Compression Pumps Now Need Prior Auth — And Most Files Can't Prove the Four-Week Trial (E0651, E0652)

Two rule changes hit the same product line, eighteen months apart

Suppliers who bill compression pumps have absorbed two structural changes since late 2024, and the second one only works if you handled the first one correctly.

The first: LCD L33829, Pneumatic Compression Devices, was retired effective November 14, 2024, along with its policy article A52488. That did not loosen anything. It moved the governing text. Coverage for E0650, E0651 and E0652 now runs off NCD 280.6 plus the Standard Documentation Requirements article, A55426, and the DME MAC coverage tips built on top of them. Intake teams that are still validating against a checklist printed from the old LCD are checking a document that no longer decides the claim.

The second: on April 13, 2026, CMS expanded the required prior authorization list to include E0651 and E0652. Segmental compressors, with and without calibrated gradient pressure, now go through a pre-claim review before delivery. E0650, the non-segmental compressor, stays outside the PA program, but it lives under the same coverage criteria and the same audit exposure.

That's the part worth sitting with. A non-affirmation costs you a delivery delay. The same gap, discovered after payment on a retired-LCD-era claim, costs you the money back plus interest. Compression pumps have been an approved Recovery Audit topic since 2019 under issue 0131, complex review, all DME MACs, aimed squarely at medical necessity and documentation. Nothing about that changed when the LCD went away.

The four criteria a reviewer is actually scoring

Every PCD determination, whether it arrives as a prior auth decision or an ADR six months after delivery, comes down to four things the medical record has to establish. They read like a coverage summary. They function like a scoring rubric.

One, the patient's diagnosis and prognosis. Two, symptoms and objective findings, including measurements that establish how severe the condition is. Three, the reason the device is required, which means the treatments that have been tried and failed. Four, the clinical response to an initial treatment with the device.

Read the fourth one again. The record has to show how the patient responded to treatment with the pump. And the DME MAC guidance is specific about timing: initial treatment with the PCD must be documented by a clinician on or before the date of delivery. It has to include measurements. It has to be part of the medical record, signed by the treating practitioner, with concurrence if the note came from a licensed or certified medical professional.

So the sequence most intake workflows assume is backwards. You do not deliver the pump and then chart the response. Someone has to trial the patient on the device and document the measured result before the delivery ticket is dated. If your PA package goes out with a diagnosis, an order, and a promise that response data is coming, you have submitted three of four criteria.

Where the four-week trial goes wrong

For lymphedema, coverage sits behind a conservative therapy trial of at least four weeks that failed to produce adequate improvement. Most suppliers know this. Far fewer can prove it from the file.

The trial is not one thing. It has three components, and the record has to show all of them for the full duration:

Compliant use of a compression bandage or garment, and the note has to establish that the garment delivered adequate compression at sufficient pressure, was prefabricated or custom fabricated, and was graduated. A referral note that says "patient tried compression stockings" satisfies none of those four qualifiers. Regular exercise. Elevation of the limb.

Then the failure has to be visible. "Patient did not improve" is a conclusion. Circumferential or volumetric limb measurements at the start of the trial and at the end of it are evidence. If the file has no starting measurement, there is no arithmetic a reviewer can do, and no amount of clinical narrative fixes that.

The E0652 problem nobody wants to hear

E0652 is the calibrated gradient pressure segmental compressor, and it is the code with the narrowest path to coverage. It requires everything E0651 requires, plus one more finding: the individual has unique characteristics that prevent satisfactory treatment with a non-segmental device used with a segmented appliance, or with a segmental device that lacks manual pressure control in each chamber.

The DME MAC guidance on this is unusually direct. The documentation should include a description of the unique characteristics. Listing only "unique characteristics" in the medical record will result in denial.

That note exists because reviewers kept seeing the phrase copied verbatim out of the policy and pasted into progress notes as if quoting the criterion satisfied it. It doesn't. The record has to say what is unique about this specific patient's limb, wound, or anatomy, and why the lesser device could not do the job for them. Sensitive skin over a healing ulcer, a limb contour that prevents uniform pressure distribution, documented intolerance of an uncalibrated device during the trial. Something a reviewer can point at.

What the PA package has to contain

The submission itself is short, which is deceptive. Three things: the PCD prior authorization coversheet, the Standard Written Order, and medical record documentation supporting the coverage criteria. Submit through the Noridian Medicare Portal and the coversheet is not required, since the portal fields replace it. Fax, mail, and esMD all need it.

Route by the beneficiary's permanent address, not by where the referring practice sits. Jurisdiction errors are one of the named rejection reasons, along with submitting a code that isn't subject to PA and duplicating a request you already have in flight. Those rejections don't get reviewed. They just come back.

Where PCD claims and PA requests actually fail

Sorted roughly by how often each one clears intake and then falls over:

Notice how few of these are billing errors. Almost every one is a fact that either exists in the clinical record at intake or does not, and the intake coordinator is the only person in the chain positioned to find out while it's still fixable.

What a non-affirmation actually costs you

Less than people assume, if you catch it early. You can resubmit as many times as you need to, and a resubmission after gathering the missing note is the normal path. The five-day clock restarts, but the claim is still alive.

The expensive version is different. A patient discharged Friday with a pump ordered for Monday, no trial documentation in the chart, and a family expecting delivery. Now you're choosing between a delivery you can't bill, an ABN conversation nobody prepared the patient for, and a delay you have to explain to a referral source. The DME MAC guidance is blunt about prevention: work with referral sources to establish necessity before the item becomes urgent, and start gathering discharge documentation on admission rather than on discharge day. Expedited review exists, at two business days, but only when the standard timeframe would jeopardize the patient's life or health. Scheduling pressure isn't that, and requests that don't meet the bar get downgraded to standard anyway.

Pre-submission checklist

Run this before the PA package goes out, and again before the claim goes out after affirmation.

The gap is upstream of billing

Look at what the four criteria require and it becomes obvious where the failure sits. Diagnosis and prognosis, objective measurements, therapies tried and failed, response to initial treatment. Every one of those is a clinical fact created by someone outside your organization, in a note you receive as a fax or a portal export, weeks before anyone bills anything.

An intake coordinator working through eighty referrals a day is being asked to read a twelve-page clinical packet and determine whether a four-week trial is documented with all three components and bracketing measurements, whether the response note predates the delivery ticket, and whether an E0652 justification describes anything or merely names it. That is not a fast read. Under a five-day PA clock with a 60-day delivery window behind it, it is also not optional.

Which is why the practical move is not more training. It's putting the check earlier, where a missing baseline measurement costs a phone call to the referring office instead of a non-affirmation, a delayed delivery, and an awkward conversation with a patient who was told the pump was coming.

What to do this week

1. Pull your last 90 days of E0651 and E0652 submissions and count non-affirmations by reason

The decision letters tell you exactly which criterion failed. If trial documentation is your top reason, the fix is a referral-source conversation, not a billing process change. If it's the response-to-treatment note, the fix is your delivery sequencing.

2. Find every intake template that still references L33829 or A52488

Retired almost two years ago and still quietly in circulation at a lot of suppliers. Anything validating against them is validating against nothing.

3. Audit the E0652 files you already delivered

These predate prior authorization, which means nobody checked them before the money moved. If the phrase "unique characteristics" appears without a description behind it, you have located your extrapolation risk. Better to find it now than to have a UPIC find it in a sample and apply the rate to the population.

Prior authorization changed when you find out. It didn't change what the record has to say. The suppliers who will move through the five-day window without friction are the ones whose intake process was already reading for those four criteria before April.