AI applications to reduce medical cost

COMPLEX HOSPITALIZATION BILLS · LATIN AMERICA

Six in ten findings held on bills the team had already closed.

10x
faster than the five days a manual review takes
6 in 10
findings the team’s own reviewer validated
100%
of accounts reviewed, with no value floor

Background

The team decides which lines of a hospital bill the record supports.

The client runs clinical review on complex hospitalization bills. Its reviewers work through a bill item by item against the medical record, the clinical guideline and the contract that priced each service.

Every item the review clears or challenges carries the reason behind it. What the team can cover is set by how many accounts it opens and how long each one takes.

Challenge

Under a certain amount, the account closed unread.

A complex hospitalization bill runs to thousands of pages. To decide one line on it, a reviewer reads the clinical record behind it, which can run to a thousand pages of its own, then the clinical guideline, the regulation, and the contract terms that priced the service. One bill takes about five days of one person working on nothing else.

The bills that carry that cost are the ones at a hundred thousand dollars and up. Below a certain amount, the team reviewed nothing at all. The reading cost more than the account could return, so those accounts were closed without a reviewer opening them.

Solution

The agent reads every line and returns a reason for each finding.

Falcon’s agent connects to the sources the team already holds: the bill, the clinical record, the guidelines, the regulation, and the team’s own clinical rules. It reads the whole bill against all of them and returns each finding with the rule it applied and the evidence in the record behind it.

The reviewer accepts or rejects each finding, and every decision is written back into the model, so the next bill runs on it. Bills run in parallel, so how much the team can review depends on how many findings its reviewers confirm.

Mechanism

Where each finding comes from.

CLINICAL RECORD GUIDELINE CONTRACT MEDICAL BILL CLINICAL RULES ITEM BY ITEM FINDINGS REVIEWER EVERY VERDICT UPDATES THE MODEL CLINICAL RECORD GUIDELINE CONTRACT MEDICAL BILL RULES ITEM BY ITEM FINDINGS REVIEWER VERDICTS UPDATE THE MODEL

Fig. 01 · One bill, read against the record, the guideline and the contract

“It evaluates items a normal audit does not reach, given the time available to find them: routine lab work, low-cost medications, low-cost supplies.”

HEAD OF CLINICAL REVIEW · LATIN AMERICA

Results

01
The team’s own reviewer validated six in ten findings
The team’s clinical reviewer checked every finding against the record, the scientific evidence and the fee schedule. Six in ten held.
02
Four in ten of those were new to the team
Four in ten of the validated findings were absent from the team’s original review of the same bills. Each one sat in an account the team had signed off.
03
Charges the record did not yet support
Reading the bill against the clinical record catches charges the record does not yet support, item by item, with the gap named.

For providers

Every account gets read before it leaves the building.

A hospital runs the same agent on its own accounts inside the bill hold window. It reads each item against the clinical record, the guideline and the contract, and returns the same finding with the same evidence behind it, while the account can still be corrected.

Most of what it returns is care the hospital delivered and documented thinly: charges the record does not yet support, and services that never reached the account at all. Correcting those inside the bill hold window takes a note in the record. Correcting them after the account goes out takes an appeal.

Next steps

Start reducing medical cost through contracts,
authorizations and bills.

Get a demo

See how Falcon helps healthcare organizations reduce medical cost.

Thank you. We will be in touch at the email you shared.