Nurses Have Waited a Century for Billing Codes. AI May Get Them First

Part of Nurse.org’s Nursing AI Watch, our ongoing investigation into how artificial intelligence is reshaping nursing practice. In this piece we look at the money: the AMA is building billing codes that would pay hospitals for AI’s clinical work, while nursing, a century after being folded into the room charge, still has none of its own.

The American Medical Association is developing a new class of billing codes that would let hospitals get paid for clinical work performed entirely by software, with no physician involved at the point of care. They’re called Clinically Meaningful Algorithmic Analyses, or CMAA codes, and they could reshape how American healthcare pays for artificial intelligence.

Here’s what makes that remarkable for nurses: the nearly 1.9 million registered nurses working in U.S. hospitals still have no billing codes of their own, which means the payment system is now closer to paying directly for an algorithm’s clinical judgment than for a nurse’s. Nursing leaders warn that if AI becomes billable while nursing remains a cost, hospitals gain a financial incentive to buy more algorithms and no new incentive to hire more nurses.

And there’s a deadline attached: the AMA’s code-setting panel accepts written public comments through August 10, and a parallel CMS payment rule is open to public comment from any nurse through September 14 — the rules are being written right now, and there’s still time to weigh in. 

Want to see more Nurse.org articles in your Google results? Add us as a preferred source.

First, a 30-second translation of the billing system. CPT codes are the standardized codes that describe nearly every service in American healthcare, more than 11,000 of them, maintained by a panel convened by the AMA. They determine what hospitals and providers can charge for. If a service has a code, it can generate revenue. If it doesn’t, the payment system essentially can’t see it.

CMAA is a proposed new category of CPT codes for situations where an algorithm analyzes patient data, such as a scan, a lab result, or a heart rhythm, and produces a result that changes care, even when no physician does traditional work at the point of service.

Read the AMA’s own description of who the code set is evolving to serve. The panel says the new codes will help CPT keep pace with medicine “whether services are performed by physicians, other qualified health care professionals or solely provided by AI-enabled algorithms.”

Physicians. Advanced practice providers. Algorithms. Bedside nurses do not appear in that sentence.

“If the AMA can build an entirely new billing category for AI, it can build one for nursing. For decades, health systems have been told nursing work is too complex to unbundle nurses from the room rate and bill. However, the very process used to develop AI’s CMAA codes allows us to develop a CMAA equivalent for nursing. We must act now, for if we don’t, AI will be reimbursed for the very work nursing has long done but was never able to bill for. Nursing will remain a cost, AI will become a revenue stream, and the implications will be devastating to nursing.”

Rebecca Love

See also  Nursing’s own general election manifesto launched

RN, BS, MSN, FIEL, Founder Commission for Nurse Reimbursement

Here’s how the CMAA effort has unfolded so far:

• 2021: The AMA adds Appendix S to the CPT code set, a framework that classifies clinical AI as assistive, augmentative, or autonomous based on how much of the work the software does.
• September 2025: The CPT Editorial Panel holds its first formal discussion of the CMAA framework, aimed at services that don’t require physician work at all. The American College of Radiology confirms the pathway is under active development.
• January 1, 2026: The CPT 2026 code set takes effect with 418 changes, including new codes for AI-enabled services like coronary plaque analysis and algorithmic ECG reading. Early forms of AI billing are already live.
• May 2026: The panel accepts revisions to Appendix S that sharpen the definition of a “clinically meaningful output.” Health policy trade press reads the update as groundwork for CMAA codes that could eventually reach Medicare.
• August 10, 2026: Deadline for written stakeholder comments ahead of the next panel meeting.
• September 2026: The panel meets again, with decisions published by early October.

This is not a slow-moving hypothetical. The AMA says the panel is already receiving a steady increase in applications for services that rely entirely on algorithmic analysis, without traditional physician work at the point of care. The demand for a way to bill for autonomous AI is real, and the infrastructure is being built to meet it.

>>Listen to The Latest Nurse News Podcast

To understand why this stings, you have to understand how nursing gets paid for today. In most U.S. hospitals, it doesn’t, at least not directly. For about a century, nursing care has been bundled into the room-and-board charge as a fixed daily cost, priced the way a hotel prices housekeeping. Hospitals bill for the bed. The nurse comes with it.

The Commission for Nurse Reimbursement, a nonprofit working to modernize how nursing is paid for, put a number on that invisibility. Reviewing price transparency data from one academic medical center with more than 1,000 beds, the Commission found 158,475 individual billable line items and not a single inpatient nursing charge.

The scale of what’s missing is enormous. Hospitals are by far the largest employer of RNs in the country, and by the Commission’s analysis, nurses account for roughly 30% of hospital labor spending, about $266 billion a year. That entire contribution appears in hospital finances as a cost to be managed, not a service that earns revenue.

One important nuance: advanced practice nurses, including nurse practitioners and CRNAs, can and do bill under CPT codes. The gap is bedside RN care with a few exceptions: Oregon allows RN billing under Medicaid, and North Carolina’s Institute of Medicine has recommended RN billing codes. But as a rule, the work of the largest clinical workforce in the country has no presence in the billing architecture of American hospitals.

For decades, the standard answer to direct nursing reimbursement has been that it’s too hard. Nursing work is too continuous, too interwoven with everything else that happens at the bedside, too complex to break into billable pieces.

CMAA is the AMA solving essentially that same problem for software. According to the AMA, the proposed codes would describe services in which algorithms process clinically relevant data to produce “medically actionable outputs,” even when no physician is involved at the point of service.

