While large language models (LLMs) have shown promise in diagnostic dialogue1, their capabilities for effective management reasoning—including disease progression, therapeutic response, and safe medication prescription—remain under-explored. We advance the previously demonstrated diagnostic capabilities of the Articulate Medical Intelligence Explorer (AMIE)1−3 through a new LLM-based agentic system optimized for multi-visit clinical management and dialogue. To ground its reasoning in authoritative clinical knowledge, AMIE leverages Gemini’s long-context capabilities4, combining in-context retrieval with structured reasoning to align its output with up-to-date clinical practice guidelines and drug formularies. In a randomized, blinded virtual Objective Structured Clinical Examination (OSCE) study, AMIE was compared to 21 primary care physicians (PCPs) across 100 multi-visit case scenarios designed to reflect UK NICE Guidance and BMJ Best Practice guidelines.
AMIE was non-inferior to PCPs in management reasoning as assessed by specialists and scored better in both preciseness of treatments and investigations, and in its alignment with and grounding in clinical guidelines. To benchmark medication reasoning, we developed RxQA, a multiple-choice question benchmark derived from two national drug formularies (US, UK) and validated by board-certified pharmacists. Though AMIE and PCPs both benefited from the ability to access external drug information, AMIE outperformed PCPs on higher difficulty questions. While further research would be needed before real-world translation, AMIE’s strong performance across evaluations marks a significant step towards conversational AI as a tool in disease management.
You have full access to this article via your institution. These authors jointly supervised this work: Alan Karthikesalingam, Mike Schaekermann These authors contributed equally: Valentin Liévin, Anil Palepu Valentin Liévin, Khaled Saab, David Stutz, Yong Cheng, S. Sara Mahdavi, Joëlle Barral, Ryutaro Tanno & Tao Tu
Anil Palepu, Wei-Hung Weng, Kavita Kulkarni, Dale R. Webster, Katherine Chou, Avinatan Hassidim, Yossi Matias, James Manyika, Vivek Natarajan, Adam Rodman, Alan Karthikesalingam & Mike Schaekermann Correspondence to Valentin Liévin, Anil Palepu, Alan Karthikesalingam or Mike Schaekermann. Supplementary discussion, methods and results (Sections 1-16). Contains related work, details on the system design for the Mx agent and Dialogue agent, details on the OSCE evaluation study (inter-rater reliability analysis, clinician metadata, scenario metadata, ablation analysis), and methods details and further results for the RxQA medication reasoning benchmark.
Detailed view of two sample scenarios with AMIE and PCP output and evaluation gradings. Full details for two sample scenarios used in the OSCE evaluation study, including scenario information, AMIE-patient-actor conversations, PCP-patient-actor conversations, specialist physician gradings and patient actor gradings for all three visits per scenario. Details for all 120 OSCE scenarios with AMIE output (PDF). Scenario details and AMIE output for all 120 scenarios used either in the OSCE evaluation study (100) or for validation purposes (20), in human-readable PDF format.
Details for all 120 OSCE scenarios with AMIE output (CSV). Scenario details and AMIE output for all 120 scenarios used either in the OSCE evaluation study (100) or for validation purposes (20), in machine-readable CSV format. Liévin, V., Palepu, A., Weng, WH. et al. Towards Conversational AI for Disease Management.
https://doi.org/10.1038/s41586-026-10764-5 Anyone you share the following link with will be able to read this content: Sorry, a shareable link is not currently available for this article. Provided by the Springer Nature SharedIt content-sharing initiative
