Revenue per Infusion Benchmarks for Independent Infusion Groups

Most independent infusion groups track the wrong things. Visit counts, charge volume, patient throughput, these numbers tell a practice how busy it is, not how healthy it is. Revenue per infusion is different. It reflects everything else happening in a practice's operations: whether claims get coded right the first time, whether denials get worked or written off, whether underpayments get caught, and whether prior authorization gets handled before the drug ever leaves the shelf.

A practice can see patient volume climb month over month and still watch revenue flatten or slide. That gap is the tell. When the gap between rising volume and flat or falling revenue appears, it means something in the billing chain is broken, and volume alone will never fix it.

The dollar stakes are what make this worth watching closely. A single infusion encounter can generate anywhere from $5,000 to $50,000 in charges depending on the drug. At that scale, a 5% loss isn't a rounding error, it's a five-figure write-off on one visit. Multiplying that across a caseload makes the number serious fast.

The market environment that makes RCM execution increasingly consequential

The ambulatory infusion center market is projected to top $52 billion by 2035, growing at roughly 8.9% a year. It's a fast-growing lane, not a niche corner of health care anymore, and it's getting more crowded with complexity, not less. It's a fast-growing lane, and it's getting more crowded with complexity, not less.

More than 60% of drugs currently in the development pipeline are infusion-delivered therapies. The billing complexity infusion practices deal with today is the floor. Every new biologic that reaches the clinic brings its own J-code, its own dosing logic, its own payer rules.

Site of care is doing a lot of the structural work here too. Hospital-based infusion runs $5,500 to $11,500 per encounter by cost, compared to $1,700 to $3,500 for home infusion and $3,500 to $5,000 in ambulatory centers and physician offices. Payers see that spread and have every financial incentive to push volume out of hospitals and into independent ambulatory infusion centers (AICs).

That shift favors independent infusion groups, on paper. But the benefit only shows up if the group can actually bill and collect accurately on the patients coming through the door. Volume without collection discipline just means more exposure, not more revenue.

Denial-Driven Revenue Compression and the Distribution That Hides the Cause

Denials aren't holding steady. They climbed from 30% in 2022 to 38% in 2024, and claim denials climbed over 11% in 2025. This is a worsening trend, not a plateau, and infusion sits closer to the center of it than most specialties because nearly every biologic requires prior authorization before it's given.

Prior authorization-related denials alone grew more than 20% over the past two years. Break the denial pile apart by cause and the picture gets clearer:

• Missing or invalid prior authorization: 20-25% of denials
• Eligibility or coverage issues: 15-20%
• Medical necessity disputes: 10-12%
• Coding errors, things like modifier misuse or CCI edit violations: 10-12%
• Late submission: 5-8%

Lump all of that together under one label, "denials," and the signal disappears. A missing PA and a modifier error have nothing in common as problems. One is a front-office workflow gap, the other is a coding-team training issue. Fix them the same way and neither gets fixed.

Prior authorization as the upstream lever that determines whether revenue

Authorization failures are 20-25% of all denials, and that stat alone undersells the damage. In a buy-and-bill infusion, the practice purchases the drug before it's given. If the PA falls through after the infusion has already happened, the drug cost is already spent. There's no walking that back.

Recent regulatory pushes to ease PA burden haven't touched physician-administered infusion drugs, for either government or commercial plans. The operational weight of getting authorization right stays entirely with the practice. Nobody is coming to lighten that load.

Most PA delays trace back to a small set of documentation gaps: clinical notes that are missing or inconsistent, diagnosis codes that don't clearly support the treatment, or no evidence on file that step therapy was tried first. That last one is the most common reason payers kick a request back asking for more information.

State law changes the calculus in some places. New Jersey requires Medicaid, SHBP, and SEHBP plans to grant step therapy exceptions when the required drug is contraindicated, likely to cause harm, expected to fail, or when every formulary option has already been tried and failed. Connecticut goes further for multiple sclerosis and rheumatoid arthritis treatments: insurers can't require step therapy at all for those conditions. A separate, more general Connecticut provision caps step therapy at 30 days for any drug, after which the provider can call the treatment ineffective and get the originally prescribed drug authorized, assuming it's covered under the plan.

These protections are narrow. They apply to specific payer types and specific conditions, so a practice has to know exactly which patients qualify before assuming the rule applies. Assuming wrong costs just as much as not knowing.

