Fixing the $31.2B Gap: How AI Can Help Cut Costs and Improve Accuracy

COMMENTARY

Every year, billions of taxpayer dollars vanish into the cracks of Medicare’s complex claims system. In 2023 alone, improper payments accounted for a staggering $31.2 billion — a financial strain exacerbating an already stressed health care budget.

These errors aren’t mere numbers, they reflect inefficiencies that burden providers, inflate patient costs and erode public trust. 

As the health care system faces daunting challenges, addressing high government spending, policy changes, access to care and workforce shortages, it is clear that transformative solutions are needed. Artificial intelligence presents a powerful opportunity to reduce these errors, enhance data accuracy, and streamline claims processing — paving the way for a more sustainable and transparent health care future.

Behind the $31.2 Billion Loss

The staggering amount isn’t just a financial concern, it reflects the systemic challenges within medical coding and documentation. These errors are not merely administrative glitches, they also impact hospital revenues, delay patient care and inflate health care costs. 

The shortage of skilled medical coders exacerbates the problem. The American Medical Association reported a 30% gap in these roles, with 50% of chief financial officers struggling to fill the positions, as many medical coders are near retirement and few replacements are trained to handle the complex, evolving coding standards. Experts say there should be at least one coder for every 10 physicians, a reality that many health care centers can’t adhere to.

This shortage doesn’t just create administrative headaches, it triggers a cascade of negative effects that ripple through every level of the health care system:

• Delayed Reimbursements: Errors and backlogs in medical coding slow down claims processing. Twenty-six percent of health care providers reported worsening medical coding due to staffing shortages, leading to delayed payments and increased financial pressure on their organizations.
• Increased Administrative Costs: Hospitals must allocate additional resources to correct coding errors, conduct audits and manage appeals, driving up overall administrative expenses.
• Reduced Care Quality: Staff shortages force medical coders to handle heavier workloads, increasing the likelihood of errors. Over 35% of rejected claims result from basic data entry errors, leading to incorrect patient records and impacting treatment decisions.
• Challenges in Adapting to New Standards: As the industry transitions to complex systems like ICD-11, the shortage of trained coders raises concerns about the smooth implementation of new standards, potentially exacerbating coding errors. Yet only 11% of provider organizations have begun preparing for this shift, making the need for action more urgent.
• Increased Health Care Costs: Coder shortages lead to inefficiencies that upscale health care costs, as hospitals pass on administrative burdens and revenue losses to patients and insurers. Recent data shows that administrative costs now make up more than 40% of total hospital expenses.

How AI Can Bridge the Gap

AI models offer significant potential to streamline medical coding by automating routine tasks and reducing human error. They can quickly analyze vast amounts of data and identify patterns, improving efficiency in claims processing. 

However, AI is not infallible. Recent studies show that even advanced AI models struggle with tasks such as accurately reproducing medical codes, often with error rates exceeding 50%.

This underscores the critical need for human oversight to ensure reliability and accuracy — a gap effectively addressed by human-in-the-loop machine learning.

Real-world cases demonstrate the benefits of this collaboration. Health care systems that combine AI-driven tools with expert oversight have seen coding accuracy improve dramatically — up from 85.5% to 98% — reducing errors and increasing efficiency. By leveraging AI’s processing power alongside human expertise, providers can address coder shortages, enhance accuracy and lower administrative costs, creating a more sustainable and efficient health care system.

While AI is not a complete solution on its own, its ability to complement human expertise offers an effective way to bridge the gap in the medical coding workforce. By embracing AI-powered tools to support coders, billers and auditors, health care organizations can address ongoing challenges. With the right balance of technology and human oversight, the health care industry can move toward a more accurate, efficient and cost-effective future.

Building a Data-Driven Future Through Policy

As the health care industry increasingly turns to AI for medical coding and claims processing, policies must ensure data governance is transparent and accountable. 

Establishing clear frameworks for how medical data is used by AI systems can help maintain trust while also addressing concerns around privacy, bias and accuracy. Strong regulations should ensure that AI models are assessed and refined, minimizing errors and fostering a continuous improvement cycle. This ongoing assessment is necessary to adapt to evolving medical standards, such as the shift to ICD-11, and to handle new challenges that may arise as AI tools become more deeply integrated into health care processes.

Policymakers must prioritize the creation of standards that govern data usage, ensuring AI systems are tested rigorously and consistently. Transparent governance will not only protect patients but also improve health care efficiency and reduce costs. Regulations should encourage collaboration between AI developers, health care providers and regulatory bodies to build systems that can evolve as the landscape changes. Focusing on the HITL/ML model ensures AI systems are always used with human expertise to verify and correct errors. 

As technology evolves, policies should adapt to ensure AI continues to serve the best interests of both patients and health care providers. Ongoing assessments should be institutionalized within health care regulations, ensuring that AI tools develop along with the demands of the industry. By implementing regular oversight, health care organizations can enhance the reliability of AI systems and ensure their long-term sustainability.

Data governance frameworks emphasizing transparency and continual refinement can help build trust in AI, not only from health care providers but also from the patients they serve.

John T. Bright is a distinguished health care technology executive and the founder and CEO of Med Claims Compliance Corporation. With over three decades of experience, he has driven the development of innovative medical claims processing systems, including VetPoint™, CliniPoint™ and RemitOne™. His deep expertise spans EMR systems, medical device sales, FDA 510K applications and health information standards. He can be reached on LinkedIn.