Why Your Coverage Analysis Undercounts Imaging, And Audits Notice

THE BILLING GRID
A weekly column on research billing and coverage analysis, by Krishma Shah

The call that nobody wants comes about eighteen months into a study: the compliance team is reviewing the research billing log and the CT scans are not adding up. The coverage analysis (the document, sometimes called an MCA or medical coverage analysis, that assigns each protocol event a billing designation) says imaging is a research expense at visit two. The site billed nine scans. The MCA defended one.

That gap is not a documentation technicality. It is the fact pattern that produces audit findings, repayment demands, and, in the worst cases, the kind of institutional settlement that ends careers in research administration. The billing risk here is specific and preventable, and it starts with a structural problem in how imaging gets entered into the MCA in the first place.

How the Decomposition Error Happens

A typical oncology protocol writes imaging requirements the way a clinician thinks about them: CT chest, abdomen, and pelvis every eight weeks for nine cycles. That is one instruction with a defined frequency and a defined duration. The site reads it as a package.

But many MCA templates are built around procedure codes, not protocol schedules. A CT of the chest is one CPT line. A CT of the abdomen is another. A CT of the pelvis is a third. When a billing analyst or a template-driven tool enters that imaging requirement, it decomposes the scan set into three separate rows, one per anatomical region. Each row gets a designation: research-paid (RES), standard of care (SOC), or non-billable (NB). The designation is applied once, to a single row entry, without a visit count attached.

The parent schedule, every eight weeks times nine, disappears. It never makes it into the defensible record. The MCA now says CT chest is research-paid at the imaging visit. It does not say which imaging visits, how many, or that there are nine of them across the study calendar.

A human reviewer who knows the protocol catches this. A template does not. And at sites running eight or ten protocols simultaneously, with coordinators and billing staff under real time pressure, the template answer is often the final answer.

Why Auditors Find It

The defensible unit in a coverage analysis is the visit-grid cell: the intersection of a specific procedure and a specific visit. The same CBC, which maps to NCD 190.15 under Medicare billing rules, can be a standard-of-care item at screening and a research-required item at a mid-cycle safety visit. The designation is visit-specific, not procedure-global. This principle comes directly from the CMS guidance on routine costs in qualifying clinical trials under NCD 310.1, which governs which costs in an IND-supported, therapeutically-intended trial are billable to Medicare versus which must stay with the sponsor.

When an auditor pulls the billing record and sees nine CT scan charges across the study period, they look at the MCA to confirm each charge was properly designated at each visit. If the MCA has one row for CT chest with a single RES designation and no visit-count, the auditor cannot reconcile the nine charges to nine approved billing decisions. The documentation does not exist. The site has billed nine times on the authority of one entry.

This is structurally similar to the billing compliance failures that produced institutional repayment obligations at academic medical centers over the past two decades, including the fact pattern behind Rush University Medical Center’s 2005 settlement, where research items were billed to Medicare without the designation work to support the charges. The amounts at issue in those cases were not trivial, and the reputational damage to research programs outlasted the financial settlement.

What a Correct Imaging Entry Looks Like

Take the same protocol instruction: CT chest, abdomen, and pelvis every eight weeks, nine cycles. A defensible MCA entry preserves the schedule in the grid. That means a visit column for each imaging timepoint, visit two, visit four, visit six, and so on through visit eighteen in an every-eight-weeks cadence, with the CT procedure row carrying an explicit designation at each column. If all nine are research-paid, all nine cells say RES. If a subset overlaps with a diagnostic workup that would have occurred regardless of the trial, the SOC designation at those specific visits requires its own clinical rationale and NCD citation.

The anatomical decomposition can stay for coding purposes. A CT of the chest, a CT of the abdomen, and a CT of the pelvis may each carry a separate CPT code in the charge master. But the visit-count logic must survive the decomposition. Nine visits times three scan regions equals twenty-seven cell entries in the grid. Auditors can follow twenty-seven cells. They cannot follow one row with an unexplained billing history of nine charges behind it.

The Protocol Read That Prevents the Error

The fix is not a better template. Templates cannot read protocols; they receive inputs. The fix is a protocol read that is indication-aware and schedule-aware before a single row is entered.

Indication-aware means understanding why the imaging is in the protocol. In an oncology study with RECIST-based response assessment, every eight-week CT is a required efficacy measure, making it research-paid by definition. In a cardiovascular study where an echocardiogram appears at the same frequency as a standard monitoring interval for the enrolled population, the SOC argument is stronger and requires the NCD 310.1 desk criteria: therapeutic intent confirmed, Medicare-benefit-category item confirmed, deemed status under IND confirmed.

Schedule-aware means reading the visit schedule table and the eligibility criteria together. Protocols frequently state the imaging cadence in one section and then adjust it in the amendment history or the deviation management section. An MCA built from the original protocol and never reconciled against amendment two has the wrong visit count. Sites I work with have flagged this as a standing checkpoint: every MCA gets a final read against the most current protocol version and the most current amendment log before it goes to the budget team.

The budget reconciliation step matters too. Once the sponsor budget arrives, the INVOICEABLE section of the MCA gets priced at the greater of the sponsor rate or the charge master rate for research-paid items. If the visit count in the MCA is wrong, the invoiceable total is wrong, and the site is either leaving money on the table or overbilling the sponsor. Both outcomes are problems. The imaging cadence is the line item most likely to drive the discrepancy because it is the highest-frequency, highest-unit-cost procedure in most oncology and imaging-heavy protocols.

The MCA sits on the site activation critical path, alongside the contract and the IRB submission. Getting it right the first time is not a compliance formality. It is the operational foundation for every billing decision the site makes for the life of the study.