A ghost network is a parity violation

The claim

If a health plan tells you it covers psychiatric care, hands you a list of psychiatrists, and a third of that list is wrong, then the plan has not covered psychiatric care. It has covered the appearance of psychiatric care, which is cheaper and passes inspection.

We don't call that a parity violation. We call it a directory problem, which sounds like something an intern fixes in a spreadsheet. That framing is doing enormous work, and it's doing it for the plans.

A directory is a limit

Parity law exists to stop plans from putting harsher restrictions on mental health care than on medical care. When the crude limits were banned, the restriction migrated into things that are harder to see: prior authorization, medical necessity criteria, and network composition.

Network composition is on that list. It's an established category of nonquantitative treatment limitation. So the question isn't whether an unusable network could be a parity issue. It's why we so rarely treat it as one, when the Senate Finance Committee's secret shopper study found staff could book an appointment 18 percent of the time. Imagine that finding about cardiology. It would not be filed under data hygiene.

Why the framing decides the outcome

Framing determines who has to answer.

Call it a data problem and it belongs to an operations team, gets a quarterly cleanup project, and never changes, because the plan bears no cost when a listing is wrong. The entire cost lands on a person with depression making their eleventh phone call.

Call it a parity problem and it belongs to a regulator, becomes evidence in a comparative analysis, and carries consequences. Same facts. Completely different gravity. The reason this matters right now is that the enforcement architecture is already wobbling: as of July 2026 the federal departments have said they won't enforce the newest parity rule pending litigation. If the strongest available frame is going unused while the tools themselves are being weakened, the practical result is a coverage promise nobody has to keep.

The case against this view

Here's the strongest version of the other side, and it isn't weak.

Plans don't fully control their networks. Psychiatrists leave panels because reimbursement and administrative burden make participation irrational, and a plan can't compel a clinician to stay or to answer a phone. Calling every stale listing a parity violation punishes plans for a supply problem they didn't create, and in the worst case they respond by shrinking directories to only the provably reachable, which makes networks look thinner without making them better. There's also a real argument that the Senate study was small, 120 calls in a specific slice of plans, and can't carry the weight of a national legal theory.

All fair. My answer is that plans do control the reimbursement that empties the panel, they do control whether the directory is verified, and they do benefit when it isn't. A party that profits from an inaccuracy and has the power to correct it doesn't get to call the inaccuracy an accident. But I hold that with less certainty than the rest of this piece, and the person who thinks I'm stretching parity doctrine past what it can bear may simply be right.