Electronic Health Records

Large language models (LLMs) show great potential for clinical decision-making, yet most applications remain narrow, task-specific chat tools rather than systems integrated into clinical workflows1,2. However, building physician copilots will require models that operate within the electronic health record (EHR), with governed access to patient data and the ability to initiate permitted EHR actions within defined safety constraints. Yet it remains unproven whether such a system can manage patient cases with physician-level performance. Here we show that MIRA (Medical Intelligence for Reasoning and Action), an autonomous artificial intelligence agent operating in a sandboxed EHR environment, can navigate a large clinical action space to obtain patient histories; order and interpret laboratory, imaging and microbiology tests; generate differential diagnoses; and formulate treatment plans such as prescribing medications, scheduling surgical procedures and planning admissions. In simulations on real patient cases spanning multiple diagnoses, MIRA outperformed physicians in diagnostic accuracy and made guideline-concordant, medication-safe and appropriate admission decisions. Compared with previous LLM applications that addressed isolated subtasks or provided free-text advice, these results suggest that an EHR-integrated artificial intelligence agent can turn clinical intent into structured, actionable EHR operations, possibly making it a more effective decision-support partner for physicians. Further work is needed to establish generalization, safety and governance through prospective, real-world studies.

BackgroundLarge language models (LLMs) have emerged as transformative technologies, revolutionising natural language understanding and generation across various domains, including medicine. In this study, we investigated the capabilities, limitations, and generalisability of Generative Pre-trained Transformer (GPT) models in analysing unstructured patient notes from large healthcare datasets to identify immune-related adverse events (irAEs) associated with the use of immune checkpoint inhibitor (ICI) therapy.MethodsWe evaluated the performance of GPT-3.5, GPT-4, and GPT-4o models on manually annotated datasets of patients receiving ICI therapy, sampled from two electronic health record (EHR) systems and seven clinical trials. A zero-shot prompt was designed to exhaustively identify irAEs at both the patient level (main analysis) and the note level (secondary analysis). The LLM-based system followed a multi-label classification approach to identify any combination of irAEs associated with individual patients or clinical notes. System evaluation was conducted for each available irAE as well as for broader categories of irAEs classified at the organ level.FindingsOur analysis included 442 patients across three institutions. The most common irAEs manually identified in the patient datasets included pneumonitis (N = 64), colitis (N = 56), rash (N = 32), and hepatitis (N = 28). The GPT models demonstrated generalisable abilities in identifying irAEs across EHRs and clinical trial reports. Overall, the models achieved relatively high sensitivity and specificity but only moderate positive predictive values, reflecting a potential bias towards overpredicting irAE outcomes. GPT-4o achieved the highest F1 and micro-averaged F1 scores for both patient-level and note-level evaluations. Highest performance was observed in the haematological (F1 range = 1.0-1.0), gastrointestinal (F1 range = 0.81-0.85), and musculoskeletal and rheumatologic (F1 range = 0.67-1.0) irAE categories. Error analysis uncovered substantial limitations of GPT models in handling textual causation, where adverse events should not only be accurately identified in clinical text but also causally linked to immune checkpoint inhibitors.InterpretationThis study demonstrated that GPT models can automate the detection of immune related adverse events in varied healthcare datasets, reducing the burden on physicians and other healthcare professionals by limiting the need for manual review. This capability will accelerate the generation of safety insights across large healthcare datasets and facilitate the characterisation of patient-level drivers of toxicities, thus enhancing safety monitoring and ultimately improving patient care.FundingNational Institutes of Health, Roche, National Health and Medical Research Council of Australia, Stevens-Johnson Syndrome Foundation, Angela Anderson Research Fund, Larry L Hillblom Foundation and UCSF Research Allocation Program.

BackgroundAmbient artificial intelligence scribes are increasingly used to reduce clinician burnout and cognitive load, although their impact on documentation time remains inconsistent across studies. Most existing real-world impact studies have been conducted in the United States and rely on electronic health record time stamps, which may not accurately reflect actual documentation time.ObjectiveThis study aimed to quantify the impact of ambient scribe technology on clinician efficiency and patient engagement in an outpatient setting using direct observational methods and patient surveys.MethodsWe conducted a prospective, within-clinician quality improvement study at a large academic medical center in Singapore from December 2024 to May 2025. Nine clinicians participated in matched observation sessions with and without an in-house ambient scribe tool. Five trained observers used standardized time-motion methodology to capture documentation time, proportion of eye contact, and consultation duration across 169 consultations. Patient surveys were administered following consultations using the ambient scribe. Linear mixed-effects models were used to account for clustering within clinicians.ResultsAmbient scribe use was associated with a 15.0% reduction in documentation time per consultation (mean 5.3, SD 2.7 minutes to 4.5, SD 2.4 minutes; P=.04); a 10.6% increase in proportion of eye contact time (mean 69.6%, SD 18.6% to 77.1%, SD 17.7%; P=.009); and no significant change in consultation duration (mean 11.5, SD 6.9 minutes vs 10.9, SD 5.6 minutes; P=.42) or total cycle time per patient (mean 13.7, SD 7.7 minutes vs 13.2, SD 6.3 minutes; P=.57), where total cycle time includes preclerking, consultation, and postconsultation documentation. Effects were consistent across new and follow-up patients. Patient acceptance was favorable: of 39 surveyed patients, 27 (69.2%) agreed that their physician focused on them more during the consultation, and none expressed discomfort with the technology.ConclusionsThis evaluation of ambient scribe use in an Asian health care setting demonstrates that experienced users achieved reduced documentation time and improved patient engagement without affecting consultation duration or total time per patient. The technology was also well accepted by patients. Taken together, these findings suggest that ambient scribes reallocate clinician effort toward patient interaction rather than enabling faster patient turnover, supporting their implementation across diverse health care settings.