In 2024, at 12 weeks pregnant, Kathleen Clark didn’t expect her OB-GYN to ask for a payment of nearly $1,000 before she could leave her appointment. It was a bill she hadn’t budgeted for and didn’t understand (as reported by CBS news). Like most expectant mothers, she assumed her pregnancy would be billed the way pregnancies have been billed in this country for decades: as one global obstetric package, tallied up and settled with her insurer only after her baby was born.
In 2024, her experience was the exception. Beginning January 1, 2027, it may become the rule.
That’s the date CPT 2027 takes effect, and with it, American obstetrics retires the global maternity package for a maternity fee-for-service model. Instead of one bundled rate covering an entire pregnancy, providers will submit claims for individual services as they happen—an ultrasound here, an office visit there, and labor and delivery as their own line items. Think of the difference the way you’d think of a vacation package versus a hotel bill. A bundled package charges one price no matter how many meals you order or amenities you enjoy. An itemized bill charges for exactly what happened, when it happened. Maternity billing is moving from the first to the second.
For patients, that will take some getting used to. For payers, it represents an operational shift that arrives with a set of risks the industry’s existing edit sets and itemized review models were never built to catch.
Why providers have wanted this for years
For providers, the shift from a global obstetric package to a maternity fee-for-service model comes as a relief, as flat reimbursement rates fail to account for the growing complexity of modern pregnancies.
The data backs this up. A 2020 analysis of the CDC’s natality database found that between 1989 and 2018, hypertensive disorders in pregnancy rose 149%, diabetes mellitus rose 261%, and that the share of pregnancies involving very advanced maternal age rose 194%. The average pregnancy today requires more monitoring, more counseling, and more judgment calls than the one a global package rate was designed for 30 years ago. Under a model that pays the same whether a pregnancy is straightforward or high-risk, the additional visits, screenings, and time that complex cases require have effectively amounted to free labor for the providers delivering them.
That doesn’t mean fee-for-service billing is a clean win for patients. Under the global package, patients like Kathleen Clark typically received one bill after delivery—a known, estimable number, arriving after the baby was already home. Billing visit by visit moves those costs earlier, into the middle of a pregnancy, at a moment when an unexpected charge might be enough to make a patient skip their next prenatal appointment. Delayed or missed prenatal visits let complications go undetected. Untreated complications tend to become more expensive and more dangerous the longer they’re overlooked.
The revenue cycle industry is already positioned for CPT 2027
Revenue cycle management companies exist to make sure providers are paid for what they do, and they view CPT 2027 as an opportunity to capture reimbursement that the global package left on the table. Medical Billers and Coders, one such firm, describes the transition as “both an EBITDA opportunity and a risk mitigation imperative,” and estimates that practices updating their global exception billing could recover $100,000 to $300,000 per month with their support. Their 90-day OB-GYN billing audit helps to capture these funds.
For a provider managing truly high-risk pregnancies, that recovery may be nothing more sinister than getting paid—finally—for work that was always being done. For a payer, that same shift looks different. Every service that becomes separately billable is also a service that must be separately validated, checked against documentation, against clinical guidance, and against the terms of the contract. Without systems built to do that validation at scale, payers are looking at potentially tens of millions of dollars a year in improper payments they have no way of catching.
How CPT 2027 could expose payers and patients to fraud, waste, and abuse
CPT 2027 could expose payers and patients to greater fraud, waste, and abuse by replacing bundled maternity reimbursement with individually billable services, creating new payment decisions that require validation that the previous services are not built to address. From a payment integrity lens, the change maps directly onto the three best-quantified improper billing practices (E/M upcoding, unbundling/fragmentation, and modifier abuse), and many payers struggle to counter an RCM vendor ecosystem that is already being paid to maximize captured revenue ahead of the effective date.
Under the traditional global obstetric package, most of the common forms of improper billing were largely irrelevant. Simply put, coding a prenatal visit at a higher level didn’t increase the payment because the payment was fixed. Adding another diagnosis code didn’t generate anything extra, like the vacation package analogy above. Even sloppy documentation of a medication dose was a compliance problem, not a revenue opportunity, because nothing about it touched the size of the global fee. The financial incentive behind upcoding, modifier stacking, and duplicate billing—schemes that show up elsewhere in healthcare—had no foothold in obstetrics.
