cs.LGMar 4, 2026

A Multi-Agent Framework for Interpreting Multivariate Physiological Time Series

Authors: Davide Gabrielli, Paola Velardi, Stefano Faralli, Bardh Prenkaj

Organizations: Sapienza University of Rome Rome, Italy · ISTC-CNR & Sapienza University of Rome Rome, Italy · Technical University of Munich Munich, Germany

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Continuous physiological monitoring is central to emergency care, yet deploying trustworthy AI is challenging. While LLMs can translate complex physiological signals into clinical narratives, it is unclear how agentic systems perform relative to zero-shot inference. To address these questions, we present Vivaldi, a role-structured multi-agent system that explains multivariate physiological time series. Due to regulatory constraints that preclude live deployment, we instantiate Vivaldi in a controlled, clinical pilot to a small, highly qualified cohort of emergency medicine experts, whose evaluations reveal a context-dependent picture that contrasts with prevailing assumptions that agentic reasoning uniformly improves performance. Our experiments show that agentic pipelines substantially benefit non-thinking and medically fine-tuned models, improving expert-rated explanation justification and relevance by +6.9 and +9.7 points, respectively. Contrarily, for thinking models, agentic orchestration often degrades explanation quality, including a 14-point drop in relevance, while improving diagnostic precision (ESI F1 +3.6). We also find that explicit tool-based computation is decisive for codifiable clinical metrics, whereas subjective targets, such as pain scores and length of stay, show limited or inconsistent changes. Expert evaluation further indicates that gains in clinical utility depend on visualization conventions, with medically specialized models achieving the most favorable trade-offs between utility and clarity. Together, these findings show that the value of agentic AI lies in the selective externalization of computation and structure rather than in maximal reasoning complexity, and highlight concrete design trade-offs and learned lessons, broadly applicable to explainable AI in safety-critical healthcare settings.

Figures & tables

Figure 1. System architecture as a five-scene Emergency Department narrative. The area on the right represents the Shared Memory Buffer (SMB) that Vivaldi manages, orchestrating communication with it and the agents, or between agents, as an intermediary: e.g., think of this as API calls where read (dashed lines) and write (dotted lines) operations happen in the SMB. The circles at the edge of the shaded area on the right illustrate the agents’ entry point for communicating with Vivaldi. We remove the details of read/write interactions between the SMB and Vivaldi for illustration purposes, and defer them to the single scenes. Same-colored lines represent a single logical flow whose completion might depend on multiple interactions between Vivaldi and the agents therein (e.g., 5.b, 8, 9, and 10).
MetricThinkingNon-Thinking
 Zero-shotAgenticZero-shotAgentic
Factuality88.6 ±\pm 21.483.3 ±\pm 28.987.0 ±\pm 22.391.0 ±\pm 19.4
Justification93.2 ±\pm 17.683.3 ±\pm 28.988.9 ±\pm 21.296.2 ±\pm 13.5
Relevance95.5 ±\pm 14.781.0 ±\pm 29.587.0 ±\pm 22.397.4 ±\pm 11.2
Trust90.9 ±\pm 25.183.3 ±\pm 28.990.7 ±\pm 19.893.6 ±\pm 16.9
Table 1. Aggregated expert explanation quality (0–100 scale; Yes=100, Partial=50, No=0). Green shading marks the best strategy per category; bold denotes the global best. Thinking models: Gemini 3 Pro, Claude 4.5 Opus. Non-Thinking: GPT 5.2, Llama 4 Maverick, MedGemma 27B.
 FactualityJustificationRelevanceTrust
 Zero-shotAgenticΔ\mathbf{\Delta}Zero-shotAgenticΔ\mathbf{\Delta}Zero-shotAgenticΔ\mathbf{\Delta}Zero-shotAgenticΔ\mathbf{\Delta}
Gemini 3 Pro88.981.8-7.1%94.481.8-12.6%94.477.3-17.2%88.981.8-7.1%
Claude 4.5 Opus88.585.0-3.5%92.385.0-7.3%96.285.0-11.2%92.385.0-7.3%
GPT 5.285.784.6-1.1%92.996.2+3.3%92.9100.0+7.1%85.788.5+2.7%
Llama 4 Maverick90.0100.0+10.0%90.096.2+6.2%85.096.2+11.2%90.092.3+2.3%
MedGemma85.088.5+3.5%85.096.2+11.2%85.096.2+11.2%95.0100.0+5.0%
Table 2. Explanation quality comparison for Thinking models (top) versus Non-Thinking models (bottom) on a normalized 0–100 scale. Δ\mathbf{\Delta} represents the difference in percentage points between Agentic and Zero-shot performance.
Clinical MetricThinkingNon-Thinking
 Zero-shotAgenticZero-shotAgentic
ESI (F1 ↑\uparrow )61.064.640.765.4
Pain Score (MAE ↓\downarrow )1.91.82.42.8
LOS (MAE ↓\downarrow )1.51.91.62.0
qSOFA (F1 ↑\uparrow )63.6100.048.1100.0
SI (MAE ↓\downarrow )0.10.00.10.0
Table 3. Average model performance on triage-related clinical metric estimation tasks under zero-shot and agentic settings. Green-shaded cells indicate improved performance within the model type; bold indicates the best across the board per metric. (ESI=Emergency Severity Index; LOS=Length of Stay; MAP=Mean Arterial Pressure; SI=Shock Index).
Figure 2. ESI Level Confusion Matrices. Zero-shot models (left) show a critical failure to detect Level 1 (highest acuity) emergencies, misclassifying all such cases.
Figure 3. Performance shifts in Chart Comprehensibility (left) and Clinical Utility (right) across models. Each arrow represents the transition from zero-shot (open circle) to the agentic workflow (solid circle). Green arrows indicate performance gains; red, degradations.
Figure 4. Time efficiency analysis of Zero-Shot versus Agentic pipelines. Stacked bars illustrate the temporal distribution of agents, each annotated with its share of total execution time. Gray bars represent baseline Zero-Shot latency.
Figure 5. Prompt and Completion token usage by model.
More in this topic

Explainable Artificial Intelligence

Explainable artificial intelligence develops ways to make AI decisions understandable, testable, and useful to people.

