an end-to-end autonomous coding solution driven by AI that supports all medical specialties, including hospital-based inpatient claims, while still hitting quality and compliance.
Evolution of coding technology
Computer-assisted coding (CAC) was a turning point for medical coding. These solutions use predefined coding rules and guidelines to automate and streamline coding workflows in the face of growing complexity and the need for enhanced accuracy and efficiency.
The CAC market has matured and grown over the last 10 to 15 years, Ray stated. However, he asserted that healthcare has "advanced beyond the rules set."
Autonomous coding has enabled healthcare organizations to move past strictly rules-based solutions. Some autonomous coding solutions may incorporate rules-based logic, but their innovation lies in their ability to analyze clinical data to determine appropriate codes with minimal human intervention. In contrast, CAC aims to support human coders.
Healthcare organizations have realized the benefits of autonomous coding, including greater coding workflow automation, the ability to scale output easily, fewer human errors and faster chart and claim processing.
Still, autonomous coding has its own challenges. For example, solutions can struggle with navigating the complexity and ever-changing rules and guidelines in coding and billing, especially in complex cases.
However, advancements in AI and logic behind autonomous coding may be able to overcome these challenges and more.
"The core AI technology itself, with its ability to ingest various forms of discrete and non-discrete data, process it very efficiently and make sense of it, is a huge unlock in potential," Ray said. "Its continual learning nature also moves us from a rules-based system to one that can learn, and not just from the clinical documentation and the coding, but the downstream impacts of what gets denied and why it gets denied."
How AI will transform autonomous coding
AI is pushing autonomous coding to the next level by addressing the complexity within coding and billing to drive coding efficiency and accuracy. In particular, the technology can help healthcare organizations overcome the difficulties of applying autonomous coding to inpatient care.
Coding for inpatient cases is more challenging than outpatient or physician care. Documentation tends to be far more extensive and intricate to cover a patient's journey that can last days to weeks, involving multiple diagnoses, comorbidities and various services.
Medical coders need a deep understanding of not only coding requirements but also medical language for a more contextual understanding of an inpatient stay. AI can struggle to deliver this level of contextual understanding to fully capture the patient's journey and translate it into billable charges.
AI is helping us to minimize that complexity, and it will continue to do that.
Jason Burke, VP of revenue cycle solutions, Solventum
Autonomous coding that can integrate professional and facility coding using AI is driving innovation in inpatient care. It can also escalate complex or unclear cases to a human coder to manually review AI's suggestions, make edits and send feedback to the AI model to improve in future cases.
But Ensemble and Solventum intend to combine the former's robust client base (28 health systems, representing about $40 billion in net patient revenue) and AI capabilities with the latter's autonomous coding solutions and coding compliance expertise to make autonomous coding work for inpatient, outpatient and beyond.
Autonomous coding solutions have largely been verticalized, targeting a specific care setting or specialty. However, healthcare organizations are looking for more comprehensive autonomous coding to truly drive coding efficiencies and billing accuracy, especially as volumes increase, complexity grows and staffing shortages worsen.
"AI is helping us to minimize that complexity, and it will continue to do that," Burke said.
AI-driven autonomous coding can analyze clinical data from across a hospital more efficiently than individual coders. The feedback loop through audits and rule-based checks also ensures AI models are constantly learning and improving coding practices.
From coding to revenue integrity
AI-powered, scalable autonomous coding isn't just for improving coding accuracy and efficiency, although the technology is doing that. These solutions are just as much about supporting clinical documentation improvement (CDI), compliance and denials prevention, Ray and Burke explained.
"Revenue cycle is evolving, and we think of it more as revenue integrity," Burke stated.
Revenue integrity focuses more on payment accuracy, compliance and ethical standards compared to more traditional revenue cycle management approaches, according to the National Association of Healthcare Revenue Integrity. As such, it homes in on appropriate documentation and financial practices that can withstand audits and prevent revenue leakage.
AI autonomous coding is a natural first step in smarter revenue integrity. But these technology experts are taking a more longitudinal approach to AI transformation as it relates to coding.
"Coding is historically vertical, but there are the upstream pieces like documentation, then downstream pieces like denials and aligning payment," Ray said. "AI capabilities allow us to create reinforcement learning and feedback mechanisms through all those pieces that have historically been siloed and difficult to crack through."
In other words, AI models can identify patterns in denials that may stem from coding practices, then surface those insights to coding teams to correct through CDI, for example. This can also happen closer to real time thanks to advancements in AI, Burke suggested.
"AI application for coding is still only as good as the input, and the input in this case is clinical documentation," Ray added.
Healthcare organizations should have more resources to focus on CDI efforts with effective AI autonomous coding throughout the enterprise, too. After all, this technology frees up staff to focus on revenue integrity, particularly CDI and physician training.
"There's a new world possible that we've not really contemplated previously in terms of the completeness, quality and compliance outcome side of this, as well as where some of those resources can be poured even more heavily into the clinical documentation upfront to further enhance that," Ray said.
This all plays into the "big picture vision for the future of coding," Ray continued.
"With improved coding quality outcomes and turnaround times for completion of clinical documentation and coding, there is an acceleration that'll drive provider revenue, getting claims out the door and getting them out the door cleanly."
Jacqueline LaPointe is a graduate of Brandeis University and King's College London. She has been writing about healthcare finance and revenue cycle management since 2016.
