RCM Performance Metrics for Multi-Site Infusion Platforms

Adonis's 2026 survey of more than 120 RCM leaders across 18 specialties found payer behaviors have surpassed internal challenges like staffing as the largest risk to revenue growth, a decisive reversal from prior years when internal optimization was the primary lever. That reversal matters more for infusion than for almost any other specialty, because a single payer's adjudication rule change can degrade revenue performance across an entire multi-site platform without any individual site doing anything wrong. A staffing shortage is visible, local, and fixable with a hire. A payer policy shift is none of those things: it can hit every site simultaneously, it often arrives without warning, and it can look, on a summary dashboard, exactly like a performance problem rather than an external one. That distinction is the starting point for everything this piece covers. If the dominant threat to revenue now originates outside the organization, then the metrics an infusion platform relies on have to be built to detect external, payer-driven shifts and separate them from internal execution failures, not just track internal throughput.

What generic RCM benchmarks miss in infusion billing

The RCM metrics most platforms already report, overall denial rate, days in AR, net collection rate, were built for specialties where the average claim carries modest financial weight. Infusion breaks that assumption, because a single claim can carry a five-figure drug acquisition cost, and standard benchmarks are too blunt to catch the failure modes specific to that exposure. A blended denial rate, for instance, rolls every claim type into one number, but infusion denials cluster disproportionately in high-dollar biologics and J-code drug claims, where a denied buy-and-bill claim leaves the practice holding the acquisition cost with nothing to show for it, exposure that runs from $5,000 to $50,000 per infusion encounter, a financial exposure per denial that is categorically different from a routine E&M claim.

The same blindness occurs in AR aging, where standard RCM benchmarks are too blunt to surface the failure modes specific to infusion billing, since a single claim can carry five-figure drug acquisition cost, and a blended aging report hides the specific risk inside drug-cost AR by mixing it in with low-value administrative claims that carry none of the same urgency. Generic platforms compound the problem by tracking denial volume without denial root cause by payer, so a multi-site operator has no way to tell whether a spike is a documentation problem at one clinic or a payer-wide adjudication change hitting every site at once. The result is a platform that can look healthy on a summary dashboard while individual sites quietly bleed revenue through authorization lapses, underpayments, and site-of-care denials that never surface as their own line item, and needs its own metric framework for exactly this reason, one built around prior authorization cycle time, denial rate by payer and root cause, J-code and buy-and-bill integrity, and ASP-based underpayment detection, each covered in turn below.

Prior authorization cycle time as a leading indicator of infusion revenue health

Prior authorization cycle time deserves to be read as a leading indicator of revenue health, not filed away as an administrative throughput number. For a chronic infusion population, an authorization lapse doesn't just slow down a workflow, it stops the treatment and the reimbursement for that treatment at the same time. Authorization in infusion is rarely a one-time event: ongoing treatments require reauthorization once the initial approval period runs out, and some Marketplace and Medicare Advantage plans demand a fresh approval every year even for patients who have been stable on the same therapy for years. That makes authorization a recurring operational cycle that has to be tracked as one, rather than a box checked once at intake.

The gap between good and average performance here is stark. That figure only has diagnostic value when it's broken out by payer and by drug rather than reported as one blended number, because a 95% approval rate platform-wide can still be masking a single payer or a single drug class where approvals are consistently slow or getting denied. Tracking first-pass approval rate, average calendar days from referral to authorization, the reauthorization gap rate among ongoing patients, and the share of authorizations started before versus after the treatment appointment gives a multi-site operator the resolution needed to tell a systemic payer problem from a localized process failure at one site.

Regulatory momentum here is real but incomplete. The National Association of Insurance Commissioners has pushed for prohibiting retroactive denials on preauthorized care and for requiring authorizations to stay valid for at least a year regardless of dose changes, and for the full duration of treatment in chronic condition patients. That reform has not taken effect uniformly across payers, leaving multi-site platforms to manage a patchwork of payer-specific authorization timelines rather than a single standard. Some operators still treat auth cycle time as a scheduling concern rather than a revenue metric. Because authorization has to begin before the treatment cycle starts, a lapse that only lives in scheduling data does not reach the billing system until the denial occurs, and by then recovery is expensive and sometimes impossible.

Denial rate by payer and root cause in infusion

A blended denial rate is a lagging number that tells an infusion platform almost nothing actionable on its own. The number that matters is denial rate segmented by payer and by root cause, because that is what actually shows where revenue is leaking and why. Top-quartile ambulatory infusion centers maintain denial rates below 7%, while mid-market operators and the broader industry average run considerably higher, and the spread between average and best-practice represents recoverable revenue. That target number is close to meaningless, though, unless it's broken apart by root cause, because medical necessity and documentation failures, authorization-related denials, site-of-care denials, J-code and coding errors, timely filing misses, and payer adjudication rule changes each demand a completely different fix upstream.

