How Optum’s Claims Editing System Got So Big the DOJ Forced a Divestiture That Became Lyric, Why Its PPS Pricer Touches Medicare Advantage Facility Claims, and Why Data Plus Distribution Matters
Video Preview
🎧 Podcast episode for paid subscribers only. Also available on Spotify.
How Optum’s Claims Editing System Got So Big the DOJ Forced a Divestiture That Became Lyric, Why Its PPS Pricer Touches Medicare Advantage Facility Claims, and Why Data Plus Distribution Matters
The DOJ tried to block a $13B healthcare deal because two companies together would have controlled ~95% of first-pass claims editing. Most people have never heard of that market. Here is why it matters.
To listen to paid episodes in Apple or Spotify, link your Substack subscription via the show settings on those platforms (instructions inside the Substack app under Subscriptions → Podcast).
Table of Contents
The 95 percent problem that made antitrust lawyers nervous
What these platforms actually do once the marketing stops
The data underneath, and the gap between public rules and private scar tissue
Distribution, or why being the plumbing beats being the product
The flywheel that quietly eats great models for lunch
Where the AI venture money keeps bouncing off
The cracks worth watching, because no moat lasts forever
Abstract
In 2022, the DOJ tried to block UnitedHealth’s roughly $13B purchase of Change Healthcare, partly because Optum’s claims editing system (around 25% share) plus Change’s ClaimsXten (around 70%) would have meant about 95% of first-pass claims editing under one roof.
Judge Carl Nichols let the deal through on a fix-it-first remedy: ClaimsXten was carved out and sold to TPG Capital for $2.2B. It later rebranded as Lyric, hired Raj Ronanki as CEO, and kept the ClaimsXten engine humming.
Separately, Optum owns the PPS pricing stack (EASYGroup, PSI, the federal and state pricers). By Optum’s own numbers, more than 80% of Medicare Advantage facility claims get priced through it every year.
Both businesses sit on the same two-layer data foundation: public CMS rule sets that anyone can download, plus decades of proprietary adjudicated-claims outcomes that nobody hands over.
Thesis: the moat is not the model. It is multi-payer claims data, deep integration into adjudication systems, and legal defensibility. AI lowers the cost of building a model, not the cost of earning trust or acquiring the data, which is why even top-tier AI investors keep struggling to crack this category.
The 95 percent problem that made antitrust lawyers nervous
Most people who get a claim denied never learn the name of the company that wrote the rule behind the denial. That is the whole story in one sentence. When UnitedHealth went to buy Change Healthcare for about $13 billion, the Department of Justice looked at one narrow market, first-pass claims editing, which is the automated check that runs the instant a medical claim hits a payer, and did some arithmetic that spooked everyone. Optum already ran a claims editing system with roughly a quarter of the market. Change owned ClaimsXten, which had something like 70%. Add those together and the combined entity controls about 95% of the thing that decides whether a claim is even allowed to proceed. The court filings put the post-merger concentration numbers well past every threshold that normally gets a deal killed on sight, with an HHI in the high 8,000s and a delta that screamed unilateral effects.
The fix here is the part that matters for the rest of this essay. Rather than walk away, UnitedHealth ran the classic fix-it-first play: agree to divest ClaimsXten before the trial even ends, so the judge can rule on a world where the overlap is already gone. The buyer was TPG Capital, the price was $2.2 billion, and roughly 375 employees went along with it. Judge Carl Nichols of the D.C. District Court found the divestiture good enough to preserve competition in first-pass editing and let the merger close in October 2022. TPG, in unsealed filings, said it planned to more than double the R&D budget from around $14 million to roughly $30 million by 2026, which is the polite way of saying the asset had been run for cash and could use some love.
ClaimsXten then did the most 2023 thing imaginable and rebranded as Lyric, a name that sounds like a feature inside a music app rather than the engine quietly denying portions of your colonoscopy claim. Raj Ronanki, fresh from running Carelon’s digital platforms at Elevance, came in as CEO. Carolyn Wukitch, who had run ClaimsXten since the year 2000, slid over to COO. The funny part, and the reason the DOJ partly lost, is that forcing the divestiture created a competitor on the editing side while leaving Optum with everything else: the pricing stack, the data exhaust from one of the largest payers on earth, and the rest of the OptumInsight machine. The government argued that even after the carve-out, Optum’s access to rival insurers’ data would chill innovation. The court basically shrugged and noted that Optum already sits on an enormous pile of data, and the government never quantified how much more it would get. That shrug is the tell. The data was the prize all along, and the editing logo was almost a sideshow.
