The great unbundling

What happens when health care outgrows the system to pay for it?

I practiced medicine before the Affordable Care Act was signed into law, before the hot mic in the White House East Room memorialized it as a “big f*cking deal.”

It was. And it wasn’t.

Back then my clinic sat on a worn stretch of our post-industrial New England city. The waiting room was rarely empty. Strollers lined the walls. Coats piled onto chairs in winter. These were not patients with options.

Now, most had insurance. But high deductibles meant that getting pregnant still cost thousands out of pocket. Coverage did not mean protection. It meant entry into a system that could still ask more than they could afford.

Meanwhile, costs have kept rising. In 2025, the United States spent over $5 trillion on health care—nearly one-fifth of the economy. But no one goes to medical school to treat the economy.

In the clinic, I would focus on the patient in front of me. The rest faded into the background. Billing codes were just strings of letters and numbers, clicked through between visits, notations I barely registered. I moved through my days directing vast sums of health care spending, like a conductor leading an orchestra I couldn’t hear.

In my first year out of medical school alone, that tally reached into the tens of millions—an unfathomable contrast to my $50,000 intern salary. For patients, those costs would surface months later as a weighty invoice spread across a kitchen table.

Somewhere between those clicks and that bill is a system. It has a language.

The American Medical Association maintains a codebook called the Current Procedural Terminology (CPT®) Professional Edition. It is a translation manual containing 11,525 codes that doctors choose from to bill for our services.

Billing is a hard language to learn. Even a routine physical can get complicated. One set of codes is required for a new patient, another for an established patient. One set of codes is used if they are 25 years old, another if they are 65. One set of codes is needed if there is a problem to be addressed, another if there is a vaccine to be administered or a lab panel to order.

Maternity care has long taken a simpler approach. We use just a handful of codes, a single fee, paid as a bundle that covers prenatal visits, delivery, and postpartum care altogether. But the bundled payment approach is now under strain and on the verge of being abandoned—a seemingly obscure, technical change that could shape the care families receive for a generation

Seismic shift

That clinicians have so little visibility into how their decisions shape the flow of dollars can be forgiven. The system operates as much outside the exam room as within it.

Few could trace how a billing code becomes a negotiated rate, then fragments into a co-pay, a deductible, and some opaque category of “patient responsibility,” with the rest assigned to an employer or insurer. Even the most diligent accountant would struggle to follow it. But Dr. Barbara Levy has spent a career deciphering the way a code becomes a bill.

In the 1980s Dr. Levy was among the first gynecologists to perform surgery without large incisions. A camera entered through the navel. Instruments followed through openings no larger than the width of a fingernail. Operations that once required days in the hospital could begin in the morning and end at home by evening.

The problem was that the system had no language for it. “If you don’t have a code, you don’t get paid,” she told me.

“If you don’t have a code, you don’t get paid,” she told me.

When she pointed out the gap, her mentor enlisted her to help fix it. That request drew her deep into the Byzantine world of medical billing—and into the two American Medical Association committees that define how care is described and paid for. Together, these committees shape how roughly $400 billion in physician fees are allocated each year. Their work is foundational, and rarely publicized.

Dr. Levy is who I called after seeing the announcement from the American College of Obstetrics and Gynecology (ACOG) about unbundling. And she explained that the simplicity of the bundled payment system had long ago begun to fray.

The model took hold in the early 1980s, when obstetricians often cared for the same patients in the clinic and at the bedside. Dr. Levy explained that today, the clinician that a pregnant person sees in the office is rarely the one present at delivery. Care is distributed across teams—midwives, obstetricians, specialists—and increasingly shaped by technology.

At the same time, the clinical profile of pregnancy has shifted. More than 30 percent of pregnant people now have obesity, diabetes, or hypertension, yet the payment remains the same as for a straightforward pregnancy.

The bundled model also created a timing problem. Physicians are typically paid only after delivery, months after care begins. For a nine-month episode that may include dozens of visits, practices carry the financial burden while payers reconcile claims long after the fact.

