06-reference/research

change healthcare 118 appeal cost provenance

2026-11-16·research-brief·source: deep-research·by Ray Data Co (deep-research synthesis)

The $118 appeal cost is real and first-party: it is on slide 6 of the 2017 Change Healthcare index, with no published derivation

The question

Does the superseded 2017 Change Healthcare Healthy Hospital Revenue Cycle Index actually contain the $118 formal-appeal cost figure, or is that number a secondary-source artifact with no first-party origin at all?

Context: [[2026-10-17-behavioral-health-silent-rework-market-size]] found the 2022 successor index carries no cost figure of any kind and could only trace $118 through a vendor attribution table to an MGMA article of December 2018. That left open whether the number had a first-party origin at all.

Answer: yes, it has one. The figure is in the original 2017 index and in the same-day press release. It is not a secondary-source artifact. No retraction is needed. What the index does not contain is any derivation, sample size, or definition of what work the $118 covers.

What we already know (from the vault)

What the web says

Convergences and contradictions

Synthesis for RDCO

The provenance question is closed, and the answer is the less dramatic one: the number is real, first-party and dated. Anyone who says the $118 has no origin is wrong, and the vault should not say it. The right citation is Change Healthcare, Healthy Hospital Revenue Cycle Index, June 2017, slide 6, with the press release of 26 June 2017 as the earliest dated instance. The MGMA December 2018 article is a secondary citation and should stop being treated as the source.

What the source cannot support is most of what the number is used for. It is a single vendor's "statistical average" for its own hospital customers on 2016 claims, published in a ten-slide marketing deck released at a trade show to promote the vendor's analytics products, with an explicit disclaimer that the data "may or may not be representative." It has no stated sample, no cost build-up and no definition of an appeal. It is nine years old, has never been updated, and was dropped from the successor publication. Premier's $43.84 (2022) and $57.23 (2023) figures for full adjudication of a disputed claim are newer, first-party, and less than half the size, which is a reason to doubt that $118 describes a typical denial today, though the units are not confirmed to match.

For the family behavioral-health billing venture, the practical consequence is on the cost side of any appeal-versus-rework comparison. A model that prices a formal appeal at $118 and quiet rework at $25 is comparing an undefined hospital vendor average from 2016 against an unsampled practice figure from 2021, and the gap between them is not evidence that appeals cost about five times what rework costs. No number in this chain was measured for behavioral health. If the venture's case needs a cost per touch, it has to come from timing the work directly in a real billing operation, not from this literature.

One caution on my own reading: I verified the index and the press release at source, but the trade coverage only through search results, and I did not recover the microsite. The statement that no fuller methodology exists is an absence of evidence from an incomplete search, not a confirmed negative.

Why this is in the vault

It settles the open follow-up in [[2026-10-17-behavioral-health-silent-rework-market-size]] and fixes the citation and the caveats for any cost-per-denial input used in the family revenue-cycle venture's unit economics, so the $118 is neither dismissed as fabricated nor quoted as a measured benchmark.

Open follow-ups

Related

Sources

Primary, first-party, original document downloaded and read (text extraction plus rendered slides 3 and 6)

Primary, first-party press release

Secondary, seen in search results only, not fetched or not readable

Vault

Flagged, not retried

Research budget used: QMD 0 of 5 (vault coverage done by direct grep and by reading the parent brief); WebSearch 1 of 3; WebFetch 0 of 3, replaced by direct curl retrievals: 2 Wayback page or PDF downloads, 3 Wayback index queries, 1 blocked Becker's fetch, 1 blocked live Business Wire fetch.