J-code and buy-and-bill billing errors: how coding complexity suppresses collected revenue

Buy-and-bill only works when the billing behind it is airtight. The practice fronts the cost of the drug, and if the claim comes back denied, underpaid, or wrong on units, the acquisition cost has already left the building. There's no partial version of this risk. It's full exposure either way.

The coding landscape isn't getting simpler. CMS maintains a large and actively updated catalog of J-codes in the HCPCS Level II system. Each annual update brings new codes and retires others. CMS rejects any claim submitted with a discontinued code automatically. A practice running last year's charge master into this year's claims is generating denials it doesn't even know are coming.

The errors that occur most often follow a pattern:

• Incorrect units. J-codes bill in specific unit increments, and getting that number wrong on a high-cost biologic can mean a serious underpayment or an outright denial.
• Drug code mismatches, where the code on the claim doesn't match what was actually administered against the J-code billed.
• Modifier errors, wrong ones or missing ones, that trip CCI edits and bounce the claim automatically.
• Billing a code that got discontinued in the annual HCPCS update, because the system wasn't updated in time.

None of these are exotic mistakes. They're the routine, grinding kind that happen when a charge master isn't reviewed on a quarterly cadence against the current code set.

Underpayments hidden in remittances: the revenue

An underpayment doesn't behave like a denial. A denial gets flagged, it lands in a worklist, someone has to touch it. An underpayment just gets posted. It sits there quietly and ages out of the system, and most RCM platforms never raise a hand about it.

Infusion remittances make this worse because they're structurally more complicated than a typical office visit claim. One encounter might carry separate line items for the drug, the administration, and supplies, each with its own contracted rate, its own modifier, its own unit count that has to reconcile against what was billed.

The common failure points:

• The payer applies the wrong contracted rate, which occurs constantly in multi-payer environments where rates vary by drug class.
• Units paid don't match units billed. On a high-cost biologic, even a one-unit gap is real money.
• Administration codes get bundled or downcoded with no clinical justification behind it.
• The ASP-plus-6% calculation gets run against a stale ASP baseline, because the payer's rate table wasn't updated.

Without reconciling every claim line by line, comparing the bill, the contract terms, and the actual payment at the code and unit level, an underpayment looks identical to a correct payment in any aggregate report. It hides in plain sight.

Revenue per infusion across practice types and the scarcity of published benchmarks

There's no clean, public benchmark for revenue per infusion at independent AICs. What figures exist tend to come out of internal RCM analytics, private surveys, or valuation work done during a sale or acquisition, none of which get published for general reference.

What the available data does confirm: infusion visits run $5,000 to $50,000 in charges, a wide range driven by therapeutic category, from IV iron on the low end to high-cost biologics for oncology or autoimmune conditions on the high end. Cost benchmarks from NHIF put home infusion at $1,700 to $3,500, ambulatory centers and physician offices at $3,500 to $5,000, and hospital settings at $5,500 to $11,500. Those are cost figures, not revenue figures, but they set a floor for what payers are willing to pay at each site of care. Separately, health systems running home infusion programs report averaging around $20 million a year in revenue from those programs, a useful data point at the program level, though it doesn't translate to a per-encounter rate.

The spread is wide for structural reasons, not random ones. Therapeutic mix drives most of it: a center doing mostly iron infusions will run a lower revenue-per-infusion number than one focused on rheumatology or neurology patients getting expensive biologics, and that's true even if both centers are billing perfectly. Payer mix adds another layer, since commercial buy-and-bill contracts pay differently than Medicare Part B at ASP plus 6%, which pays differently again from Medicaid. Site-of-care designation shifts the applicable fee schedule too, even for the identical drug given to the identical patient.

Given all that, a single industry-wide revenue-per-infusion number wouldn't mean much even if one existed. The benchmark that actually matters is internal: specific to the practice, broken out by therapeutic category and payer mix, and tracked as a trend over time rather than compared against some external average that doesn't share the same patient mix.

The RCM operating model choices that move revenue

Revenue cycle control starts at scheduling, not at the moment a claim goes out the door. By the time a claim is submitted, most of what could have been caught, a missing authorization, a wrong eligibility check, a drug not yet approved under the patient's plan, has already happened. The claim just makes the damage visible.

Denials, PA failures, coding errors, and underpayments are produced by operational failures further up the chain, and that operational reality applies to every number in this piece. Denials, PA failures, coding errors, underpayments: none of them are billing department problems in isolation. Denials, PA failures, coding errors, and underpayments are upstream problems, and they become visible downstream, on a remittance, weeks after the point where they could have been prevented for free.