At a Becker’s Healthcare webinar on May 7, Dr. David Kirk, CMO of Regard, moderated a conversation on where the documentation problem actually starts and what it takes to solve it there. He was joined by Dr. Carolyn Bauer, Medical Director of Clinical Documentation at Montefiore, and Fran Jurcak, an independent consultant with more than twenty years in CDI and revenue cycle.
Key takeaways
- Every retrospective query is a documentation failure that already happened. The infrastructure built to recover missed revenue — CDI queries, denial management, chart reviews — exists because the clinical picture wasn’t complete at the point of care.
- The gap between how clinicians document and what payers require is a system design failure. There are more than twenty ways to document heart failure, and most physicians were never taught which one the record needs.
- Query volume is the wrong productivity metric for CDI. A team incentivized to send more queries will find more to send. What matters is whether case mix, severity scores, and diagnoses are accurate and defensible from the start of the encounter.
- Payers are already moving. Diagnoses added through retrospective queries are increasingly being denied as not native to care.
- The case to the hospitalist and the case to the CFO are the same argument. CC and MCC capture, severity of illness, query agreement rates: each is simultaneously a clinical and a financial measure. Health systems that treat them as separate conversations are solving for the wrong thing.
Where the breakdown actually starts
Health systems have spent years building infrastructure around the back end of the revenue cycle: CDI queries, denial management, retrospective recovery. That investment made sense given the pressures of the time. But it’s organized around fixing a problem that could have been prevented.Physicians aren’t trained in what documentation requires.
Medical education focuses on care delivery, not on the specific language that translates clinical thinking into billable codes. There are more than twenty ways to document heart failure, and what’s most accurate to the patient in a clinician’s mind may not map to any of them. As Jurcak noted: “Words are just words. How accurate those words are is dependent upon what is the clinical picture of the patient.”The problem compounds across a hospitalization. Copy-paste notes and templated documentation mean that what was ruled out on day one can still appear on day ten. A condition present at admission can drop out of the record entirely by discharge.
Dr. Bauer described it as a failure that runs the full length of a stay: “Every document that the doctor writes is an opportunity to correct and get the scope of what’s happening to the patient into the medical record.”
Clinicians also tend to wait for a definitive answer before documenting a working one. A physician treating a patient for what is likely sepsis won’t commit that to the record until certain, even as care is actively being delivered for that condition. “We need to capture it from day one in the treatment,” Jurcak said, “not at the end of the treatment.” By the time a discharge summary is written, the window to defend what happened is already closing.
The true cost of retrospective infrastructure
The infrastructure built to recover missed documentation carries costs that rarely appear in a business case. Dr. Bauer itemized them: physician time spent answering queries, the FTEs required to fight denials even when the fight is eventually won, and the delay between care delivered and claim submitted. “You have all the touches post-discharge, and then that takes time away from when the claim can be submitted for that case.” Every touch adds cost. And most of them exist because the documentation wasn’t right the first time.
Payers are paying close attention to when documentation happens, and the pressure is accelerating. Diagnoses captured at the point of care, backed by clinical evidence already in the record, hold up. Diagnoses added through retrospective queries increasingly don’t. As Dr. Kirk noted, “It’s gonna come down to payers denying anything that was queried because it’s not native to care.” The window to defend a diagnosis is during the encounter. Retrospective tools can only ever recover a fraction of what was missed, and that fraction is shrinking.
What ambient documentation solves — and what it doesn’t
That narrowing window has pushed many health systems toward ambient documentation as a solution. The panel was frank about its limits. Capturing more of what’s said during an encounter produces a more complete record of the conversation, but it doesn’t include reasoning across the chart. The technology doesn’t interpret a lab value in the context of a patient’s medication history or flag a diagnosis that was never spoken aloud. “Ambient is really just capturing what was said,” Jurcak said. “Unless the physician says the lab result, ambient doesn’t have the capability to consider it.”
Dr. Kirk drew on his own experience as an ICU physician. The thing that kept him up at night wasn’t missing something said in the room. It was the fear of something buried in the chart, in a prior record, a lab trend, an outside note, that had he known it, would have changed the care. Volume of documentation and completeness of the clinical picture are different problems. Tools that solve the first don’t address the second.
What changes for CDI
Removing the most common, avoidable queries from CDI’s plate doesn’t reduce the function. It elevates it. Low-complexity, high-volume queries take up precious bandwidth that experienced specialists should direct toward complex clinical validation, quality alignment, and the denial patterns that require genuine expertise.
CDI programs also face persistent staffing problems that rarely factor into these conversations. “Almost all of them will always tell you, ‘I constantly run a headcount under what my FTE allotment is,'” Jurcak said. Freeing experienced staff from repetitive query work doesn’t eliminate the need for CDI but directs that expertise where it matters.
Both speakers pushed back on query volume as a productivity metric. A team measured on output will find queries to send, regardless of whether they’re the right ones. The outcome is what matters: whether case mix reflects care complexity, whether severity scores are accurate, whether diagnoses are present and defensible from the start of the encounter. “If they are, and I don’t have to query,” Jurcak said, “then somewhere else along the way we did it right.”
One case, two audiences
Getting CDI, clinical leadership, and finance aligned on the same investment doesn’t require telling three different stories. It requires telling one story clearly. Dr. Bauer’s observation was that the gap between clinical and financial audiences has narrowed considerably: “There was a time when you couldn’t talk about reimbursement to doctors. But that has shifted. Doctors understand the financial pressures.”
CC and MCC capture, severity of illness, risk of mortality, query agreement rates: each is simultaneously a clinical and a financial measure. A physician who answers fewer queries and finishes a shift confident the clinical picture was complete is also a physician whose documentation is generating cleaner claims. Getting it right at the point of care changes what every downstream team has to do.
The stakes of getting it wrong extend beyond the revenue cycle. One of the planned panelists, Nio Quiero, was absent due to a family medical emergency. At the time of the webinar, she was navigating multiple health systems without a shared medical record, trying to assemble a clinical picture across institutions that couldn’t communicate with each other. Dr. Kirk raised her situation deliberately. The documentation problem the panel spent an hour discussing wasn’t abstract to her. Getting it right at the point of care matters to the families on the other side of it, not just the teams managing the revenue cycle.