Fee-for-service changes the math. Antepartum visits, labor management, delivery, and postpartum care used to be one fee. Now, each becomes its own payment decision, with its own code, its own modifier, its own chance to be billed wrong or billed aggressively. It’s a pattern that has played out in every other corner of medicine where itemized billing replaced bundled payments.
That does not mean the global obstetric package was free of waste; it was just a different kind of waste. A practice offering four prenatal visits and one offering 12 were paid identically, which made it nearly impossible for a payer to tell whether care was being delivered efficiently or being cut short. Fee-for-service billing provides greater visibility into what care actually looks like, along with a much larger burden of claims to check.
Most payers don’t yet have payment integrity programs built for obstetric billing specifically. CPT 2027 introduces new risks before many organizations have developed the systems needed to detect and address inappropriate billing patterns accurately and at scale. With revenue cycle management vendors positioned to defend providers’ interests, payers are in a time crunch to catch up.
AI helps payers plan for CPT 2027 changes
For payers anticipating CPT 2027, AI adds value by expediting the rate at which they can make accurate, compliant decisions. At Machinify, our work with payers has shown us how much of that opportunity comes down to giving experts the right information, context, and signals at the moment a payment decision needs to be made.
Think of AI as a research team of a thousand analysts working at the speed of light. They don’t make the call, but they give the expert everything they need to make it in seconds instead of weeks.
That is a sharp contrast from how claims review has worked until now. The traditional process starts with humans who received a claim, manually searched for supporting evidence through hundreds of pages of documentation, compared it to written contract rules, and made a judgment call. But humans make mistakes, especially when the volume of claims grows and the time to review each one shrinks.
AI helps payers identify the right claims for review, surfacing discrepancies, anomalies, and inconsistencies at a scale that’s not limited by the number of reviewers. Using natural language processing (NLP) and optical character recognition (OCR), it extracts key data from messy source files in seconds and reassembles them around the patient’s story. A human reviewer still makes the clinical call, but now they’re making it with the full record in front of them instead of a fraction of it. Patterns that used to take months or years to surface become visible immediately because the system can compare one claim against thousands like it at once.
Here’s a sampling of what AI is trained to catch in obstetric claims:
- Upcoding routine visits: Every prenatal visit gets coded at a level of complexity, and higher levels pay more. It doesn’t take a bad actor to describe a routine 28-week checkup as slightly more complicated than it was by, say, citing the patient’s body mass index and bumping the code up a notch. Multiply that by a dozen visits per pregnancy, and Medicare’s own auditors have already seen how it adds up. A review of 2010 claims found 21% of E/M payments improper, amounting to $6.7 billion.
- Stacking modifier 25: A provider can bill an extra office-visit fee on the same day as a procedure, but only if that visit was truly separate from it and not just part of the routine check. This is indicated by modifier 25. It’s something we expect to see with ultrasounds inside routine appointments, which don’t qualify for a separately billed claim.
- Fragmentation: Almost every corner of healthcare sees bills where the ordinary contents of a single visit are broken into separate charges. For maternal care, it could look like a heartbeat check, a fundal-height measurement, a urine test, or a few minutes of counseling. There is no maternity-specific rulebook yet telling a claims system which of these combinations shouldn’t be billed separately.
- Billing rare scans as routine ones: Ultrasound comes in tiers: a standard scan, and a more detailed, higher-paying scan meant only for a suspected problem. It’s easy to bill the detailed scan as though it were routine, and payers are already discovering this in their audits.
- Testing more than pregnancy calls for: Certain monitoring makes sense for high-risk pregnancies, but ordering these tests twice a week for someone whose only risk factor is her age, and each visit adds inappropriate charges.
- Getting paid twice at the seams: Many pregnancies will begin in 2026 under the old rules and end in 2027 under the new ones. The same care could be coded once under each system, amounting to duplicate charges.
- Phantom, undocumented counseling: Mental health screening, lactation counseling, and care coordination are all real, valuable services. But they’re easy to code without a note thorough enough to prove they happened.
None of these examples require a bad actor. Most of the time these errors are simply mistakes. AI tools are capable of detecting them before they become patterns that scale.
