Scorecard for Evaluating Operator-Embedded RCM Models in Multi-Site Infusion Platforms

An infusion-only RCM platform for buy-and-bill practices should be evaluated on whether experienced operators own revenue-protecting decisions at every site. This scorecard tests six capabilities, from recurring authorization through cross-site analytics. Local judgment and what sites learn from one another must work together before the same revenue loss happens again.

Why multi-site infusion platforms expose RCM failures that single-site evaluation frameworks miss

A payer’s new site-of-care restriction or tighter authorization criterion for a biologic changes more than billing. It affects intake, benefits verification, authorization, clinical documentation, claims, and follow-up. Each site must make the change part of its daily work. If one coordinator follows the old process while another interprets the requirement differently, claims enter the revenue cycle with inconsistencies that delay reimbursement. A single-site evaluation can miss this distribution problem because it tests whether expertise exists at one location.

The practice has already purchased and administered the drug when an authorization lapse, administration coding error, or payment below acquisition cost becomes visible. An administered biologic dose cannot be taken back. The AleraCare and Pure Healthcare merger into a 77-site ambulatory infusion network illustrates the scale at which consistent local execution must now work.

What “operator-embedded” actually means, and why the definition matters for evaluation

An operator-embedded model puts experienced human judgment into the workflow at the point where people make decisions. It does not reserve that judgment for a central team reviewing exceptions afterward. A site coordinator without the knowledge or authority to catch an authorization mismatch does not provide the same protection as an operator who owns that decision. Both arrangements may be described as embedded, but only one puts accountability where the error can be prevented.

That definition shifts evaluation upstream of claim submission. Billing can uncover incomplete benefits verification, missing medical necessity documentation, or a treatment change that never reached the authorization record. It cannot undo the treatment already delivered. Rules engines can flag known errors, but experienced operators must make sense of unfamiliar denial reasons, how each payer behaves, and when an appeal needs escalation. Those operators also need to share analysis across sites so that what they know can help other locations.

Prior authorization lifecycle management as the first scorecard dimension

Prior authorization is a recurring obligation. Site-of-care reauthorization differs by payer and clinical exception category, and it needs up-to-date documentation that shows the patient still meets the criteria. The scorecard should therefore test whether deadlines are tracked by patient and site, with alerts before expiration. Waiting for a denial is too late. For a buy-and-bill practice, a missed renewal can put reimbursement at risk after the drug has been purchased and administered.

The next test concerns treatment changes before the next infusion. Evaluators should establish whether a change in dose, frequency, or product triggers a defined authorization review, and who owns that review. Authorization status should be visible to the site operator in real time rather than depend on a manual lookup that busy staff may skip. The model should also compare billed codes and units with the authorization before submission.

Cross-site reporting should then show authorization denials by payer and diagnosis. When a documentation problem keeps happening at one location, it should shape how other sites do the work. The evidence should include a workflow for preventing expired approvals. That distinction separates lifecycle management from retrospective counting.

J-code and drug unit accuracy as the second scorecard dimension

Once authorization matches the treatment, the claim must correctly show which drug was administered. The scorecard should test how the operator selects the J-code, matches the National Drug Code, and checks drug units, and whether the operator has up-to-date reference data before submission. CMS updates its ASP files and crosswalks quarterly. A mismatch between claim data and those files can produce denials, payment discrepancies, or audit inquiries. The evaluation should establish who updates the crosswalk and how that update reaches each site.

New biologics create a separate test of judgment. A product may launch without a permanent J-code and require a not-otherwise-classified code or temporary Q-code. Even after a permanent code is assigned, reimbursement may still be based on wholesale acquisition cost until average sales price data catches up, usually two quarters later. The operator must track both the code transition and the reimbursement basis. A vendor should be able to explain who owns each transition and how other sites receive the change.

Accuracy also depends on the encounter. NDC validation should reflect the drug actually administered, including substitutions, while units and JW or JZ modifiers must reflect documented waste correctly. Administration sequencing requires judgment because chemotherapy takes priority over therapeutic infusion, which takes priority over hydration. Legitimately delivered services can still produce bundling denials when sequenced incorrectly. Biosimilar and branded products can also have different margins, making product-specific billing accuracy a financial concern.

Denial pattern recognition and payer-specific response as the third scorecard dimension

An appeal win resolves a claim, but it does not necessarily repair a workflow. The scorecard should test whether denials are grouped by payer, denial type, and the root cause in the work as well as by denial code. An expired authorization needs a different prevention response from missing medical necessity documentation. A site-of-care denial from UnitedHealthcare is different again from an NDC mismatch denial from a MAC. One combined denial rate hides those distinctions.

Evaluators should establish who turns a root cause into a prevention protocol, who distributes it, and whether platform leadership sees the pattern. Overturn rates should also be available by payer and denial type so that appeal approaches reflect what works with each payer. Local success should become shared operating knowledge. Otherwise, each site can become proficient at appealing the same preventable problem.

