Patient-Zero: A Unified Framework for Real-Record-Free Patient Agent Generation

2025-09-13

AI Summary

Existing virtual patient generation methods rely on real-world electronic health records (EHRs), posing critical challenges including patient privacy leakage, clinical inaccuracy, and rigid, non-adaptive interactions. To address these limitations, we propose the first EHR-free, multi-step virtual patient generation framework. Our approach ensures clinical plausibility through hierarchical medical knowledge injection, while real-time clinical logic validation and dynamic state updating guarantee controllable generation quality. Furthermore, we design an adaptive dialogue policy to enable high-consistency, context-aware physician–patient interaction simulation. Extensive evaluation demonstrates that our method significantly outperforms existing baselines in accuracy, response diversity, and contextual coherence. On the MedQA benchmark, it improves downstream model performance by a statistically significant margin. This work establishes a novel paradigm for training privacy-preserving, clinically verifiable, and interactive medical AI systems.

Technology Category

• Humans and AI: Game Design — Virtual Humans, NPCs and Autonomous Characters
• Natural Language Processing: Generation
• Machine Learning: Privacy

Application Category

• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Synthetic data generation using large language models (LLMs) has emerged as a promising solution across various domains, particularly in medical field, to mitigate data collection challenges. However, existing studies mainly utilize LLMs to rewrite and complete existing medical records, where the limitations in data privacy, accuracy, and diversity sill exist, and additionally lack the ability to interact like real patients. To address these issues, we propose a realistic patient generation framework, Patient-Zero, which requires no real medical records. Patient-Zero first introduces a medically-aligned multi-step generation architecture, which builds comprehensive patient records through hierarchical medical knowledge injection without real medical records. Then, to optimize the virtual patient's interaction abilities with humans, Patient-Zero designs a dynamic updating mechanism to improve the consistency and conversational performance. Our framework enables the generation of contextually diverse patient records while maintaining strict medical coherence, supported by adaptive dialogue strategies and real-time clinical plausibility verification. Experimental results demonstrate that our model achieves good performance in accuracy, diversity, and consistency. After training with our generated virtual patients, existing models show significant improvements on the MedQA dataset.

Problem

• Generates synthetic patient data without real medical records
• Improves interaction abilities of virtual patients with humans
• Ensures medical coherence and diversity in patient records

Innovation

• Multi-step architecture with medical knowledge injection
• Dynamic updating mechanism for interaction consistency
• Real-time plausibility verification for medical coherence