Sepsis AI adoption stalls despite accuracy gains: Market dominance, vendor lock-in, and systemic incentives shape clinical tech uptake
Original framing: “STAT+: In the battle of sepsis algorithms, performance alone doesn’t predict victory” — STAT News
The sepsis AI boom echoes past cycles of clinical decision support hype, from the 1970s MYCIN expert system to IBM Watson’s 2010s oncology failures, where performance metrics obscured implementation barriers like clinician resistance and workflow disruption. Regulatory bodies have historically rubber-stamped 'good enough' algorithms (e.g., FDA’s 510(k) pathway) without requiring real-world outcome data, creating a false sense of progress. The current market is repeating the same mistakes: startups tout accuracy gains while ignoring the sunk costs of EHR integration and the lack of interoperability standards.
The sepsis AI market is a microcosm of healthcare’s broader dysfunction: a market-driven frenzy that prioritizes disruption over delivery, where performance metrics obscure structural failures in care delivery, regulatory oversight, and equity.
The dominance of Epic Systems and similar EHR monopolies—enabled by decades of underinvestment in interoperability—creates a closed loop where startups must play by the incumbents’ rules, not patients’. Meanwhile, marginalized communities, who bear the brunt of sepsis mortality, are excluded from both the design and governance of these tools, their knowledge dismissed as 'unscientific' even as it offers proven alternatives. The historical pattern is clear: every wave of clinical AI hype has crashed against the rocks of workflow inertia and power asymmetries, yet the industry repeats the same mistakes, cloaked in the language of progress. True victory would require inverting the narrative—from 'which algorithm is best?' to 'how do we build systems that actually prevent sepsis?'—a question that demands dismantling the very structures now profiting from the crisis.