Medicaid plans keep two lists of maternity doctors. The regulator's copy has more ghosts.

The finding, in one paragraph: Medicaid pays for more than 40 percent of the births in the United States, mostly through managed care plans that states pay per enrollee to maintain networks of doctors.▸ Whether a network is real is checked against a list the plan itself sends the state. The HHS Office of Inspector General sampled the maternal-health providers on those lists — the OB/GYNs and nurse-midwives of the three largest Medicaid insurers, in five states — and telephoned them. Nearly one in four said they were not in the network that listed them; some had retired years earlier, some had never heard of the plan, some practiced in a different state. The plans' public directories had mostly already dropped these people. The copy that overstates the network is the copy the regulator relies on.

Five documents, read directly — two of them released as a matched pair on June 11, 2026.▸ The first is the HHS OIG evaluation Inaccurate Medicaid Managed Care Network Lists May Compromise State Oversight of Access to Maternal Health Care (OEI-05-24-00091) — the regulator-facing list. The second is its companion, Inaccurate Medicaid Managed Care Provider Directories May Limit Enrollees' Access to Maternal Health Care (OEI-05-24-00090) — the patient-facing directory. Both surveyed the same sample: 453 OB/GYN physicians and certified nurse-midwives drawn from the network lists of Centene, Elevance, and UnitedHealthcare plans in Louisiana, Missouri, Nevada, New Jersey, and Washington. The precedent document is the OIG's October 2025 data brief on behavioral health networks (OEI-02-23-00540), which found the same disease in a different organ. The remedy documents are CMS's 2024 managed-care access rule, as described in both reports, and CMS's formal concurrence with all three recommendations.

The $450 billion states paid managed care plans in fiscal 2024 is capitation: a fixed payment per enrollee, in exchange for which the plan promises a network adequate to deliver the care.▸ Adequacy is enforced through quantitative standards — an OB/GYN within 30 miles or 30 minutes of every enrollee, say — and states test compliance against the plans' own network lists. Some states attach money to the test: enrollment freezes or monetary penalties for inadequate networks. That makes the network list a financial document. Inspectors called every provider on the sample and asked one question — are you in this network?

On top of the ghosts, the plans supplied no working phone number for 25.6 percent of sampled providers — the OIG notes some of the unreachable may be additional ghosts it could not confirm.▸ A network list where a quarter of the entries deny membership and another quarter cannot be dialed is the instrument states use to certify that mothers on Medicaid can find a doctor.

The companion report walks the same ground as an enrollee would — a pregnant woman searching her plan's online directory:

The directory fails in both directions.▸ Nine percent of its listings are ghosts (22 percent in Centene's directories) — providers a woman might select a plan for, then discover she cannot use. And 22 percent of genuinely in-network providers are missing from it — real capacity, invisible, which the OIG notes matters most "in communities with maternal health provider shortages, such as many rural areas." A third of the accurate listings carry a wrong phone number or address. The report's plain summary of the stakes: timely prenatal and postpartum care prevents deaths, and the United States has "worse outcomes than in any other high-income country."

The two reports together catch the plans' own two documents contradicting each other.▸ The parent companies told OIG the network list and the directory "should generally have the same providers." Measured: 46 percent of network-list providers were absent from the directories. And the direction of the discrepancy is the finding: among providers who confirmed they were not in-network, 77 percent had already been removed from the public directory — but remained on the list sent to the state. Where the plans' data was corrected, it was corrected in the copy customers see, not the copy regulators grade. The OIG stops short of assigning motive; the arithmetic doesn't need one. A network that is 24 percent ghosts passes a 30-mile adequacy test it might otherwise fail.

The remedy has the same asymmetry.▸ CMS's 2024 access rule orders states to run independent "secret shopper" surveys — but they begin with rating periods starting on or after July 10, 2028, two years from today, and they verify only the online directory: network status, address, phone, and new-patient acceptance. The OIG's central point is that the rule "does not address the accuracy of network lists" — the regulator-facing copy, the one with more ghosts, stays unaudited. This is also not the first warning: the OIG's October 2025 behavioral-health brief found plans whose listed networks barely existed in claims data — in one Medicare Advantage plan, 353 of 356 listed providers delivered no service to any enrollee in a year.

CMS concurred with all three recommendations across the pair of reports: work with states to improve the accuracy of the data used for network-adequacy reviews, support states in holding plans accountable for network-list accuracy, and support states in holding plans accountable for directory accuracy.▸ The concurrences carry no deadlines. The enforceable machinery arrives with the secret-shopper surveys in rating periods from July 10, 2028 — scoped, as of today, to the directory only. The OIG suggests the structural fix in passing: a single common data source for both oversight and enrollee search, so there is no second copy to diverge.

All figures are from the documents cited in-line, each read directly. Estimates generalize only to the three companies and five states studied; per-company differences on phone-number accuracy were not statistically significant.