White-bagging denials need their own response because they affect the buy-and-bill model itself. When the payer’s designated specialty pharmacy supplies the drug directly, the practice does not purchase or bill for that drug and earns no drug margin. The model should distinguish these denials and support the relevant clinical documentation and contractual arguments for maintaining buy-and-bill where state law permits. Evaluation should establish who owns that response rather than assume general denial staff have the required expertise.

Underpayment identification and remittance reconciliation as the fourth scorecard dimension

A paid claim can still lose revenue. If a commercial payer reimburses below the contracted rate, claim acceptance says nothing about whether the amount is correct. Without line-level reconciliation, the shortfall can be posted as payment and the remaining balance written off. No denial appears in the denial report.

The scorecard should test whether reconciliation checks each drug against the rate expected from that payer. Evaluators should check whether the expected reimbursement schedule follows contract terms and the quarterly ASP updates that apply, and whether each claim payment is enough to meet that schedule. When a line pays below expectation, the model should identify and quantify the variance and route it to recovery. Manual discovery is not a defined recovery process. Experienced operators still need to interpret the discrepancy and take the appropriate payer-specific action.

Reports on recovery should break it down by payer and drug category. That detail allows repeated underpayments to be considered for contract renegotiation rather than handled indefinitely as unrelated collection tasks. A view across sites should also show whether the same biologic is being repriced below contract at other locations. Without it, a closed claim can conceal a recurring loss.

AR aging discipline and high-dollar claim prioritization as the fifth scorecard dimension

Infusion accounts receivable requires triage. A small number of high-dollar biologic claims can represent a disproportionate share of outstanding revenue. Sorting only by age can lead operators to spend time on smaller balances while claims that carry drug costs get closer to payer deadlines. The scorecard should test whether daily queues reflect dollar value and drug category as well as age. The order of work affects the money at risk.

Separate tracking for high-dollar biologics and routine infusions makes that risk visible. Evaluators should look for days-in-AR by drug category, denial patterns by drug category and provider, and alerts that leave enough time to act before timely filing deadlines. A good overall average can hide how long it takes to collect payment for expensive drugs. Days-in-AR should be assessed against high-performing infusion organizations rather than treated as strong because it resembles an industry average.

Escalation tests whether senior operators are there to take on the work. When a high-dollar claim reaches the sixty-to-ninety day bucket unresolved, the vendor should have a defined route to someone able to take payer-specific action. Leaving it in the same queue is not escalation. Platform reporting should also distinguish a local follow-up lapse from a payer that repeatedly slow-pays a particular biologic, allowing payer relations to address the broader pattern.

Cross-site analytics and platform-level visibility as the sixth scorecard dimension

Operators at different sites who lack shared analytics can repeat single-site problems across more locations. Each location may manage its own denials competently while missing a payer change affecting the entire platform. Centralizing every decision is not the answer. The model needs site-level ownership of execution, a normalized view of patterns, and a defined escalation route when a local finding points to a wider problem. That combination distinguishes a connected platform from a collection of billing teams.

The scorecard should test a platform-level view of denials, underpayments, and AR aging by payer and drug category. Separate site scorecards that require manual synthesis do not provide the same visibility. Leaders should be able to spot authorization lapses, aging biologic balances, or poor underpayment recovery at specific sites and send operator support where it is needed. Operators should do the work while leaders see outcomes. Performance over time should show whether a protocol change improved results across locations.

External payer data could strengthen this comparison. CMS’s April 2026 proposed rule, Federal Register 2026-07205, would extend Access API availability to drug authorization data, including denied and expired authorizations for at least one year after the last status change. The proposal also sets out public reporting of new authorization metrics beginning in 2028 for 2027 data, with payer-specific reporting scope. An analytics model able to use that information could compare internal performance with payer-reported approval rates, denial rates, and decision times to investigate whether a problem originates in local operations or broader payer behavior.

Operator accountability structures and how to test them during vendor evaluation

For reauthorization, the evaluator should ask who owns tracking at each site and what happens when a deadline is missed. For coding, the vendor should identify who manages a novel biologic’s transition from a Q-code to a permanent J-code and how that expertise reaches other locations. A central expert’s name is not enough. The answer must connect expertise to the site workflow.

Denial diligence should require an example of a payer behavior change identified across sites and the specific prevention protocol that followed. The vendor should explain how root causes were categorized, who received the analysis, and who changed the work. For a white-bagging requirement, the discussion should establish the clinical and contractual response process and its owner. Automation can support the decisions these questions test, but it does not substitute for someone accountable for making them.

Financial diligence should be equally concrete. The evaluator should establish whether remittances are reconciled at claim or line level, which expected rate schedule is used, and what underpayment recovery rates the vendor achieves by payer on accounts it manages. A platform-level denial report should show how findings guide where operators spend their time, and leaders should be able to compare sites without having to ask for a custom report. Software-led and operator-led models should have to meet the same requirements for evidence. The scorecard gives buyer judgment something firmer than a vendor’s promise.