6 Places Your Revenue Is Leaking Before Anyone Sends a Claim

• By James
• 15-09-2026
• Technology

I was sitting in a conference room with the leadership team of an HME provider in Texas - about $30 million in annual revenue, eight locations, respiratory-heavy. Their CFO had pulled up a slide showing collections trending down for the fourth straight quarter, and the room was doing what rooms do: blaming payers.

And then their billing supervisor, a woman named Deb who'd been there nineteen years and had said nothing for forty minutes, finally spoke.

"It's not the payers," she said. "By the time a claim gets to me, it's already dead. I'm just the person who has to watch it die."
Silence. The kind where everyone realizes the smartest person in the room has been sitting quietly the whole time.

Deb was right. And her point is the single most under-appreciated truth in HME finance: your revenue cycle does not start when you submit a claim. It starts the second a referral hits your fax queue, and by the time anyone thinks of it as "billing," the outcome is largely determined.

So let's go find the leaks where they actually are - upstream, before billing ever touches anything.

1. The referral that sat in a queue for two days

The first leak isn't a billing error. It's a clock.

Every hour a referral sits unworked, three things get worse. The patient's eligibility window drifts. The referring physician starts wondering if they should have sent it elsewhere. And the documentation gets harder to gather, because the discharge planner who had context on Monday has moved on to Thursday's patients.

Most providers have no idea what their actual referral-to-intake time is. They have an aspiration ("same day!") and a reality (a fax queue where things fall behind the newer things). Nobody measures it because it doesn't appear on a financial statement.

But it's the highest-leverage number in the entire cycle. A referral worked in two hours converts at a dramatically higher rate than the identical referral worked in two days - same patient, same payer, same equipment. The only variable is elapsed time.

Go measure it this week. Pull thirty referrals, timestamp arrival and first meaningful action. The gap will surprise you.

2. Eligibility checked once, at the wrong time

Almost everyone checks eligibility. Fewer people check it again.

Here's the failure pattern: you verify coverage at intake, everything's clean, and then delivery happens eleven days later because of a documentation delay. In those eleven days the patient's plan changed, or their Medicaid recertification lapsed, or their commercial coverage terminated when they left their job - which, given they're a patient, is not exactly a rare event.

You deliver equipment to someone who was covered when you checked and isn't covered when you billed. That's not a denial you can appeal. That's just a loss.

Eligibility isn't a checkbox at intake. It's a state that needs re-verification at the moment of service, automatically, every time. If a human has to remember to do it, it won't get done on the days that matter - which are the busy days, which are most days.

3. Documentation collected reactively instead of by requirement

This is the big one, and it's structural.

Most HME providers gather documentation in response to problems. Claim gets denied for a missing CMN, so someone goes and gets the CMN, and resubmits. That loop works, technically. It also takes 45 days, consumes staff time, and only recovers claims someone remembered to follow up on.
The alternative is that documentation requirements are known and enforced at order entry, by product and by payer. HCPCS code plus payer equals a specific required document set. That's a lookup table, not a judgment call. There's no reason a human should be remembering which payer wants a face-to-face within how many days for which product category.

When providers move this upstream, denial rates don't improve incrementally. They collapse - because the vast majority of HME denials aren't disputes about medical necessity. They're paperwork that was knowable and missing.

This is precisely the class of problem revenue cycle management software exists to solve: encoding payer rules as gates the order has to pass, instead of tribal knowledge in the heads of your three most experienced people.

Which brings up the risk nobody talks about. If your compliance depends on Deb's nineteen years of memory, then Deb's retirement is a material financial event. Documented rules are also a succession plan.

4. The delivery that happened but never got confirmed

Equipment goes out. Driver delivers it. Patient signs. And then the proof of delivery lives on a piece of paper in a truck for three days, or gets photographed and texted to a supervisor, or gets entered Monday for a Friday delivery.

Every day between service and billable-confirmed is a day of float you're financing for free. At scale, it's enormous. A provider doing $30 million with an average four-day POD lag is carrying roughly $330,000 in permanently deferred cash - not lost, just always a few days behind, forever.
Worse, unconfirmed deliveries are where claims get lost entirely. Nobody bills what nobody knows happened.

The fix is unglamorous: delivery confirmation happens on the device, at the door, and drops straight into the billing-ready queue. No paper, no re-entry, no Monday.

5. Rentals that don't know what month they're in

HME rental billing is genuinely complicated - capped rentals, month counts, maintenance intervals, conversion to purchase, patient responsibility changes at specific thresholds. It's arithmetic, but it's arithmetic with a lot of edges.

And when it's tracked manually, or half-manually in a spreadsheet next to the software, two things happen. You bill months you shouldn't, which becomes a refund and an audit risk. Or you stop billing months you should, which is silent revenue you'll never notice missing.

I've never once audited a manually-tracked rental book and found it correct. Not once. The errors go both directions and they roughly cancel out on the P&L, which is exactly why nobody catches them.

6. Nobody owns the number end to end

The final leak isn't a process. It's an org chart.

Intake reports to operations. Billing reports to finance. Delivery reports to logistics. Each one is measured on their own slice - referral volume, clean claim rate, on-time delivery percentage - and every single one of them can hit their target while the overall cycle degrades.

That's how you get a room full of competent people whose metrics are all green and whose collections are down four quarters running.

Real HME revenue cycle management means one person owns referral-to-cash as a single number, with the authority to change what happens in intake because of something they saw in denials. Without that, you're not managing a cycle. You're managing six departments that happen to be adjacent.

What Deb's team actually did

They didn't buy their way out of it, though they did eventually replace their platform. They started smaller.

They measured referral-to-intake time and found the median was 31 hours. They moved eligibility re-verification to the morning of delivery. They built a documentation matrix - payer by product - and made it a hard gate at order entry rather than a suggestion. They put PODs on tablets. And they made their VP of Operations, not the CFO, accountable for days-in-AR, which caused approximately three weeks of productive conflict.

Eighteen months later: denial rate from 12% to under 5%, days in AR down nineteen, and collections up meaningfully on roughly flat volume. Same patients. Same payers. Same market everyone had been blaming.

Deb, for the record, is still there. She now spends her time on genuine payer disputes - the actual hard cases where her nineteen years of judgment are worth something - instead of chasing CMNs that should never have been missing.

That's the real return. Not just the recovered margin, though that's nice. It's that your most experienced people stop doing work that a rule engine should have done, and start doing the work only they can do.

Go find out what your referral-to-intake time is. Start there.