Hospital price transparency arrived with a consumer theory attached. Publish what procedures cost, and patients would shop, and competition would do the rest. The files were duly published, in formats that ran to enormous size and required technical skill to parse, and consumers overwhelmingly did not use them.
The theory was wrong about the user, not about the value. The data is being used extensively, by employers who fund health benefits and by the advisers who negotiate on their behalf.
What the files actually revealed
The finding that changed conversations was variation. Negotiated rates for the same procedure at the same facility differ substantially between payers, and rates between facilities in the same market differ by multiples in ways that do not track quality, teaching status or case mix in any way the data supports.
A self-funded employer looking at that discovers something uncomfortable: they have been paying rates negotiated on their behalf without knowing how those rates compared to what others paid, and in some cases the plan with the largest network discount was not delivering the lowest total cost.
That has produced concrete behavior. Employers have used the data to renegotiate, to steer employees toward specific facilities with meaningful cost differences, and in some cases to move to reference-based pricing, which sets what the plan will pay against a benchmark rather than accepting a negotiated schedule. Each of those was available before. What was missing was evidence, and negotiation without evidence is just asking.
Hospitals have responded predictably, arguing the files are being read without context, and they have a point that is worth taking seriously. Rate variation reflects real differences in contract structure, service bundling, risk arrangements and volume commitments, and a comparison that ignores those can mislead. The trouble is that the magnitude of variation exceeds what those factors plausibly explain, which is a difficult position to argue from.
The consumer-facing failure remains real and is not fixed by employer sophistication. A patient facing a procedure still cannot readily determine what they will owe, because the answer depends on their plan's benefit design, their deductible status and the facility's billing, and no published file resolves that. Employers acting as purchasers can improve prices in aggregate without doing anything for the individual at the point of care.
This is broadly the same pattern visible across employer health strategy. Subscription primary care, on-site clinics and the spreadsheet scrutiny now applied to mental health programs all reflect employers behaving less like benefit administrators and more like purchasers of a service they are financing directly.
Compliance has improved but remains uneven, and enforcement has been mild relative to the stakes. Files that are technically posted but structured to resist analysis satisfy the letter of the requirement while defeating its purpose, and the sophistication now needed to use them is itself a barrier that favours large purchasers.
Whether that improves the system depends on a question the data cannot answer. Employers negotiating harder redistributes cost among payers. It does not by itself reduce what care costs to deliver, and a hospital that concedes on commercial rates will look to make it up elsewhere.
Earlier reporting on employers acting as purchasers rather than administrators appeared in Food as Medicine Moves From Pilot to Payment Code.
