A physician’s most valuable resource is time. Yet today’s reality is sobering: on average, doctors spend over three hours each day on documentation. Most concerning? Much of this labor happens during “pajama time” – those late hours when physicians catch up on paperwork from home instead of resting or being with family. Besides the impact to physician satisfaction, the cost of this administrative burden is staggering: Organizations spend $82.7 billion annually on documentation, coding, and other administrative tasks, straining both budgets and staff.

Throughout the conference circuit this year, health system and physician group leaders have been gathering and discussing ways to make a dent in this burden. While various solutions have been proposed, one approach is gaining traction: autonomous medical coding. But before we explore this AI technology, we need to understand a puzzling dynamic: Why are physicians performing any coding in the first place?

The current state of coding

Physician coding stems from both tradition and necessity. First, in many small practices and departments within larger systems, providers have been used to handling their own coding – a norm that persists largely through institutional inertia. Today, this is especially common for primary care, internal medicine, and similar specialties. While having physicians tackle their own coding may have made sense in the past, it’s increasingly stood out as an unnecessary burden in the context of a broader ballooning of administrative duties on physicians.

Second, given constraints on coding labor – for example, a recent study estimated a 30% shortage in certified medical coders – many clinicians have been asked to bridge the gap themselves. Although physicians are highly trained, they are not prepared to be experts at coding in medical school. As a result, this seemingly pragmatic approach creates downstream problems as coding errors cascade into denied claims, delayed reimbursements, and costly rework.

At the same time, the complexity of medical coding has grown. Today, there are over 69,000 ICD-10 diagnosis codes, over 10,000 procedure codes, and dozens of other coding elements that must be precisely determined for each patient encounter. This increasing challenge strains traditional manual processes, with even experienced coders requiring weeks or months to adapt to guideline changes. For physicians trying to keep up, achieving coding accuracy benchmarks becomes even more difficult.

How to get out of this spiral? Some organizations have tried the obvious path: expanding human coding teams. But this approach presents its own set of challenges. For one, training new coders requires months of education and testing, during which organizations must absorb both the training costs and reduced productivity. And what about offshore teams? While they might seem like a cost-effective solution, organizations typically discover hidden costs in the form of higher error rates and increased overhead required to stay on top of quality. What starts as a cost-saving measure often becomes more expensive and riskier than anticipated. 

The AI alternative

This is where technology enters the picture. Unlike its rule-based predecessors that merely suggest codes for users to validate, today’s AI can fully automate coding for most encounters at high accuracy, consistency, and scale. And to keep up with the changing regulatory landscape, it can adapt to guideline changes almost instantly, avoiding painful and costly ramp periods.

Many clinical leaders understandably express skepticism about AI coding at first. Common concerns include accuracy on complex cases, maintaining compliance standards, and the impact on existing coding teams. These are important considerations – after all, coding accuracy affects both reimbursement and patient care. However, the data is compelling: early adopters have found that strong AI can actually reduce coding errors, which currently cost the industry $10.6 billion annually. As a result, claims denials, typically taking 90+ days to resolve, decrease dramatically – and so does staff time spent on appeals.

For skeptical leaders, the evidence is compelling and reassuring: AI technology maintains or exceeds compliance standards while reducing operational costs and complexity. The AI approach to coding finally frees physicians from one administrative burden that pulls them away from patient care.

A necessary change

For healthcare leaders, AI coding is a powerful solution to a growing crisis around physician burnout and retention. Simply put, physicians shouldn’t be coding, the complexity of coding continues to increase, and traditional solutions aren’t working. As organizations grapple with rising error rates and physician burnout, AI offers a clear path forward: better coding accuracy, reduced clinician burden, and the opportunity to expand clinical focus where it belongs – on patient care.

Credit: smolaw11, Getty Images

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.