Site-of-care denial deserves isolation as its own category because it is growing and illustrates the whole argument for disaggregation. As of July 1, 2025, CareFirst expanded the list of medications requiring prior authorization and site-of-care management for commercial members. The physical location where a patient receives an infusion is now itself something a payer has to approve. Failure to obtain that site-of-care approval results in a denied claim and can mean the patient gets redirected to a home, ambulatory, or physician-office setting instead of the site originally scheduled. In a blended denial rate, a policy change like that is invisible: the number might tick up slightly platform-wide, with no way to tell what caused it.

Documentation-driven denials deserve the same isolated tracking, since inadequate or inaccurate records that fail to connect clinical reasoning to the billed code are among the most common causes of denial in infusion, and payer scrutiny of medical necessity is only intensifying. Adonis's 2026 survey found RCM teams spending substantial hours each week on denial-related work, and without root-cause segmentation that labor gets spread reactively across whatever claim is in front of someone rather than aimed at the categories producing the highest volume or the highest dollar impact. Comparing denial rate by payer across sites shows where to act: a spike that appears at every site carrying the same payer points to a payer-driven issue that needs to be escalated at the contract level, while a spike concentrated at a single location points to a local process correction instead.

J-code error rate and buy-and-bill claim integrity as a distinct financial risk metric

J-code unit accuracy and buy-and-bill claim integrity form a risk category that has no real equivalent anywhere else in general RCM. Buy-and-bill billing means the practice purchases the drug before it ever submits a claim, so the financial exposure is in the acquisition cost itself, held against a claim that hasn't resolved yet, and that exposure needs its own metric separate from standard denial tracking because by the time a denial shows up in AR, the practice has already paid for the drug. A denied buy-and-bill claim leaves the practice holding that acquisition cost with no reimbursement at all, and at typical per-encounter drug costs running from the thousands into the tens of thousands of dollars, a cluster of J-code denials can turn into a cash-flow emergency rather than an ordinary billing backlog.

The scale of the underlying error problem is large enough to demand its own benchmark. The estimated J-code unit error rate for a typical oncology or infusion practice runs 15% to 20% of drug claims, while best-practice performance with systematic dosage review in place sits well under that threshold, making this one of the single most actionable numbers available to a practice that hasn't yet built unit verification into its workflow. The code set itself is also a moving target: CMS intensified enforcement around product-specific J-codes and quarterly ASP pricing submissions in 2026, a requirement directed primarily at manufacturers, not billing practices, alongside additions and deletions to the HCPCS Level II code set, so a practice that isn't tracking its own code-level accuracy against the current list risks denials simply from using a deleted code, separate from any billing error in the traditional sense. Modifier compliance adds another layer: CMS requires the JW modifier for discarded drug amounts and JZ when there's no wastage, and a single-dose claim missing either one becomes an audit target, while NDC format errors, specifically converting a 10-digit code to 11 digits incorrectly, trigger electronic rejection before the claim even reaches pricing validation.

Electronic prior authorization for Part B J-code drugs has been proposed under the CMS-0062-P rule, but it has not been finalized and wouldn't take effect until October 1, 2027 at the earliest. Practices still running manual PA workflows for these drugs are carrying both authorization risk and compliance risk at the same time, with no clear end date on that exposure. The metrics to establish at the platform level are J-code unit error rate broken out by drug and by site, modifier compliance rate for JW and JZ, NDC format rejection rate, and buy-and-bill claim days-to-resolution, tracking the time from submission to either payment or final denial as a proxy for how long the practice's own capital is tied up in limbo.

ASP-based underpayment rate and line-level remittance reconciliation in infusion

Revenue leaks out of infusion platforms through underpayment on claims that were never denied at all, and standard payment posting only catches that leak if remittances get reconciled at the line level against contracted and ASP-based rates. A paid claim looks like a success in a standard billing report regardless of whether the payer actually paid what the contract called for, so an underpayment is structurally invisible unless someone compares expected reimbursement against actual reimbursement, line by line, by CPT and by J-code. Nothing about a shortfall like that trips an alarm on its own. It has to be found.

The gap between average and best-practice performance here is wide. ASP mismatch is the chronic mechanism behind a large share of this leakage: many infusion drugs get reimbursed based on Average Sales Price, and when a drug's acquisition cost climbs faster than its ASP figure gets updated, the provider loses margin on every single treatment using that drug until the pricing catches up. Guarding against that erosion takes regular internal audits of ASP pricing updates alongside a monthly review of infusion CPT and HCPCS changes, since both move on their own schedules and neither waits for a practice to notice.

For a multi-site platform, comparing underpayment rate by payer and by site reveals whether a specific payer is systematically short-paying across the whole network or whether the shortfall is concentrated at sites that simply haven't built line-level reconciliation into their process yet. The metrics to establish are underpayment detection rate, the share of paid claims actually verified against the contracted rate; underpayment recovery rate, how many identified underpayments get successfully appealed; and net underpayment loss as a percentage of net revenue broken out by payer and by site. Payer performance scorecards that track expected against actual reimbursement by CPT and J-code are the operational tool that makes these patterns visible in the first place. Without them, underpayments simply age until they're indistinguishable from an ordinary contractual write-off, and revenue that was owed to the practice disappears without ever being counted as lost.