A text-to-tabular approach to generate synthetic patient data using LLMs

2024-12-06

arXiv.org

AI Summary

To address the scarcity of medical data due to privacy constraints, data-sharing barriers, and high acquisition costs, this paper proposes a novel paradigm for synthesizing patient tabular data—requiring neither raw data nor domain-specific training, but only natural-language database descriptions. Methodologically, it leverages large language models (LLMs) integrated with medical knowledge injection, structured prompt engineering, and in-context learning to enable end-to-end, text-driven generation. Its key contributions are: (i) the first systematic validation that LLMs can generate tabular clinical data achieving high fidelity, privacy preservation, and utility while maintaining clinical relevance; (ii) a three-dimensional quantitative evaluation framework and ablation studies identifying critical prompt components. Experiments demonstrate that our approach rivals state-of-the-art models trained on real data across multiple metrics, faithfully preserves disease–symptom–test associations, and supports rapid customization and educational use.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Linked Open Data, Knowledge Graphs & KB Completion

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Abstract

Access to large-scale high-quality healthcare databases is key to accelerate medical research and make insightful discoveries about diseases. However, access to such data is often limited by patient privacy concerns, data sharing restrictions and high costs. To overcome these limitations, synthetic patient data has emerged as an alternative. However, synthetic data generation (SDG) methods typically rely on machine learning (ML) models trained on original data, leading back to the data scarcity problem. We propose an approach to generate synthetic tabular patient data that does not require access to the original data, but only a description of the desired database. We leverage prior medical knowledge and in-context learning capabilities of large language models (LLMs) to generate realistic patient data, even in a low-resource setting. We quantitatively evaluate our approach against state-of-the-art SDG models, using fidelity, privacy, and utility metrics. Our results show that while LLMs may not match the performance of state-of-the-art models trained on the original data, they effectively generate realistic patient data with well-preserved clinical correlations. An ablation study highlights key elements of our prompt contributing to high-quality synthetic patient data generation. This approach, which is easy to use and does not require original data or advanced ML skills, is particularly valuable for quickly generating custom-designed patient data, supporting project implementation and providing educational resources.

Problem

• Generates synthetic patient data without original data access
• Uses LLMs to create realistic data with clinical correlations
• Addresses privacy and scarcity issues in healthcare databases

Innovation

• LLM-based text-to-tabular synthetic data generation
• No original data needed, only database description
• Leverages medical knowledge and in-context learning