FDA posts most serious recall of a HIT antibody assay; ONC rule drops AI model card requirements; female physicians' inpatients have lower mortality

• FDA Oct 30, 2026
  FDA lists HemosIL AcuStar HIT antibody assay recall as most serious type over bias that may mask heparin-induced thrombocytopenia
  Instrumentation Laboratories (Werfen) confirmed a negative bias of 20% to 35% in affected lots of the HemosIL AcuStar HIT-IgG(PF4-H) assay, which could cause results to be reported below the 1.00 U/mL cut-off, according to the Food and Drug Administration (FDA). The error may delay recognition of heparin-induced thrombocytopenia, and continued heparin exposure may raise the risk of thrombotic complications. Laboratories are told to identify an alternative HIT antibody test and, with treating clinicians, decide on reviewing previously reported results. As of Sept. 3, the company reported no serious injuries or deaths.
  Source: FDA
• Policy Oct 26, 2026
  ONC final rule drops 14 EHR certification criteria and removes model card requirements for AI decision support tools
  The Office of the National Coordinator for Health Information Technology (ONC) released the HTI-5 final rule on Oct. 26, removing 14 certification criteria and revising seven, according to ONC. The rule removes the model card requirements for source attribute support and for source attribute access and modification in the decision support interventions criterion, ending transparency, documentation and annual review requirements from HTI-1, according to Holland & Knight. Hospital and clinician commenters said removing the disclosures shifts the burden of vetting AI onto providers, and ONC said it may revisit AI oversight, according to the firm.
  Sources: Federal Register (public inspection), ONC, Holland & Knight, STAT
• Research Oct 30, 2026
  Care by female physicians tied to lower 30-day mortality in 958,150 Ontario medical inpatients, adjusted odds ratio 0.88
  A population-based cohort of 958,150 adults hospitalized in Ontario, Canada, from 2008 to 2022 with one of seven common acute medical conditions, including heart failure, chronic obstructive pulmonary disease, pneumonia and stroke, found adjusted 30-day mortality of 12.3% with female physicians versus 13.7% with male physicians (adjusted odds ratio 0.88, 95% CI 0.85 to 0.91), according to JAMA Network Open. Readmission was 18.6% versus 18.9% (adjusted odds ratio 0.98, 95% CI 0.97 to 1.00). Each 10% increase in hospital days managed by a female physician was associated with lower 30-day mortality (adjusted odds ratio 0.93).
  Source: JAMA Network Open

The weekly innovation roundup for hospital medicine, published every Sunday for the week before; every item links to its sources. Latest hospital medicine news · All specialties