A third of Medicaid maternal-care provider listings are wrong

Medicaid pays for over 40% of births in the United States⧉, almost all of them through managed care plans that states pay more than $450 billion a year⧉ to run.▸ A federal watchdog tested whether the online directories those plans publish -- the tool a pregnant enrollee uses to find an in-network doctor -- actually match reality. In the three largest national Medicaid managed-care companies' plans, across five states, 22.4% of maternal-health providers confirmed to be in-network didn't appear in the plan's own directory at all, and a third of those who did appear had a wrong phone number or address⧉.

HHS's Office of Inspector General built its sample from the network lists Centene, Elevance, and UnitedHealthcare plans submit to five states -- Louisiana, Missouri, Nevada, New Jersey, and Washington⧉ -- pulling a stratified random sample of 453 OB/GYN physicians and certified nurse-midwives⧉, 151 per company.▸ Investigators surveyed the providers directly by phone, completing 356 surveys, a 79% response rate⧉, and separately looked each one up in the plan's public online directory.

Of the 248 providers who confirmed to OIG's surveyors that they really were in-network with a given plan, 22.4% did not appear anywhere in that plan's own online directory⧉ -- meaning an enrollee searching that directory would see them as out-of-network, or not see them at all.▸ The rate varied sharply by company: an estimated 34.7% for Centene, 25.3% for UnitedHealthcare, and 11.0% for Elevance⧉ -- a difference OIG's own statistical test found significant, not noise.

The reverse error also showed up.▸ Among providers who did appear in a directory, 9.3% told OIG's surveyors they were not actually in-network with that plan at all⧉ -- what the report calls "ghost providers." Some had never contracted with the plan; others practiced in a different state and didn't take that state's Medicaid plans. Here too the rate varied by company, from 22.2% at Centene down to 4.1% at Elevance and 5.6% at UnitedHealthcare⧉.

A third failure mode hits providers who are genuinely in-network and correctly listed as such: the contact details are simply wrong.▸ Of the 188 providers in that category, 33.3% reported at least one inaccurate phone number or address in their directory listing, and 20.7% said they didn't actually work at one or more of the locations listed for them⧉. For a directory whose entire purpose is letting an enrollee search by location and call to book care, that's the failure that turns a correct listing into a dead end anyway.

CMS already knows how to set a hard bar for provider-access data: its 2024 rule requires states to treat a plan as noncompliant on appointment wait times unless at least 90% of sampled providers meet the state's maximum wait window⧉.▸ No such numeric standard exists for directory accuracy. Starting in 2028, states must run annual "secret shopper" calls that check some directory data -- but OIG concludes that requirement alone "might have limited impact," since it carries no measurable pass/fail bar and can't catch in-network providers who are missing from a directory entirely⧉ (a secret-shopper call can only catch what's listed, not what should be listed and isn't). CMS concurred with OIG's recommendation⧉ to work with states on plan accountability, and OIG notes an earlier, still-open recommendation on the same problem from a prior report.