See all 326 papers

Explore similar work

CardsList
Jul 28, 2026

Agentic AI in medicine: architectures, applications, evaluation, and challenges for clinical translation

Large language models and multimodal foundation models are enabling medical artificial intelligence (AI) systems to move beyond isolated prediction and undertake multistep clinical tasks that require planning, tool use, memory, iterative correction, and coordination among specialized agents. However, the scope of agentic AI in medicine remains unsettled, and current evaluation practices are not yet aligned with the requirements of clinical use. We conducted a scoping review with systematic evidence mapping across five electronic sources, screened 1,649 exportable records, and provisionally included 557 unique studies that met predefined criteria for goal-directed task execution, tool use, interaction with external resources, feedback-based refinement, or multi-agent collaboration. The included studies describe single agents that use external tools, workflows supported by retrieval and external knowledge, multimodal agents, and multi-agent systems applied to medical question answering, image interpretation, electronic health record analysis, drug safety, and clinical trial prediction. The evidence base remains dominated by public benchmarks, simulated settings, retrospective datasets, and small-scale expert evaluation. Process reliability, evidence traceability, uncertainty, safety, workflow impact, and external validity are evaluated less consistently. Clinical translation will depend on clearer definitions, reproducible evaluation, auditable oversight, interoperable system design, and prospective validation in real-world clinical workflows.

Zheng Tong, Yang Liu, Wanshu Fan +6

School of Software Engineering, Dalian University, Dalian, China · Department of Surgery, Affiliated Zhongshan Hospital of Dalian University, Dalian, China · The Chinese University of Hong Kong, Shenzhen, China +3

Medical AI AgentsAI Agent EvaluationAgentic WorkflowsMultimodal AgentsMulti-Agent LLMsAgentic AI
AskPDF
Mar 2, 2026

CARE: Towards Clinical Accountability in Multi-Modal Medical Reasoning with an Evidence-Grounded Agentic Framework

Large visual language models (VLMs) have shown strong multi-modal medical reasoning ability, but most operate as end-to-end black boxes, diverging from clinicians' evidence-based, staged workflows and hindering clinical accountability. Complementarily, expert visual grounding models can accurately localize regions of interest (ROIs), providing explicit, reliable evidence that improves both reasoning accuracy and trust. In this paper, we introduce CARE, advancing Clinical Accountability in multi-modal medical Reasoning with an Evidence-grounded agentic framework. Unlike existing approaches that couple grounding and reasoning within a single generalist model, CARE decomposes the task into coordinated sub-modules to reduce shortcut learning and hallucination: a compact VLM proposes relevant medical entities; an expert entity-referring segmentation model produces pixel-level ROI evidence; and a grounded VLM reasons over the full image augmented by ROI hints. The VLMs are optimized with reinforcement learning with verifiable rewards to align answers with supporting evidence. Furthermore, a VLM coordinator plans tool invocation and reviews evidence-answer consistency, providing agentic control and final verification. Evaluated on standard medical VQA benchmarks, our CARE-Flow (coordinator-free) improves average accuracy by 10.9% over the same size (10B) state-of-the-art (SOTA). With dynamic planning and answer review, our CARE-Coord yields a further gain, outperforming the heavily pre-trained SOTA by 5.2%. Our experiments demonstrate that an agentic framework that emulates clinical workflows, incorporating decoupled specialized models and explicit evidence, yields more accurate and accountable medical AI. Project page: https://xypb.github.io/CARE-Project-Page/

Yuexi Du, Jinglu Wang, Shujie Liu +2

Microsoft Research Asia · Department of Biomedical Engineering, Yale University · Department of Radiology & Biomedical Imaging, Yale University

Medical VLMsMedical ReasoningMedical VQAMultimodal AgentsMultimodal Grounding
AskPDF
Mar 11, 2026

HeartAgent: An Autonomous Agent System for Explainable Differential Diagnosis in Cardiology

Heart diseases remain a leading cause of morbidity and mortality worldwide, necessitating accurate and trustworthy differential diagnosis. However, existing artificial intelligence-based diagnostic methods are often limited by insufficient cardiology knowledge, inadequate support for complex reasoning, and poor interpretability. Here we present HeartAgent, a cardiology-specific agent system designed to support a reliable and explainable differential diagnosis. HeartAgent integrates customized tools and curated data resources and orchestrates multiple specialized sub-agents to perform complex reasoning while generating transparent reasoning trajectories and verifiable supporting references. Evaluated on the MIMIC dataset and a private electronic health records cohort, HeartAgent achieved over 36% and 20% improvements over established comparative methods, in top-3 diagnostic accuracy, respectively. Additionally, clinicians assisted by HeartAgent demonstrated gains of 26.9% in diagnostic accuracy and 22.7% in explanatory quality compared with unaided experts. These results demonstrate that HeartAgent provides reliable, explainable, and clinically actionable decision support for cardiovascular care.

Shuang Zhou, Kai Yu, Song Wang +11

Division of Computational Health Sciences, Department of Surgery, University of Minnesota, Minneapolis, MN, USA · Department of Computer Science, University of Central Florida, Orlando, FL, USA · College of Science and Engineering, University of Minnesota, Minneapolis, MN, USA +6

Clinical Decision SupportDifferential DiagnosisExplainable Artificial IntelligenceAgentic Reasoning
AskPDF