Problem
Research questions and friction points this paper is trying to address.
clinical decision support
large language models
structured data reasoning
retrieval hallucination
EMR analytics
Medi-Gemma: A Hybrid Clinical Decision Support System Integrating Deterministic EMR Analytics and Retrieval-Augmented Generation
🤖 AI Summary
This work addresses critical limitations of large language models (LLMs) in high-stakes clinical settings—namely, their propensity for structured hallucinations, inability to perform deterministic reasoning over tabular electronic medical records (EMRs), and susceptibility to missing key information via vector retrieval. To overcome these challenges, the authors propose a decoupled clinical decision support architecture coordinated by a ClinicalOrchestrator that separates clinical perception from data orchestration. A novel Ground Truth Injection Module injects verified, up-to-date structured clinical states prior to generation, while a ProtocolManager and SafetyVerifier jointly enforce adherence to evidence-based guidelines and safety constraints. The system integrates a PandasQueryEngine for deterministic EMR analysis, a CPU-optimized ClinicalRAGEngine, type-enforced data sanitization, and hierarchical intent routing. This design effectively mitigates semantic drift, prevents database failures, and significantly enhances factual consistency between LLM outputs and backend clinical data, establishing a new paradigm for safe and reliable clinical AI deployment.
📝 Abstract
Deploying Large Language Models (LLMs) in high-stakes clinical settings remains limited by structural hallucinations, weak deterministic reasoning over tabular patient data, and omissions in vector retrieval. This paper presents the architecture and validation of Medi-Gemma, a Clinical Decision Support System (CDSS) for wound pathology triage and workflow automation. The platform introduces a decoupled framework that separates clinical perception from data orchestration while preserving traceable reasoning. Medi-Gemma uses a multi-stage pipeline coordinated by a centralized ClinicalOrchestrator. Data requests are handled without generative inference by a DataManager that cleans unstructured Electronic Medical Record (EMR) files through type coercion. Natural language queries are processed by a hierarchical IntentRouter, which routes requests to deterministic analytics paths executed by a PandasQueryEngine or to patient-specific reasoning managed by a ClinicalRAGEngine using a CPU-optimized vector store. A key contribution is the Ground Truth Injection Module, which intercepts patient-specific queries, extracts numeric identification tokens, queries the structured dataframe via Pandas, retrieves the latest validated clinical state, and embeds this snapshot as an overriding context block in the LLM prompt before generation. Safety compliance is enforced by a deterministic ProtocolManager that maps clinical terminology to fixed evidence-based risk pathways, while a SafetyVerifier phrase filter prevents output rule violations. Validation shows that this architecture eliminates semantic context drift, prevents database compilation crashes, and improves factual adherence to backend clinical repositories. These results support Medi-Gemma as a safer pattern for LLM-based clinical decision support where structured data fidelity, retrieval grounding, and deterministic safeguards are essential.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Retrieval-Augmented Generation
Deterministic Reasoning
Clinical Decision Support System
Ground Truth Injection
Structured EMR Analytics
🔎 Similar Papers
No similar papers found.
