Health Outcomes and Biomedical Informatics, College of Medicine, University of Florida, Gainesville, FL, United States.
Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, TX, United States.
JMIR Aging. 2024 Nov 27;7:e58980. doi: 10.2196/58980.
Transcatheter aortic valve replacement (TAVR) is a commonly used treatment for severe aortic stenosis. As degenerative aortic stenosis is primarily a disease afflicting older adults, a frailty assessment is essential to patient selection and optimal periprocedural outcomes.
This study aimed to enhance frailty assessments of TAVR candidates by integrating real-world structured and unstructured data.
This study analyzed data from 14,000 patients between January 2018 and December 2019 to assess frailty in TAVR patients at the University of Florida. Frailty was identified using the Fried criteria, which includes weight loss, exhaustion, walking speed, grip strength, and physical activity. Latent Dirichlet allocation for topic modeling and Extreme Gradient Boosting for frailty prediction were applied to unstructured clinical notes and structured electronic health record (EHR) data. We also used least absolute shrinkage and selection operator regression for feature selection. Model performance was rigorously evaluated using nested cross-validation, ensuring the generalizability of the findings.
Model performance was significantly improved by combining unstructured clinical notes with structured EHR data, achieving an area under the receiver operating characteristic curve of 0.82 (SD 0.07), which surpassed the EHR-only model's area under the receiver operating characteristic curve of 0.64 (SD 0.08). The Shapley Additive Explanations analysis found that congestive heart failure management, back problems, and atrial fibrillation were the top frailty predictors. Additionally, the latent Dirichlet allocation topic modeling identified 7 key topics, highlighting the role of specific medical treatments in predicting frailty.
Integrating unstructured clinical notes and structured EHR data led to a notable enhancement in predicting frailty. This method shows great potential for standardizing frailty assessments using real-world data and improving patient selection for TAVR.
经导管主动脉瓣置换术（TAVR）是治疗严重主动脉瓣狭窄的常用方法。由于退行性主动脉瓣狭窄主要影响老年人，因此在患者选择和最佳围手术期结果方面，虚弱评估至关重要。
本研究旨在通过整合真实世界的结构化和非结构化数据来增强 TAVR 患者的虚弱评估。
本研究分析了 2018 年 1 月至 2019 年 12 月期间 14000 名患者的数据，以评估佛罗里达大学 TAVR 患者的虚弱状况。使用 Fried 标准来确定虚弱，包括体重减轻、疲惫、行走速度、握力和体力活动。应用潜在狄利克雷分配（Latent Dirichlet Allocation）进行主题建模和极端梯度提升（Extreme Gradient Boosting）进行虚弱预测，以分析非结构化临床记录和结构化电子健康记录（EHR）数据。我们还使用最小绝对收缩和选择算子回归（least absolute shrinkage and selection operator regression）进行特征选择。使用嵌套交叉验证严格评估模型性能，以确保结果的可推广性。
通过将非结构化临床记录与结构化 EHR 数据相结合，模型性能得到显著提高，接收者操作特征曲线下面积为 0.82（SD 0.07），超过了仅基于 EHR 数据的模型的 0.64（SD 0.08）。Shapley 加法解释分析发现充血性心力衰竭管理、背部问题和心房颤动是虚弱的主要预测因素。此外，潜在狄利克雷分配主题建模确定了 7 个关键主题，突出了特定医疗治疗在预测虚弱方面的作用。
整合非结构化临床记录和结构化 EHR 数据可显著提高预测虚弱的能力。该方法具有使用真实世界数据标准化虚弱评估和改善 TAVR 患者选择的巨大潜力。
