AI is rapidly transforming coding operations in musculoskeletal (MSK) care. At Becker’s 22nd Annual Spine, Orthopedic and Pain Management-Driven ASC + The Future of Spine Conference, leaders from Veradigm and aiHealth unpacked the technology’s promise, adoption trends and how to avoid missteps in a high-stakes regulatory environment.
The panel featured Brian Hall, vice president of sales strategy at Veradigm; Ahmar Qasim, area vice president at Veradigm; and Kyle Swarts, president of aiHealth.
Here are four key takeaways:
1. Staffing shortages drive adoption
Health systems are increasingly turning to AI coding tools in response to coder shortages and tightening margins. Mr. Hall noted that while technology drives efficiency, it also enables coders to operate at the top of their skill set.
“The AI side of it has gone from a few miles an hour to a billion miles an hour,” Mr. Hall said. “It’s helping drive improved coding, improved clinical documentation and helps the coders augment their work and let them do some of the heavy lifting.”
Mr. Qasim echoed that point, noting that labor issues are the driving force behind most of Veradigm’s current conversations with clients. However, panelists agreed that organizations must address foundational problems before layering on AI — including cleaning up practice management systems, loading payer contracts and setting clear performance benchmarks.
2. From validation to value
Mr. Swarts noted that the industry is shifting from curiosity to practical implementation of AI. This requires leadership buy-in, workflow readiness and financial commitment.
To facilitate implementation, he outlined a phased approach: the first 30 days focus on validation. By days 90 to 120, a subset of high-volume CPT codes — such as total hips and knees — can reliably go direct to bill.
He emphasized that human coders remain essential. Keeping humans in the loop enables models to learn and ensures continuous improvement.
“There’s going to be exceptions, which is where you need expert coders,” Mr. Swarts said. “I would highly encourage you not to get rid of coders, but find other roles for them inside the organization. Your coders become reviewers rather than composers. You repurpose those staff to provide education and critical documentation improvement.”
3. Vet your vendor
Panelists urged healthcare leaders to be discerning when evaluating AI partners. Specialty-specific experience, integration with EHR systems and data quality are all critical.
“The first thing you need to understand and make sure is that they have experience in your specialty and understand if they can drive some of the coding results there,” Mr. Hall said. “You’ve got to really understand them before you jump in.”
He added that AI products are not one-size-fits-all. Understanding how tools integrate with existing infrastructure remains a key hurdle.
“That’s probably the single biggest challenge,” Mr. Hall said. “It’s a key challenge early on for growth, integrating with your system, whether it’s Athena, Epic or another.”
4. Future AI guardrails
Even as AI automates documentation and coding, the legal burden remains with physicians. Providers must audit AI-coded data and ensure their business associate agreements (BAA) reflect current uses of de-identified information.
“Data is king, your data is key, and how are we going to use that data under the rules and guidelines of the business associate agreement?” Mr. Swarts asked. “I will argue that the BAA of the past do not really support the ability to use de-identified data and such under the HIPAA guidelines.”
Maintaining human oversight and auditing AI-driven workflows are key to long-term success.
“Start with a simple half-page policy of ‘here’s what we’re going to use AI for,” Mr. Qasim said. “I’d say in 2025, it’s time to have an AI policy in place.”
The post AI coding in MSK care: 4 takeaways on implementation, pitfalls and potential appeared first on Becker’s ASC.
