Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Causal-Enhanced AI Agents for Medical Research Screening

Duc Ngo, Arya Rahgoza

TL;DR

The paper tackles the infeasibility and trust challenges of automated medical evidence synthesis by fusing causal graph reasoning with retrieval-augmented generation. It introduces CEAI, a framework that enforces evidence-first causal relationships grounded in retrieved literature using LightRAG's dual-level retrieval and a Think tool to construct explicit DAGs. The authors demonstrate, on a dementia-exercise corpus of 234 abstracts, that a causal graph–enhanced agent (CausalAgent) achieves 95% accuracy, 100% retrieval success, and zero hallucinations, outperforming baseline and lightweight retrieval systems. The work presents a blueprint for trustworthy, interpretable medical AI that supports high-stakes decision-making and explains mechanisms via causal graphs, with open implementation and clear directions for broader validation. Overall, the approach shows transferable principles for integrating causal reasoning into AI-assisted evidence synthesis in healthcare, improving reliability, interpretability, and clinical utility.

Abstract

Systematic reviews are essential for evidence-based medicine, but reviewing 1.5 million+ annual publications manually is infeasible. Current AI approaches suffer from hallucinations in systematic review tasks, with studies reporting rates ranging from 28--40% for earlier models to 2--15% for modern implementations which is unacceptable when errors impact patient care. We present a causal graph-enhanced retrieval-augmented generation system integrating explicit causal reasoning with dual-level knowledge graphs. Our approach enforces evidence-first protocols where every causal claim traces to retrieved literature and automatically generates directed acyclic graphs visualizing intervention-outcome pathways. Evaluation on 234 dementia exercise abstracts shows CausalAgent achieves 95% accuracy, 100% retrieval success, and zero hallucinations versus 34% accuracy and 10% hallucinations for baseline AI. Automatic causal graphs enable explicit mechanism modeling, visual synthesis, and enhanced interpretability. While this proof-of-concept evaluation used ten questions focused on dementia exercise research, the architectural approach demonstrates transferable principles for trustworthy medical AI and causal reasoning's potential for high-stakes healthcare.

Causal-Enhanced AI Agents for Medical Research Screening

TL;DR

The paper tackles the infeasibility and trust challenges of automated medical evidence synthesis by fusing causal graph reasoning with retrieval-augmented generation. It introduces CEAI, a framework that enforces evidence-first causal relationships grounded in retrieved literature using LightRAG's dual-level retrieval and a Think tool to construct explicit DAGs. The authors demonstrate, on a dementia-exercise corpus of 234 abstracts, that a causal graph–enhanced agent (CausalAgent) achieves 95% accuracy, 100% retrieval success, and zero hallucinations, outperforming baseline and lightweight retrieval systems. The work presents a blueprint for trustworthy, interpretable medical AI that supports high-stakes decision-making and explains mechanisms via causal graphs, with open implementation and clear directions for broader validation. Overall, the approach shows transferable principles for integrating causal reasoning into AI-assisted evidence synthesis in healthcare, improving reliability, interpretability, and clinical utility.

Abstract

Systematic reviews are essential for evidence-based medicine, but reviewing 1.5 million+ annual publications manually is infeasible. Current AI approaches suffer from hallucinations in systematic review tasks, with studies reporting rates ranging from 28--40% for earlier models to 2--15% for modern implementations which is unacceptable when errors impact patient care. We present a causal graph-enhanced retrieval-augmented generation system integrating explicit causal reasoning with dual-level knowledge graphs. Our approach enforces evidence-first protocols where every causal claim traces to retrieved literature and automatically generates directed acyclic graphs visualizing intervention-outcome pathways. Evaluation on 234 dementia exercise abstracts shows CausalAgent achieves 95% accuracy, 100% retrieval success, and zero hallucinations versus 34% accuracy and 10% hallucinations for baseline AI. Automatic causal graphs enable explicit mechanism modeling, visual synthesis, and enhanced interpretability. While this proof-of-concept evaluation used ten questions focused on dementia exercise research, the architectural approach demonstrates transferable principles for trustworthy medical AI and causal reasoning's potential for high-stakes healthcare.
Paper Structure (31 sections, 4 figures, 3 tables)

This paper contains 31 sections, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Problem and Motivation
  3. Research Questions
  4. Contributions
  5. Paper Organization
  6. Literature Review
  7. Retrieval-Augmented Generation
  8. Knowledge Graphs for Information Retrieval
  9. Causal Reasoning with Large Language Models
  10. Hallucination Mitigation Strategies
  11. Methodology
  12. Classifications of Paper for Knowledge Graph
  13. Knowledge Graph and Retrieval
  14. AI Agents
  15. Implementation
  16. ...and 16 more sections

Figures (4)

  • Figure 1: Causal-Enhanced AI Agents Framework
  • Figure 2: Classification Pipeline Architecture
  • Figure 3: AI Agent
  • Figure 4: Responses with Causal Graphs