Performance of four large language models in addressing family-centered questions on morbidity and outcomes in patients supported by extracorporeal membrane oxygenation - Takeaways - MDSpire

Performance of four large language models in addressing family-centered questions on morbidity and outcomes in patients supported by extracorporeal membrane oxygenation

  • By

  • Wouter Vankrunkelsven

  • Leen Vercaemst

  • Dieter F Dauwe

  • June 22, 2026

  • 0 min

Share

  • 1

    Extracorporeal Membrane Oxygenation (ECMO) provides critical support for patients, with increased usage during the COVID-19 pandemic.

  • 2

    The study evaluated four language models' responses to family-centered questions about ECMO morbidity and outcomes.

  • 3

    Gemini achieved the highest scores in content and communicative quality, while DeepSeek ranked the lowest among the evaluated models.

  • 4

    All LLMs provided disclaimers advising consultation with healthcare providers for personalized information.

  • 5

    Responses from LLMs were generally consistent with evidence, but many were rated unsatisfactory by ECMO experts.

Original Source(s)

Performance of four large language models in addressing family-centered questions on morbidity and outcomes in patients supported by extracorporeal membrane oxygenation

  • by Wouter Vankrunkelsven, Leen Vercaemst, Dieter F Dauwe

  • June 22, 2026

Related Content

From immunological mechanisms to targeted therapies: a bibliometric analysis in the domain of research concerning neutrophil extracellular traps and pulmonary diseases (2006–2025)

Editorial: Advancing sports cardiology: new frontiers in athlete screening and recovery

Characterization of post-COVID syndrome by self-perceived symptom severity stratified by infection wave: beyond COVID, a prospective, multicenter cohort study in Germany