Nurse executive and Love n’ Leary Nursing Podcast co-host Rebecca Love saw this coming. More than a year ago, when the CMAA proposal first surfaced, Love, founder of the Commission for Nurse Reimbursement, asked the obvious question in a widely shared LinkedIn post: “Why can we not bill these same codes for nursing?” Nursing had been told for decades that unbundling its work was too complex, she argued, yet the new codes’ definition aligned almost exactly with what nurses deliver. Her warning was blunt: act now, or nursing gets written out of the coming changes. 

Since that post, nearly everything she flagged has advanced. The panel held its first formal CMAA discussion, the 2026 code set added its first codes describing AI-performed services, and Appendix S was revised to sharpen the definitions the CMAA framework will rest on. CMAA itself is still a proposal, but the direction is unmistakable, and no parallel effort for nursing codes has appeared on the panel’s agenda.

“Here’s the financial reality for CEOs and CFOs that we need to consider: once AI becomes a billable line item and nursing remains a fixed cost buried in the room rate, capital allocation will follow the money. AI will show a clear ROI; nursing won’t. Not because nurses create less value, but because the financial model keeps them as a cost. No CEO will ever sign a memo that says ‘We’re cutting nurses for AI.’ But the reimbursement model will make that decision anyway, one budget cycle at a time.”

Rebecca Love

RN, BS, MSN, FIEL, Founder Commission for Nurse Reimbursement

Read the CMAA definition again, because “medically actionable outputs” could also describe a nurse’s shift. Catching deterioration before the numbers crash. Deciding when a patient needs escalation. Assessing, intervening, reassessing. Nurses produce medically actionable outputs every hour of every shift. They just aren’t called that on an invoice.

There’s a second irony inside the first, one Love also raised in her original post. Autonomous AI does not run itself in a clinical setting. Someone has to respond to the alert, act on the recommendation, and watch the patient. In hospitals, that someone is overwhelmingly a nurse. Under the emerging payment structure, the algorithm’s contribution would be billable while the nurse acting on its output remains a cost.

And the tools are already in the building. Our Nursing AI Watch database tracks 127 confirmed AI deployments across 106 large U.S. health systems, and the footprint is wide: ambient documentation tools, predictive deterioration scores, virtual nursing platforms. Most of it comes from a handful of vendors, and most runs inside a single EHR. Yet only a small number of the systems we track have any AI language in their nurses’ union contracts. The payment layer now being drafted at the AMA is one more place the rules of AI in healthcare are being written largely without nursing at the table.

The complexity argument isn’t invented. Creating and maintaining nurse-specific billing codes would be genuinely difficult, and it would add administrative work to a system drowning in it. The Commission’s own position statement weighs several possible payment models rather than declaring billing codes the single answer.

CMAA is also still a proposal. No CMAA codes exist yet, and a CPT code is not a guarantee of payment. As the Bipartisan Policy Center notes, clinical AI still lacks a clear Medicare benefit category, which means even AI reimbursement remains unsettled.

And there’s no villain here. The CPT Editorial Panel is a physician-led body doing what it was built to do: keeping billing language current with medical practice. The question this moment raises isn’t whether anyone is acting in bad faith. It’s who is missing from the table while the language of the next era of healthcare payment gets written.

“The time is now for a nursing-specific CMAA equivalent to become a billable service that finally lets health systems invest in the nursing workforce that is the backbone of patient care, instead of absorbing that cost with no return. We cannot build a reimbursement model that pays for AI’s clinical work and not for nursing’s. The same process now being used to unbundle and bill for AI is exactly the process needed to unbundle and bill for nursing. It is no longer too complicated. It is no longer too complex. A hundred years is long enough to wait for an ROI on nursing that was always there. It is time for all nurses to call for a fair process to create CMAA-equivalent codes to represent nursing.”

Rebecca Love

See also  Nursing Student, Paralyzed In a Drive By Shooting, Graduates With Twin Sister

RN, BS, MSN, FIEL, Founder Commission for Nurse Reimbursement

The AMA’s process is formally open to stakeholder input, and the written comment deadline for the next cycle is August 10, 2026. But commenting there requires applying for “Interested Party” access and AMA approval, which makes it a realistic path for nursing organizations more than for individual nurses.

The direct path for nurses to comment runs through the Centers for Medicare & Medicaid Services (CMS), the federal agency that decides what Medicare actually pays for. Whatever codes the AMA writes, CMS determines whether and how they get paid. Its annual payment rules are open to public comment from anyone, no application required. 

The timing could not be better: on July 14, CMS released its proposed 2027 payment rules, which for the first time includes a temporary payment category for clinical AI tools, called “Software as a Medical Service.” The payment side of the AI story is being written right now, in a venue where every nurse can be heard. If you want the people setting payment policy to understand what nursing contributes when AI enters the room, this is where you tell them.

For nursing organizations, that means the window to submit formal comments is open now. For individual nurses, the most practical steps are 1) submitting a comment to CMS, 2) supporting organizations already engaged in this work, including the Commission for Nurse Reimbursement, and 3) raising the issue inside their own hospitals. If your health system has an AI governance committee and no nurses on it, this is an agenda item to bring.

The next time a nursing leader hears “we can’t afford more nurses,” part of that answer will trace back to billing rules being drafted right now. Payment systems are hard to change once they’re written, but nurses still have a chance to help write this one.

🤔If your hospital could bill for an algorithm’s clinical judgment but not for yours, what would you want the people writing the billing rules to know about your work?  Share your thoughts in the comments below.

Nurse.org Analysis