arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.10109cs.AIcs.CLcs.LG
收藏

迈向通过可解释风险预测减少诊断错误

Towards Reducing Diagnostic Errors with Interpretable Risk Prediction

Northeastern University(东北大学);Brigham and Women’s Hospital(布莱根妇女医院);University of Amsterdam(阿姆斯特丹大学)

Denis Jered McInerney, William Dickinson, Lucy C. Flynn, Andrea C. Young, Geoffrey S. Young, Jan-Willem van de Meent, Byron C. Wallace

AI总结:

本研究提出利用大语言模型识别电子健康记录中的证据,结合神经加性模型提供个体化风险预测,以减少诊断错误并缓解诊断延迟。

摘要:

许多诊断错误的发生是因为临床医生无法轻松访问患者电子健康记录(EHR)中的相关信息。在这项工作中,我们提出了一种方法,利用大语言模型(LLMs)识别患者EHR数据中指示特定诊断风险增加或降低的证据片段;我们的最终目标是增加对证据的访问并减少诊断错误。具体而言,我们提出了一种神经加性模型(Neural Additive Model),在临床医生仍不确定的时间点,基于证据进行预测并提供个体化风险估计,旨在特别缓解因不完整鉴别诊断而导致的诊断延迟和错误。为了训练这样的模型,有必要推断最终“真实”诊断的时间上细粒度的回顾性标签。我们使用LLMs来实现这一点,以确保输入文本来自能够做出自信诊断之前的时间点。我们使用LLM检索初始证据池,但随后根据模型学习到的相关性来细化这一证据集。我们通过模拟临床医生如何使用该方法在预定义的鉴别诊断列表中进行决策,深入评估了我们方法的实用性。

英文摘要:

Many diagnostic errors occur because clinicians cannot easily access relevant information in patient Electronic Health Records (EHRs). In this work we propose a method to use LLMs to identify pieces of evidence in patient EHR data that indicate increased or decreased risk of specific diagnoses; our ultimate aim is to increase access to evidence and reduce diagnostic errors. In particular, we propose a Neural Additive Model to make predictions backed by evidence with individualized risk estimates at time-points where clinicians are still uncertain, aiming to specifically mitigate delays in diagnosis and errors stemming from an incomplete differential. To train such a model, it is necessary to infer temporally fine-grained retrospective labels of eventual "true" diagnoses. We do so with LLMs, to ensure that the input text is from before a confident diagnosis can be made. We use an LLM to retrieve an initial pool of evidence, but then refine this set of evidence according to correlations learned by the model. We conduct an in-depth evaluation of the usefulness of our approach by simulating how it might be used by a clinician to decide between a pre-defined list of differential diagnoses.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