The results are compelling. For one regional payer, Machinify cut administrative costs by 30% while surfacing an additional $15 million in annual savings. And for patients like Kathleen Clark, AI helps payers process claims faster so that members can better understand and predict what they owe throughout their pregnancy.
Making the system more fair for everyone
Above all, changes like CPT 2027 serve as test of whether incentives are pointed the right way. Providers deserve to be paid for the complexity of care they’re actually delivering, especially now that complexity is the norm rather than the exception. Payers have a responsibility to be careful stewards of that money, so that what members are billed reflects what they actually received. And pregnant women deserve to deliver healthy babies without a billing surprise attached to the experience.
Kathleen Clark ultimately delivered early with a high-risk specialist outside the practice she’d been seeing throughout her pregnancy, after experiencing a blood pressure spike in the hospital. Because her labor occurred with a different provider, she was entitled to a refund for a portion of her upfront payment. But tracking down that refund wasn’t easy. She recovered about half of it only after she’d met her deductible through other services and called the office to ask, and the rest after she’d resolved the bills from her baby’s delivery. Under a maternity fee-for-service model that bills each service as it happens, she wouldn’t need to request a refund at all. She would simply pay for what happened when it happened.
Getting there requires payers to be ready to not only receive an increased volume of separately-billed maternity claims, but to detect fraud, waste, and abuse inside them before it compounds.
AI is not a neutral arbiter. It reflects the priorities of whoever builds and deploys it. But built with care and pointed at the facts, it can serve providers, payers, and patients at once, moving quickly through a rising volume of claims to establish what happened and give human experts what they need to make the call.
Healthcare needs to work better for everyone. AI might be what finally makes it happen.
Darshak Sanghavi, MD, is Chief Medical Officer of Machinify. Recently, he was one of the first Program Managers at the Advanced Research Projects Agency for Health (ARPA-H), a new multibillion dollar U.S. agency that the President tasked with developing health programs “so bold no one else, not even the private sector, is willing to give them a chance.” Overseeing an investment portfolio of several hundred million dollars, his programs cover cures for rare genetic diseases, regenerative medicine, women’s health, organ transplantation, innovative payment and business models for prevention, and many other areas.
Prior, he was Global Chief Medical and Clinical Operating Officer for Babylon, the global end-to-end digital health care provider serving over a dozen countries and over 24 million people, with the mission of bringing “affordable and accessible health care to everyone on Earth.” He was a member of the senior leadership team taking the company public in 2021 and oversaw a team of 1500 in the company’s global operations in the United States, United Kingdom, and Rwanda with revenues exceeding $1B. He is the former Chief Medical Officer of UnitedHealthcare’s Medicare & Retirement, the largest U.S. commercial Medicare program with over $90B in annual revenue, where he directed major national clinical and affordability programs. Earlier, he was Chief Medical Officer at OptumLabs, the R&D hub of UnitedHealth Group, running a portfolio of industry-leading projects with dozens of academic, government, and industry partners.
Before then, he served in a senior role in the federal government, as the Director of Preventive and Population Health at the Center for Medicare and Medicaid Innovation, where he directed the development of large pilot programs aimed at improving the nation’s health care costs and quality. In this capacity, he was the architect of the Accountable Health Communities model, the Million Hearts Cardiovascular Risk Reduction model, and the Medicare Diabetes Prevention Program, impacting tens of millions of Medicare beneficiaries. He was a fellow and managing director of the non-partisan Brookings Institution, and chief of pediatric cardiology at UMass Medical School (where he still sees patients). He’s an award-winning medical educator, has worked around the world and published dozens of scientific papers on topics ranging from the molecular biology of cell death to tuberculosis transmission in Peruvian slums.
A frequent guest on NBC’s Today and past commentator for NPR’s All Things Considered, Dr. Sanghavi was a columnist with Slate, The New York Times, Boston Globe, and Washington Post. His best-seller, A Map of the Child: A Pediatrician’s Tour of the Body, was named a best health book of the year by the Wall Street Journal. He previously worked as a U.S. Indian Health Service pediatrician on a Navajo reservation.
Educated at Harvard and Johns Hopkins, he completed his residency in pediatrics and fellowship in pediatric cardiology at Boston Children’s Hospital.
The content on this page is subject to our Terms of Use.
