基于SHAP的预测建模,针对老年心力衰竭患者1年全因再入院风险:特征选择和模型解释
Hao Luo1, Congyu Xiang2, Lang Zeng1
1Department of Cardiology, Affiliated Hospital of North Sichuan Medical College, No. 63, Wenhua Road, Nanchong, 637000, Sichuan Province, People's Republic of China.
Scientific reports
|July 31, 2024
概括
一种新的人机协作模型显著改善了对老年患者心力衰竭再入院的预测. 这种方法将专家知识与机器学习相结合,以实现更准确的风险评估和个性化护理.
科学领域:
- 心脏病学 心脏病学
- 医疗保健中的人工智能
- 老年医学 老年医学
背景情况:
- 心力衰竭 (HF) 对老年人构成重大健康风险,其特点是高的再入院率.
- 现有的机器学习 (ML) 模型用于高频回收预测,往往缺乏与临床专业知识的整合.
- 优化再接收风险预测对于管理老年HF患者和减少医疗保健负担至关重要.
研究的目的:
- 开发和评估一种新的人机协作模型,用于预测老年心力衰竭患者因任何原因的再入院情况.
- 将协作模型的预测性能与仅使用机器选择或人类选择的特征的模型进行比较.
- 为了提高模型的可解释性,使用SHapley添加式扩展 (SHAP) 方法进行临床应用.
主要方法:
- 从2018年1月到2021年12月,对8396名老年HF患者进行了回顾性分析.
- 使用ML算法 (XGBoost,LASSO,随机森林) 和专家心血管教授进行特征选择.
- 使用CatBoost开发模型,通过AUC,F1得分和Brier得分评估性能.
- 对特征重要性和可解释性的SHAP分析.
主要成果:
- 人机协作模型实现了0.83617的优异AUC,0.73521的F1得分和0.16536.6的Brier得分.
- 该模型的性能优于仅基于专家或机器选择的特征的模型.
- SHAP分析确定了HGB,NT-proBNP,吸烟史,NYHA分类和LVEF作为关键预测因素.
结论:
- 在特征选择中的人机协作导致了对老年HF患者全因再入院的优异预测.
- 通过组合的特征选择,CatBoost模型提供了增强的预测准确性.
- SHAP分析为风险因素提供了有价值的见解,促进了个性化治疗策略和改善患者护理.
相关概念视频
01:20Actuarial Approach
70
The actuarial approach, a statistical method originally developed for life insurance risk assessment, is widely used to calculate survival rates in clinical and population studies. This method accounts for participants lost to follow-up or those who die from causes unrelated to the study, ensuring a more accurate representation of survival probabilities.
Consider the example of a high-risk surgical procedure with significant early-stage mortality. A two-year clinical study is conducted,...
Consider the example of a high-risk surgical procedure with significant early-stage mortality. A two-year clinical study is conducted,...
70
01:24Kaplan-Meier Approach
117
The Kaplan-Meier estimator is a non-parametric method used to estimate the survival function from time-to-event data. In medical research, it is frequently employed to measure the proportion of patients surviving for a certain period after treatment. This estimator is fundamental in analyzing time-to-event data, making it indispensable in clinical trials, epidemiological studies, and reliability engineering. By estimating survival probabilities, researchers can evaluate treatment effectiveness,...
117