A firm built around one question.

Every charge on a claimed bill resolves to two calls, and they are kept separate because they answer different questions. The first is the chart determination: does the record show this charge involves treatment of the injury complained of, does it not, or is it mixed. That call is made from the record alone and cites the page it rests on.

The second is the causation screen: would the claimant have required this care regardless of the events at issue. Related, not related, or gray. A charge can plainly appear in the treatment record and still be care the claimant needed anyway, and the two questions coming apart is often the most useful thing a review can show.

Where the record cannot answer, the call is GRAY and the analysis names the specific document that would resolve it. A gray call is not a hedge. It is an instruction.

The method comes out of more than twenty-five years defending providers in professional liability matters. The same pattern recurred in file after file: the related and unrelated charges were separable from the record itself, and nobody had the hours to separate them. An adjuster holding twelve thousand pages and a demand with a total at the bottom was being asked to test that total without the time to read what it rested on.

That is a records problem before it is a valuation problem, and it is the part of the work that a written protocol and a disciplined read can actually solve. Etiology does that part, cites it, and stops.

Every engagement runs to a written protocol. The production is inventoried and labeled, duplicates are proven rather than assumed, and pages that carry no searchable text are rendered and read rather than skipped, because keyword search over a scanned record returns nothing and reports no error.

The analysis runs on Anthropic's Claude models under zero data retention terms: nothing a client sends is kept by the model after the review, and no client record is ever used to train any model. Anthropic is the only AI provider in the pipeline, and it is named on this site for that reason.

Before anything is delivered, every quotation is verified against the page it cites, every subtotal is recomputed from the line items, and the scope is audited to confirm no valuation, recommendation, or adjusting language reached the work product. A determination that cannot survive that pass does not ship.

No settlement valuation. No recommended authority or reserves. No opinion on standard of care, breach, or liability. No recommendation about what a payer should pay, and no advocacy for either side of a claim.

This is not modesty and it is not a limitation of capability. A determination is only useful to an adjuster if it was made without regard to what the answer would cost. The moment a reviewer is reaching for an outcome, every call it makes becomes an argument rather than a finding, and the payer has to re-check all of it. The boundary is what makes the output relied upon.

The adjuster applies the law and decides the claim. Etiology supplies the factual input, and it is useful precisely because it decides nothing else.

The work is performed and supervised by the firm's founding member, a healthcare attorney holding a JD, an MHA, and Fellowship in the American College of Healthcare Executives. Etiology is a Louisiana LLC in good standing; its formation and its registered agent are a matter of public record with the Secretary of State.

But the more useful answer is that a determination here does not rest on who made it. The protocol is written down. The verification record is producible. Every call cites a Bates-labeled page, and the labeled record set is delivered with the analysis, so a reader who doubts a determination can pull that page and check it in about a minute. That is a deliberate design choice: the work is built to be audited by the person paying for it, not accepted on authority.

If a call is wrong, it is wrong in a way you can demonstrate. That is the standard the firm holds itself to, and it is the reason the citation travels with every finding.