Published in JAMA Network Open on 17 November 2026, the research evaluated an AI system developed with resident doctors at Imperial College Healthcare NHS Trust. In a blinded assessment of 1,294 summaries reviewed by 12 doctors, AI-generated summaries were rated higher for clinical acceptability, accuracy, completeness, clarity, and tone compared to those produced through usual clinical practice.
The findings suggest that AI-assisted discharge documentation could be a significant use case for generative AI in healthcare. Key findings include: 92% of AI-generated summaries met the threshold for clinical acceptability, versus 70% of clinician-written summaries. Reviewers identified fewer potentially harmful omissions and inaccuracies in the AI-generated versions.
iCARE, a data and analytics initiative between Imperial College London and Imperial College Healthcare NHS Trust, aims to drive innovation in health outcomes using secure health data. Discharge summaries are crucial for communicating patient hospital admission details to GPs and other healthcare professionals, aiding patient understanding and follow-up care. Producing these summaries is often time-pressured and complex due to the volume of patient record information.
Dr Benjamin Atkins, a resident doctor involved in the study, highlighted the tension between clinical and administrative tasks, stating, “That simultaneous demand is exactly why automation of administrative tasks would be so powerful on the wards.” The study used the Imperial Secure Data Environment to generate draft discharge summaries from unstructured electronic health record data.
Dr Fidan Damirova, also a resident doctor on the study, commented, “The real promise of this work lies in using AI to transform fragmented clinical information into clear, clinically meaningful discharge communication, with the potential to support safer transitions of care while enabling clinicians to focus more time on what matters most: their patients.”
The study's strength lay in its clinical co-design approach, involving resident doctors throughout development and evaluation. Kathleen Goldsmith, lead researcher and Senior Data scientist at iCARE, emphasised, “To build transformative clinical AI we need to develop applications alongside the clinicians who will use them, and rigorously evaluate their safety, reliability and clinical value.” This work provides a model for assessing emerging technologies in healthcare.




