Loutati Ranel, Ben-Yehuda Arie, Rosenberg Shai, Rottenberg Yakir
Department of Military Medicine and "Tzameret", Faculty of Medicine, Hebrew University of Jerusalem; and the Medical Corps, Israel Defense Forces, Israel; Gaffin Center for Neuro-Oncology, Sharett Institute for Oncology, Hadassah Medical Center and Faculty of Medicine, Hebrew University of Jerusalem, Israel; The Wohl Institute for Translational Medicine, Hadassah Medical Center and Faculty of Medicine, Hebrew University of Jerusalem, Israel.
Department of Internal Medicine, Hadassah Medical Center and Faculty of Medicine, Hebrew University of Jerusalem, Israel.
Am J Med. 2024 Jul;137(7):617-628. doi: 10.1016/j.amjmed.2024.04.002. Epub 2024 Apr 6.
Readmission within 30 days is a prevalent issue among elderly patients, linked to unfavorable health outcomes. Our objective was to develop and validate multimodal machine learning models for predicting 30-day readmission risk in elderly patients discharged from internal medicine departments.
This was a retrospective cohort study which included elderly patients aged 75 or older, who were hospitalized at the Hadassah Medical Center internal medicine departments between 2014 and 2020. Three machine learning algorithms were developed and employed to predict 30-day readmission risk. The primary measures were predictive model performance scores, specifically area under the receiver operator curve (AUROC), and average precision.
This study included 19,569 admissions. Of them, 3258 (16.65%) resulted in 30-day readmission. Our 3 proposed models demonstrated high accuracy and precision on an unseen test set, with AUROC values of 0.87, 0.89, and 0.93, respectively, and average precision values of 0.76, 0.78, and 0.81. Feature importance analysis revealed that the number of admissions in the past year, history of 30-day readmission, Charlson score, and admission length were the most influential variables. Notably, the natural language processing score, representing the probability of readmission according to a textual-based model trained on social workers' assessment letters during hospitalization, ranked among the top 10 contributing factors.
Leveraging multimodal machine learning offers a promising strategy for identifying elderly patients who are at high risk for 30-day readmission. By identifying these patients, machine learning models may facilitate the effective execution of preventive actions to reduce avoidable readmission incidents.
30天内再入院是老年患者中普遍存在的问题，与不良健康结果相关。我们的目标是开发并验证多模式机器学习模型，以预测内科出院老年患者的30天再入院风险。
这是一项回顾性队列研究，纳入了2014年至2020年期间在哈达萨医疗中心内科住院的75岁及以上老年患者。开发并采用了三种机器学习算法来预测30天再入院风险。主要指标是预测模型性能得分，特别是受试者操作特征曲线下面积（AUROC）和平均精度。
本研究包括19569例入院病例。其中，3258例（16.65%）导致30天内再入院。我们提出的3个模型在未见过的测试集上表现出高准确性和精确性，AUROC值分别为0.87、0.89和0.93，平均精度值分别为0.76、0.78和0.81。特征重要性分析表明，过去一年的入院次数、30天再入院史、查尔森评分和住院时长是最具影响力的变量。值得注意的是，自然语言处理得分，即根据住院期间社会工作者评估信训练的基于文本的模型得出的再入院概率，位列前10大影响因素。
利用多模式机器学习为识别有30天再入院高风险的老年患者提供了一种有前景的策略。通过识别这些患者，机器学习模型可能有助于有效实施预防措施，以减少可避免的再入院事件。
