Federal IDR guides
What does a bad IDR submission cost?
The entity picks one offer, so the number is the submission. What a low ask, a high ask with nothing behind it, and a sloppy packet each cost you.
Verified September 11, 2026
Short answer
The IDR entity picks your offer or the payer's. It cannot land in between. So the biggest mistake is not a thin packet. It is the wrong number.
Ask too little and you win small. Ask too much with no reason behind it and you lose, and the entity keeps your fee. The packet's job is to make the number believable.
Asked under 150% of QPA
86%
Won
Asked 300 to 500%
90%
Won
Asked over 1000%
85%
Won
Time to a decision
29 days
Business days, single dispute
Three ways to get it wrong
| The mistake | What it costs | How often it happens |
|---|---|---|
| Asking too little | The gap between your ask and what the entity would have accepted. Paid on every line, forever. | Most common, and invisible. It shows up as a win. |
| Asking too much with no support | The IDR entity fee, around $498 on a single dispute, plus $15 to CMS, plus about 29 business days of waiting | Rare when the number has reasons behind it |
| A packet that skips the factors | A weaker case at the same number. Or a loss on a number you should have won. | Common with home-made templates |
Asking too little is the expensive one
In 2025, providers who asked less than 150% of the QPA won 86% of lines. Providers who asked 300 to 500% won 90%. Providers who asked over 1000% still won 85%.
Those are different claims, not the same claim priced two ways. The data cannot tell you what happens if you raise one offer. What it does say is that across millions of lines, the groups asking a lot did not win less often than the groups asking a little. A group that always asks 140% of the QPA is leaving the question untested.
That money does not show up on any report. Nobody files a complaint about a win.
Asking too much with nothing behind it
The entity has to put its decision in writing and say what it relied on. It must weigh six things: the QPA, the provider's training and quality, market share, how complex the patient was, teaching status and case mix, and both sides' good-faith efforts. It may not use billed charges.
A high number with no reason under those six headings looks like a billed charge with a different label. That is how a strong claim loses.
What a good submission does
It picks a number the data supports for this payer, this code, this market, and this IDR entity.
It gives a short reason under each of the six factors. Not a paragraph. A reason.
It says why the QPA does not fit this case. Not that the QPA is wrong.
It never mentions billed charges.
The math on one claim
Take a claim with a QPA of $1,000. The payer offers $1,000. Say the payer has paid nothing yet.
Ask $1,400 and win, and the total the plan owes is $1,400.
Ask $4,000 and win, and the total is $4,000.
Ask $4,000 with nothing behind it and lose, and the total is $1,000, and you are out about $513 in fees.
If either offer would have won, the second is worth $2,600 more than the first. The public data does not tell you whether the higher ask wins as often on the same claim, so treat that $2,600 as what is on the table, not as money you were guaranteed.
One more thing the arithmetic hides. IDR sets the total the plan owes, not a bonus on top. Anything already paid, and the patient's cost sharing, come out of it. If the award lands below what the payer already paid, you send money back.
Recourse
Where Recourse fits
Recourse sets every offer from what the data says wins for that payer, code, and entity, and writes the packet to support that number under all six factors.
Sources
45 CFR 149.510(c)(4) and (c)(5); CMS Federal IDR Guidance for Disputing Parties, sections 8.1 and 8.2; CMS Federal IDR Public Use Files, 2025 Q1 to Q4, line level, air ambulance excluded, deduplicated to the latest quarterly record. Win rates by ask band are provider offer selected over decided lines, ask expressed as a percent of the QPA. The fee figure is the median compensation recorded on single disputes where a positive amount was recorded. The time figure is the median for single disputes, in business days from initiation to closure.
Every number on this page comes from public CMS files or the federal rule. Our methodology explains how we count. Questions? sales@recoursehealth.com
Your data, not the national average
See how your group's IDR record compares.
A free report card built from the same CMS files, measured at the group level. Nothing needed from you.