Dr. Levy tells me that an AMA workgroup met over a period of years to design an alternative. In the end, they chose to unbundle the maternity fee. Beginning January 1, 2027, each component of each visit will carry its own charge: one for an exam, another for an ultrasound, another for lab work, another for administering medication, another for counseling.

ACOG contends that this approach better reflects how modern pregnancy care is actually delivered, particularly for complex or high-risk patients. At the same time, it acknowledges some risks: greater fragmentation, increased administrative burden, and potential equity concerns with winners and losers. Dr. Levy also acknowledged the transition may be uneven, as practices adapt, insurers update their systems, and pregnancies that span the calendar year are caught up in the January 1st change.

Still, advocates insist that something had to give. As Dr. Levy put it, the bundled model had simply “become too large to hold.”

Aftershocks

My friend Dr. Warris Bokhari lives on the frontlines of consumer frustration with America’s health care billing system. When I called him last week, he described unbundling as “the original sin.”

“Look at what happened with cable TV,” he said. The shift to streaming promised flexibility and choice. Instead, it required multiple subscriptions—Hulu for The Bear, Netflix for Stranger Things, Prime for The Marvelous Mrs. Maisel. For the average consumer, prices went up.

“Look at airlines,” he continued. Airfare once included baggage, a meal, maybe even a movie. Now each is an add-on. In health care, the add-ons are less optional.

Warris was born in Croydon, the working-class southern edge of London. His mother lived with crippling rheumatoid arthritis; his father left the workforce early. The household bent around illness and constraint.

By adolescence, he was already familiar with the rhythms of the health system—the waiting rooms, the negotiations, the sense of things unfolding beyond his control. Warris spent his fifteenth birthday in an intensive care unit as his mother died. The experience pulled him into medicine, and eventually beyond it.

He trained as a critical care physician before moving to the United States, where his career spanned roles at General Electric, Apple, and Anthem. He eventually founded Claimable, a company designed to help patients challenge denied insurance claims. Roughly one billion claims are denied each year, he told me, and only about one percent are ever appealed. Claimable uses AI to make the appeals process scalable. This Robinhood approach earned him a place on the TIME100 Health list this year.

From Warris’ vantage, he sees a lot of claims—and it’s taught him that when care is itemized, each component becomes something to be separately scrutinized and thus denied. “You’re giving insurers much more visibility into every piece of the episode,” he said. “And placing critical decisions in the hands of the people least qualified to make them.”

He also worries about incentives for providers. For the change to be budget neutral in aggregate, low-risk pregnancies must generate fewer billable services than high-risk ones. Over time, that could create a quiet pull away from the steady, preventative care that keeps pregnancies low-risk in the first place, and instead towards complex, perhaps unnecessary interventions.

Dr. Kavita Patel sees that risk from the clinic. An internist at Mary’s Center, a federally qualified health center in Washington, D.C., she provides prenatal care to some of the city’s most vulnerable patients.

“My biggest concern,” she told me, “is for the plurality of births that are low-risk—where the reason they’re low-risk is very good prenatal care, very good interdisciplinary support. Those are not easy things to do.”

“I tell my patients, I want everything about your pregnancy to be boring. The best outcome is that nothing interesting happens. That’s my job.”

Across my conversations with OB/GYN practices around the country, I heard different versions of this concern. One administrator speculated that practices already operating on thin margins may reduce how many low-risk pregnancies they make time for.

“Given the choice between a patient who needs an ultrasound and a procedure, and a straightforward prenatal visit,” she added, “how do you think I’m going to fill my schedule?”

Towards stable ground

Like all superheroes, Dr. Patel has a dual identity. In the mid aughts she staffed the Senate HELP Committee under Ted Kennedy and then later shepherded the Affordable Care Act as a policy director in the Obama White House. It was during that time that ACOG first began making the case for unbundling maternity care.

The ACA pushed the system toward more bundled payments and value-based care—paying for outcomes rather than activity. But as confidence in the ACA has eroded, so too has confidence in its unfinished innovations.

The choice today is often framed as binary: return to fee-for-service or continue forward with value-based care. Dr. Patel, who went on to a decade-long stint imagining better policy as a Senior Fellow at the Brookings Institution, sees a missed third path.

