摘要：引言：个例安全报告（ICSR）数量的激增，对药物警戒中由专家主导的因果关系评估构成了巨大挑战。我们开发了 MARCA（多智能体检索增强因果关系评估，Multi-Agent Retrieval-Augmented Causality Assessment），这是一个用于基于 Naranjo 量表进行评估的四智能体框架，并评估了其与人类专家的一致性，以及非 RAG 与带 RAG 标注评分之间的探索性差异。
方法：MARCA 结合了科学文献检索、药品说明书评估、客观证据分类，以及使用 Gemini 2.5 Pro 和 DeepSeek-V3.2 进行的临床因果推理。智能体生成的合理性依据为人工辅助评分提供了参考。专家共识参考标准涵盖了来自 FAERS（2025 年第一季度至第二季度）的 215 份 ICSR 和 1,562 组药物-事件组合（DEC）。一致性通过百分比一致率和 Cohen's kappa 系数进行衡量。
结果：在问题层面上，Naranjo Q2 的一致率为 96.8%（kappa = 0.491；95% CI 0.265-0.663），Q5 为 85.4%（kappa = 0.269；95% CI 0.180-0.340），Q10 为 96.8%（kappa = 0.922；95% CI 0.885-0.951）。最终因果分类一致率为 79.4%（95% CI 71.8-85.3；kappa = 0.587，95% CI 0.450-0.690）。在包含 60 组 DEC 的回顾性子集中，带 RAG 标注的评分在 Q5（83.3% 对比 65.0%）和 Q10（100.0% 对比 18.3%）上的一致率均显著高于非 RAG 评分。
讨论：MARCA 辅助评估在各个 Naranjo 问题间展现出显著的一致性差异，其中对于替代原因的分析一致性较低。回顾性 RAG 对比属于探索性研究，尚未确立前瞻性有效性。

Abstract:Introduction: Increasing volumes of individual case safety reports (ICSRs) challenge expert-led causality assessment in pharmacovigilance. We developed MARCA (Multi-Agent Retrieval-Augmented Causality Assessment), a four-agent framework for Naranjo-based assessment, and evaluated agreement with human experts and exploratory differences between non-RAG and RAG-labelled scores.
Methods: MARCA combined scientific literature retrieval, drug-label assessment, objective-evidence classification, and clinical causality reasoning using Gemini 2.5 Pro and DeepSeek-V3.2. Agent-generated rationales informed human-assisted scoring. A consensus expert reference standard covered 215 ICSRs and 1,562 drug-event combinations (DECs) from FAERS (Q1-Q2 2025). Agreement was measured using percentage agreement and Cohen's kappa.
Results: Question-level agreement was 96.8% for Naranjo Q2 (kappa = 0.491; 95% CI 0.265-0.663), 85.4% for Q5 (kappa = 0.269; 95% CI 0.180-0.340), and 96.8% for Q10 (kappa = 0.922; 95% CI 0.885-0.951). Final causality classification agreement was 79.4% (95% CI 71.8-85.3; kappa = 0.587, 95% CI 0.450-0.690). In a retrospective subset of 60 DECs, agreement was higher for RAG-labelled than non-RAG scores for Q5 (83.3% vs 65.0%) and Q10 (100.0% vs 18.3%).
Discussion: MARCA-assisted assessment showed substantial variation in agreement across Naranjo questions, with lower concordance for alternative causes. Retrospective RAG comparisons are exploratory and do not establish prospective effectiveness.