Cardiology Doesn't Need Another Gatekeeper

When the American Board of Medical Specialties rejected the American Board of Cardiovascular Medicine's application in February 2025, it did more than deny paperwork. It reaffirmed a certification model designed for a version of medicine that no longer exists, one in which a physician's competence could be adequately captured by a periodic, high stakes, closed book exam administered by a body governing more than a dozen unrelated internal medicine subspecialties. That model was outdated in 2015. In an era when large language models (LLMs) and frontier AI models can already draft a differential diagnosis, flag a subtle ECG abnormality, or summarize a decade of a patient's records in seconds, it is not just outdated, but it is measuring the wrong thing entirely.

The ACC, together with the AHA, HRS, HFSA, and SCAI, was right to pursue the ABCVM. It should keep pursuing it, and it should not let one rejection or the prospect of a long AMA mediated process talk it out of finishing the job.

Cardiology stopped being a subspecialty of internal medicine decades ago. Interventional cardiology, electrophysiology, advanced heart failure, structural heart disease, cardio-obstetrics, cardio-oncology, and these are not footnotes to general internal medicine competence; they are distinct bodies of procedural skill, imaging literacy, and device management knowledge that ABIM's single track certification structure was never built to assess with precision. A board governed by and for cardiologists can build subspecialty specific, continuously updated assessment pathways in a way that a board also responsible for endocrinology, rheumatology, and a dozen other fields structurally cannot.

The AI era makes the case sharper, not weaker. The traditional argument for periodic recertification exams was that medical knowledge changes fast enough that physicians need to be tested on it. But knowledge retrieval is no longer the scarce skill. What's scarce and what actually protects patients is a cardiologist's judgment about when to trust an AI generated read of a coronary angiogram/echo/MPI, how to integrate a machine learning risk score (high CHADVAS and HASBLED) into a shared decision with a patient, and when the algorithm is wrong. None of that is measurable with a static, pass/fail, closed book test taken once every ten years. It requires exactly the kind of continuous, technology enabled, practice embedded assessment that ABCVM's founders proposed: tools that identify a clinician's actual knowledge gaps in real time, credit the quality improvement and lifelong learning work cardiologists already do, and adapt as the evidence base and the tools themselves evolve. A specialty owned board, unencumbered by ABIM's broader institutional inertia, is far better positioned to build and iterate on that kind of infrastructure quickly.

Physician trust in the current system has already collapsed, and that has consequences. Years of documented frustration with ABIM's Maintenance of Certification, including the cost, the opacity, and the sense that the process rewards test taking over actual clinical performance, have not been fixed by incremental changes like the Longitudinal Knowledge Assessment or the Collaborative Maintenance Pathway. When SCAI's own leadership describes the status quo as failing to reflect "how cardiovascular specialists train, practice, and maintain their skills today," that isn't a fringe complaint; it's the professional societies representing the overwhelming majority of practicing cardiologists saying the emperor has no clothes. A profession that doesn't trust its certifying body will treat certification as a compliance exercise rather than a genuine marker of competence, precisely the outcome that undermines public confidence in board certification generally.

ACC has the governance capacity to do this responsibly. This isn't a fringe breakaway effort. It carries the backing of every major cardiovascular professional society in the country, a detailed application that its authors say met ABMS's own published standards, and, as of the 2025 AMA House of Delegates meeting, a resolution directing the AMA itself to study and define principles for what board certifying bodies should look like going forward, with a report due at the 2026 meeting. That is not the posture of an effort running out of momentum. It's the posture of a specialty building the broader institutional case for self governance the right way, even after a setback.

None of this means the concerns behind ABMS's rejection should be waved away. A fragmented certification landscape carries real risks: diluted standards if new boards proliferate without rigor, confusion for hospitals and payers about which credentials matter, and the possibility that an assessment model built to be more "supportive" than ABIM's becomes so lenient it stops meaningfully distinguishing competence at all. Skeptics are also right that an unproven, technology driven assessment framework needs to demonstrate, not just promise, that it protects patients as well as a traditional exam does before it earns the same public trust ABIM has spent decades building, however imperfectly.

Those are legitimate design questions. They are not, however, reasons to preserve a status quo that the specialty itself has said doesn't work. They're reasons for ABCVM to keep building its case with data, and for ACC to keep pushing rather than folding the effort back into ABIM's incremental reforms. Cardiology has evolved into a distinct medical specialty with its own knowledge base, its own procedures, and, increasingly, its own relationship to AI augmented practice. Its certification process should evolve with it, under a board that cardiologists actually built and trust.

The ABCVM fight isn't over. It shouldn't be.