“Nobody in power,” she told me, “seems to be saying that trying bundling again—with more precision—could yield better results.”

She paused. “But somebody should.”

She recalls early data showing pregnancy care drifting away from the bell curve the bundle assumed, toward more high-risk management. The data was compelling. But at the time, unbundling ran counter to the broader push toward value-based care. Joint replacements became bundled. Aspects of cardiac and cancer care were bundled as well.

“Nobody in power,” she told me, “seems to be saying that trying bundling again—with more precision—could yield better results.”

Her view now is that reexamining the bundle follows the right clinical logic, but with the wrong execution. Ideally, she told me, we would pay for care the way we want it delivered—rewarding strong prenatal care for uneventful pregnancies. Because that is difficult, she favors experimentation.

We now have the data and technology to study maternity care in detail—what it costs, who provides it, and where complexity concentrates. We could have designed a controlled experiment, and then let CMS test, observe, and learn before scaling a new model.

“But nobody did that,” she said. “I looked before we talked. I couldn’t find it in the literature.”

And so instead, a version of that learning must happen in hindsight. After implementation, in the years to come, Dr. Levy says that the AMA and CMS will study how the new codes are used. The data will be more granular than ever. And the question question is what we ultimately do with this information.

Unbundling begins in nine months. Adjustments to payments may come—through modifiers, through future rulemaking—but promises are not mechanisms, and accountability has rarely followed intention in American health care. Now that decision is made, much work still remains.

What could go right is that better data reveals what maternity care actually requires, and we use these data to rationalize payments. Dr. Patel believes that the advent of health information exchanges and artificial intelligence to support analysis makes this more feasible than it was in 2010, when she began these conversations in the Obama White House.

But what could go wrong is that we mistake measurement for reform. That we get the new codes and call it progress. That those interpreting the data lack the experience to grasp what it misses. That we are unable or unwilling to make the hard choices better care will require.

Warris Bokhari, CEO of Claimable, told me his greatest concern.

“I worry,” he told me, “that it’s going to become very hard to make someone understand something if their salary depends on not understanding it.”

What my team is reading, building, and working against

• In 2009, I started a nonprofit called Costs of Care alongside friends Chris Moriates and Vineet Arora, to address the ignorance of many clinicians (including my own) to how our decisions impact what patients pay. Chris, Vinny, and I later wrote an article in JAMA, with a new credo for the profession called First Do No (Financial) Harm. I’m proud of the work Costs of Care continues to do with medical schools to train clinicians in how to address affordability within their own practice

• The cover story in the Sunday edition of the Boston Globe discusses how Massachusetts, despite being a model state in terms of health insurance coverage, still sees maternal complications that are 16% above the national average and c-section rates that are double what the World Health Organization considers optimal. I told reporter Sarah Rahal that if I could change one thing, it would be this: pay more for vaginal births than C-sections. Not because surgery is wrong — but because what we pay for shapes what we do. And right now, we are paying our way toward the wrong outcomes.

• In 2013, Atul Gawande wrote about why good ideas in medicine spread slowly — or don’t spread at all. It is nominally about the history of anesthesia and antiseptics, but it is really about the gap between knowing what works and actually doing it. I have thought about this piece for years. I am thinking about it again now.

• Every fifteen years or so, the American frustration with the costs of health care reaches a boiling point, catalyzing the political supermajority required to act. KFF has been tracking this pattern carefully, and their preview of health care’s role in the 2026 midterms is essential reading — not because it tells us what will happen, but because it clarifies the stakes. The unbundling of OB billing is a smaller story inside a much larger one.

• A recent New York Times opinion piece opens with a chilling lede: an Axios story on maternal health policy cited “findings” showing a majority of people trusted their doctors and nurses. Not initially disclosed: no people were involved. The poll was a computer simulation run by an AI startup, as part of a practice called “silicon sampling” AKA using LLMs to generate synthetic survey responses at a fraction of the cost. Their value prop feels both enticing and potentially serpentine in a field where our data problem is already acute. My team is watching closely.