Half Your Provider Directory Is a Rumor
In 2025, federal auditors found that 55% of the behavioral-health providers listed in Medicare Advantage networks billed zero services to enrollees all year. Not few services — zero. Seventy-two percent of the inactive providers they surveyed shouldn't have been listed at all. Six things from building a network instrument on the audit record.
1. The directory is not the network. Three consecutive CMS audit rounds found roughly half of directory locations wrong — most commonly, the provider simply wasn't there. A peer-reviewed secret-shopper study reached appointments with 26% of listed psychiatrists; 16% of the phone numbers were wrong, and one reached a McDonald's. The effective network is the listed network times the truth rate — and almost nobody measures the truth rate.
2. Adequacy is a floor, not a strategy. Time-and-distance compliance and real access are different facts: half of listed Medicaid providers couldn't offer an appointment when the OIG called; Senate staff posing as patients got appointments 18% of the time. Passing the paper test proves the paper. The regulators are converging on secret-shopper verification with 90% pass bars — the marketplace requirement is already live, Medicaid's is scheduled. Run the survey before the regulator does.
3. Behavioral health fails first. MA networks hold, on average, 16% of the local behavioral-health workforce. Two-thirds of MA psychiatry networks contain under a quarter of area psychiatrists. In more than half of studied counties, no MA-network psychiatrist was accepting patients. Every ghost-network headline lives here, and so does the next enforcement wave.
4. Say what the narrow-network savings are. Narrow networks carry 6–9% lower premiums — but the two best decompositions disagree about why. Colorado: two-thirds of the savings was selection — who enrolled, not what was paid. Massachusetts: quasi-random switchers genuinely spent ~40% less, through utilization redirection, with primary care up. Purchasable savings and a healthier population wearing a discount look identical on a premium sheet. A savings claim without a price/steering/selection split has skipped the hard part.
5. The termination has clocks. Mid-year network exits carry 45-day member notice for primary care and behavioral health, 30 for other specialties — duties tightened after an insurer cut 2,200 physicians in one state and a federal court enjoined it. The clocks and continuity duties belong in the plan before the letters go out.
6. The MA out-of-network rate you've seen is unsourced. Commercial out-of-network care is ~6% of spending and falling — that number is real. For Medicare Advantage, no published OON utilization rate exists at all. Whatever figure your vendor deck carries is a guess wearing the commercial number's clothes.
We built these into an instrument: the effective-network calculator (listed × truth rate), the savings-split check, the adequacy floors by program, and the termination clocks — every figure cited to the audit that found it.
A network is what answers the phone, not what fills the PDF.
What's your zero-biller rate, by specialty, by county — and has anyone ever looked?