Is this dataset available to health plans as well as patients?

Yes, on identical terms at identical prices. A record perceived as one side's weapon is not citable by anyone, and the payer is also the buyer with the largest budget. This is a commitment we have offered to have written into any licence.

Licence & neutrality

Licence, attribution and how to cite this

Q: Can this California IMR data be used commercially?

Unresolved, and we say so rather than assuming. The open-data portal displays a notice that commercial uses must be approved. Two things cut against it applying here: the notice is byte-identical boilerplate that also appears on datasets from a different department, and the decisions are published under an affirmative statutory mandate — Health & Safety Code section 1374.33(g) — that grants no copyright or licensing authority anywhere in the article. A written request to the Department is the first route and a Public Records Act request is the fallback. Until one of those lands, nothing derived from this file is sold.

Q: How should this dataset be cited?

Cite the Department for the data and us for the analysis, because they are different claims. The source is: California Department of Managed Health Care, “Independent Medical Review (IMR) Determinations, Trend”, via the CalHHS Open Data Portal; preferred citation “DMHC IMR Data, 2001 – Current”. The classification, the n=30 floor and the crosswalk are ours and are wrong in our name, not theirs.

Where this data comes from, what we have promised California’s regulator in writing, and how to credit it if you use it. The promises are published here because one that lives only in a letter binds nobody.

Can this data be used commercially?

resolved

Unresolved, and stated as unresolved. The CalHHS open-data portal displays a notice that commercial uses of this file must be approved and may be subject to a licence.

Two findings cut against that notice governing here. It is portal boilerplate — byte-identical text appears on datasets published by a different CalHHS department, it points at a disclaimer “in the data file” that does not exist, and the dataset’s own licence field reads “No License Provided”. And these decisions are published under an affirmative statutory mandate — Health & Safety Code § 1374.33(g)–(h) prescribes the database and its fields, and grants no copyright or licensing authority anywhere in the article. California courts have held that an agency lacks statutory authority to copyright records subject to the Public Records Act or to condition their release on a licence.

That makes the position strong and it does not make it settled. A written request to the Department is the first route; a Public Records Act request is the fallback; counsel confirms before either is relied on.

Consequence, and it is the whole reason this page exists. Until one of those lands in writing, nothing derived from this file is sold to anyone, at any price. Launching on the belief that a notice is probably unenforceable is not how this gets built.

What we have committed to

given in writing

Attribution, on every page and every file: Every page credits the California Department of Managed Health Care, cites the dataset page and uses the Department's preferred citation, “DMHC IMR Data, 2001 – Current.”
Modification labelled, never passed off as official: Any derived table states that the data is modified from the published file, is not official government data, and that the analysis is ours and not the Department's.
No re-identification, ever — a standing rule, not a best effort: No attempt to re-identify any enrollee, provider or plan. No linkage of these records to any other dataset. No reproduction of a fact pattern detailed enough to identify a person in a small community. Where a condition is rare the reasoning is paraphrased, never quoted.
No plan identification, because the Legislature removed it: Health & Safety Code § 1374.33(g) requires the plan's name to be removed before these decisions are published, and the file contains no plan identifier. We will not attempt to infer or attach one.
Both sides, the same product, the same price: The same statistics, the same reasoning extracts and the same data feed to patient-side users, provider appeal teams, health plans and researchers — identical terms, identical prices. Offered for writing into any licence.
No prediction, no advice, and free help named first: Every artifact states that it describes published outcomes, is not medical or legal advice, and does not predict any individual case. The Department's own Help Center is listed prominently in every consumer-facing product.

Both sides, same product, same price

our standard

A patient, a provider appeal team, a health plan and a researcher get the same statistics, the same reasoning extracts and the same feed, on the same terms, at the same price. There is no patient edition and no payer edition.

This is not generosity. A record perceived as one side’s weapon stops being citable by anyone, and the neutrality is what makes the record usable as a record. It is also why there are no “worst insurer” leaderboards here and never will be — and, separately, why there could not be even if we wanted them: § 1374.33(g) removed the plan’s name before we ever saw the file.

How to cite this

Cite the Department for the data and us for the analysis. They are different claims, and the second one is the one that can be wrong in our name.

Source: California Department of Managed Health Care, “Independent Medical Review (IMR) Determinations, Trend”, via the CalHHS Open Data Portal, data.chhs.ca.gov. Preferred citation: “DMHC IMR Data, 2001 – Current.” Published under Cal. Health & Safety Code § 1374.33(g).

Analysis: Claim Appeal Data, data as of 2026-09-04. Data modified from the published file; not official government data; not affiliated with or endorsed by the Department. 42,621 analysable decisions; no rate published below n=30.

If a figure here is wrong, say so. The corrections process is public and dated, and a correction the Department asks for gets priority review against the source record. How these numbers were made · whether this covers you.