Outcomes Articles

Cardiac AI Can Reach the Physician. Can the Hospital Run the Rest?

October 26, 2026

About This Report

Featuring insights from Roger Owens, Deanna Harshbarger, and Ted Ferrin hosted by our CEO, Saul Marquez.

The cardiac-AI result arrives where the physician already reads imaging studies. That implementation choice comes up in Saul Marquez's Outcomes Rocket conversation with Roger Owens, Chief Commercial Officer at Ultromics. Yet getting the result onto the right screen doesn't answer a different set of questions: who records the service, prepares the claim and checks what was paid?

Those responsibilities land on people with existing workloads, not on a reimbursement announcement. A hospital can understand the clinical application while its billing staff still need documentation, payer information or a clear escalation path. For a commercialization team, calling the service ready before those handoffs are defined can leave the buyer with unfinished operating work.

Follow one intended service from the clinical result through payment reconciliation before expanding the launch. Three separate interviews help examine the gaps: Ultromics on clinical delivery, Cala Health on distinct access milestones and Zelis on making payment information usable. They describe different businesses, but each exposes work that a product demonstration can leave out.

Make the Clinical Result Part of an Executable Service

The Ultromics conversation describes implementation across echocardiography physicians, heart-failure specialists and hospital IT teams. EchoGo processes the heart ultrasound studies, then returns results to the hospital's picture archiving and communication system, or PACS, for the reading physician. The clinical result has a destination inside an existing workflow rather than a separate screen the team must remember to visit.

That last detail matters. Physicians already have demanding schedules, established systems and responsibilities that cannot simply be rearranged around another technology. Returning information to their existing environment reduces the need to introduce a separate destination for reviewing results. EchoGo provides information that may alert a physician to signs associated with heart failure, while the physician remains responsible for making the clinical diagnosis.

The same conversation moves from that clinical destination to a less visible implementation task: making billing and utilization workable. It describes communication with payers, education for the people involved and billing processes appropriate to the service. These aren't additional capabilities of the AI model. They're responsibilities surrounding its use.

Consider what this means during a launch review. The clinical implementation lead needs to know where the result appears and who acts on it. IT must understand how information enters and returns through the hospital's systems. The revenue-cycle team needs to understand how use of the service will be documented and what information supports the billing process.

These responsibilities are connected, but they don't automatically resolve one another. A functioning integration demonstrates that information can reach its intended clinical destination. It doesn't establish that utilization is consistently captured, that claims can be prepared correctly or that reimbursement will follow.

For commercialization leaders, a useful review question is whether each responsible team can explain its work and identify what the next team needs. This is an editorial launch-review exercise drawn from the implementation challenges Roger describes, rather than a prescribed Ultromics process.

A hospital can map that work only when the commercialization team is precise about what its access milestones have actually settled.

Separate the Access Milestones Before Building Around Them

Deanna Harshbarger, CEO of Cala Health, describes a different medical-device challenge that helps explain why reimbursement readiness requires more precise questions. Cala developed a wrist-worn therapy for patients living with tremor, creating a treatment option that also required the company to establish a pathway for patients to access it.

The Outcomes Rocket conversation with Deanna recounts Cala's earlier work through the De Novo regulatory pathway, followed by developing new codes, obtaining a payment determination and securing a local Medicare coverage determination. Each represented a different hurdle. Progress on one didn't eliminate the work required for the next.

The episode describes a team repeatedly encountering a negative response and asking what lay behind it. The productive next move was to examine the reason and the explanation or evidence still needed. That is more useful to a launch team than treating every setback as a request to repeat the same pitch more forcefully.

Keep those decisions separate in the launch plan. Permission to market a device, a code used to report a service, the conditions governing coverage and the payment applicable to a claim answer different questions. Have the responsible specialists identify the decision and conditions that actually apply. None of these labels is a blanket promise that the hospital will receive payment.

Cala's experience is a historical account involving a different technology, not a reimbursement template for cardiac AI. The useful application is to identify the precise access decision still unresolved, the evidence or explanation required and the person responsible for pursuing it.

That information belongs in the hospital's implementation plan before staff build routine processes around an assumed payment pathway. Even when those decisions are understood, the people handling the actual transaction still need information they can use.

Make Payment Understandable Where the Work Happens

Ted Ferrin, SVP of Payments Innovation at Zelis, brings the discussion to a practical question: What happens when someone has to interpret and reconcile the payment? In his Outcomes Rocket interview, Ted describes a financial system where payer contracts, claims processing, remittances and analytics often operate through separate systems with different information requirements.

For providers, that fragmentation creates work. Staff may rely on manual postings, spreadsheets and even sticky notes to connect information that should explain the same transaction. These are examples from Ted's broader payments experience, not documented problems at an Ultromics hospital.

The distinction between seeing financial data and being able to use it becomes important here. A report might display payment information, but staff still need to understand how it relates to the contracted reimbursement expectation, the submitted claim and the payer's decision. The original billed charge, expected contracted amount and actual payment aren't interchangeable.

Ted's conversation emphasizes clean, normalized data, consistent financial definitions and information delivered inside the workflow where staff need it. Transparency becomes more useful when it helps someone resolve a discrepancy without repeatedly switching systems or manually reconstructing what happened.

Reducing that rework is a reasonable test of whether better financial visibility is helping. It's an objective discussed in the interview, not a measured result for Ultromics or evidence that the companies share an integration.

For a commercialization team, the question is where a hospital's revenue-cycle staff would still need missing information, interpret conflicting terms or contact another party before closing the transaction. Identifying those points gives the team something concrete to resolve.

A dashboard can show a discrepancy without giving anyone enough information to resolve it. The launch review should end with a person who owns that question and a defined source for the missing information, not another report to reconcile manually.

Trace One Service Before Expanding the Promise

Before expanding a cardiac-AI launch, bring the clinical implementation lead, IT owner, market-access lead and revenue-cycle representative together. Choose one intended service and build a handoff map showing how information moves from the clinical result to the final payment reconciliation.

Work through four checkpoints with the people who perform the work:

Result delivery: Where the result appears, who reviews it and what clinical or IT question remains.
Utilization: How service use is documented and how that information reaches billing.
Claim preparation: Which access conditions are established and who resolves payer or documentation questions.
Reconciliation: How staff compare the claim, contracted expectation, remittance and actual payment, and who investigates differences.

Record the responsible owner, required information and unresolved question at each handoff. Resolve or explicitly bound the outstanding responsibilities before deciding whether to expand the launch. This exercise identifies assumptions to test, not guaranteed reimbursement or profitability.

The result in the physician's imaging system has a clear destination. Give the remaining work the same treatment: identify who receives it, what they need and how they know their part is complete.

Before calling the service ready, ask the people at every handoff what they still need.

About the Conversations

These are three separate Outcomes Rocket conversations, not a shared case study. During the time of recording, the guest names and titles were as follows: Roger Owens, Chief Commercial Officer at Ultromics. Deanna Harshbarger, CEO of Cala Health. Ted Ferrin, SVP of Payments Innovation at Zelis.