用多模态特征+轻量模型自动识别临床试验中的用药错误
Automated Detection of Dosing Errors in Clinical Trial Narratives: A Multi-Modal Feature Engineering Approach with LightGBM
- 融合文本、语义、医学模式等3451个特征，用LightGBM建模
- 在4.9%正样本的不平衡数据上达0.8725的ROC-AUC
- 精简特征至500-1000个反而提升性能，证明特征选择很重要
临床试验需严格遵循用药方案，但用药错误仍是影响患者安全和试验完整性的长期难题。本文提出一种自动化系统，通过梯度提升结合全面的多模态特征工程，检测非结构化临床试验文本中的用药错误。方法整合了3,451个特征，涵盖传统NLP（TF-IDF、字符n-gram）、密集语义嵌入（all-MiniLM-L6v2）、领域特定医学模式及基于Transformer的评分（BiomedBERT、DeBERTa-v3），从九个互补文本字段提取（每样本平均5,400字符），覆盖全部42,112条临床试验叙事。在严重类别不平衡的CT-DEB基准数据集上（正样本率4.9%），通过五折集成平均获得0.8725测试ROC-AUC（交叉验证：0.8833 ± 0.0091 AUC）。系统性消融研究显示，移除句子嵌入导致最大性能下降（2.39%），表明其关键作用，尽管仅占总特征重要性的37.07%。特征效率分析表明，选取前500-1000个特征可实现最优性能（0.886-0.887 AUC），优于全特征集（0.879 AUC），体现特征选择对降噪与正则化的价值。结果强调在严重类别不平衡下，稀疏词汇特征与密集表示仍具互补性。
原文摘要 · Abstract (English)
Clinical trials require strict adherence to medication protocols, yet dosing errors remain a persistent challenge affecting patient safety and trial integrity. We present an automated system for detecting dosing errors in unstructured clinical trial narratives using gradient boosting with comprehensive multi-modal feature engineering. Our approach combines 3,451 features spanning traditional NLP (TF-IDF, character n-grams), dense semantic embeddings (all-MiniLM-L6v2), domain-specific medical patterns, and transformer-based scores (BiomedBERT, DeBERTa-v3), used to train a LightGBM model. Features are extracted from nine complementary text fields (median 5,400 characters per sample) ensuring complete coverage across all 42,112 clinical trial narratives. On the CT-DEB benchmark dataset with severe class imbalance (4.9% positive rate), we achieve 0.8725 test ROC-AUC through 5-fold ensemble averaging (cross-validation: 0.8833 + 0.0091 AUC). Systematic ablation studies reveal that removing sentence embeddings causes the largest performance degradation (2.39%), demonstrating their critical role despite contributing only 37.07% of total feature importance. Feature efficiency analysis demonstrates that selecting the top 500-1000 features yields optimal performance (0.886-0.887 AUC), outperforming the full 3,451-feature set (0.879 AUC) through effective noise reduction. Our findings highlight the importance of feature selection as a regularization technique and demonstrate that sparse lexical features remain complementary to dense representations for specialized clinical text classification under severe class imbalance.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。