跳到正文
arXiv cs.AI· Guanqiao Chen, Wei Jiang, Shu Zhang, Qian Lin, Guanxu Yu, Siran Li, Chunhui Wang, Tianyu Shi·· 2 小时前AI 评分54

arXiv 论文提出临床 AI 决策轨迹审计方法,最终诊断正确不等于路径安全

Right Diagnosis, Unsafe Route: Auditing Clinical AI Decision Trajectories and Corrective Oversight for Validity and Potential Harm

AI 导读

arXiv 论文(arXiv:2611.22612)提出以患者安全为目标的临床 AI 决策轨迹审计方法,逐步重放真实入院病例并评估每项决策的潜在伤害与临床有效性,用 LLM judge 打分并以医生参考标准校准。

正文 · 原文

View PDF HTML (experimental)

Abstract:A clinical AI system can reach a correct final diagnosis through intermediate decisions that could have harmed the patient, and can repair an error while introducing new risks. Most evaluations score final diagnoses; even stepwise benchmarks, which release a case's evidence over time, score whether systems reach the right diagnosis or gather the right information, not whether the decisions a patient passes through are safe. We make patient safety the target of stepwise evaluation. A trajectory audit replays real admissions checkpoint by checkpoint, lets each system decide with only the evidence released so far, and rates every test, treatment and disposition for its potential harm to the patient and, separately, for clinical validity, against the documented final diagnosis and physician-written disease rubrics yet on the evidence available when the decision was made. An LLM judge scores every decision, and a physician reference standard (triple annotation and senior adjudication of a stratified sample) calibrates the rates. An oversight audit branches each trajectory at its first actionable error and measures, on the same branches, whether corrective feedback repairs the error and whether it introduces new harm. Across 300 cases from a hospital cardiology cohort, MIMIC-IV and the public ASCENT benchmark, with up to seven AI systems per source, 55.2% of hospital trajectories reaching an acceptable final diagnosis passed through a decision with moderate or severe potential harm and 24.8% through a clinically unacceptable one; the harm rates were 76.2% on MIMIC-IV and 39.6% on ASCENT. Structured corrections resolved 73.6% of first errors versus 24.8% after bare rejection, yet new potentially unsafe decisions followed 30.2% of corrections versus 19.7% of rejections. A correct final diagnosis does not certify a safe trajectory, and a successful correction does not certify a safe recovery.
Subjects: Artificial Intelligence (cs.AI)
Cite as: arXiv:2611.22612 [cs.AI]
  (or arXiv:2611.22612v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2611.22612

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Guanqiao Chen [view email]
[v1] Tue, 17 Nov 2026 11:26:37 UTC (88 KB)

来源:arXiv cs.AI · arxiv.org