Chae Sena, Song Jiyoun, Ojo Marietta, Topaz Maxim
University of Iowa College of Nursing.
Columbia University School of Nursing.
Stud Health Technol Inform. 2021 Dec 15;284:15-19. doi: 10.3233/SHTI210653.
The goal of this natural language processing (NLP) study was to identify patients in home healthcare with heart failure symptoms and poor self-management (SM). The preliminary lists of symptoms and poor SM status were identified, NLP algorithms were used to refine the lists, and NLP performance was evaluated using 2.3 million home healthcare clinical notes. The overall precision to identify patients with heart failure symptoms and poor SM status was 0.86. The feasibility of methods was demonstrated to identify patients with heart failure symptoms and poor SM documented in home healthcare notes. This study facilitates utilizing key symptom information and patients' SM status from unstructured data in electronic health records. The results of this study can be applied to better individualize symptom management to support heart failure patients' quality-of-life.
这项自然语言处理（NLP）研究的目标是识别家庭医疗保健中出现心力衰竭症状且自我管理（SM）较差的患者。确定了症状和SM状况不佳的初步清单，使用NLP算法对清单进行完善，并利用230万份家庭医疗保健临床记录评估NLP性能。识别出现心力衰竭症状且SM状况不佳患者的总体精确度为0.86。证明了从家庭医疗保健记录中识别出现心力衰竭症状且有记录的SM状况不佳患者的方法具有可行性。本研究有助于利用电子健康记录中非结构化数据中的关键症状信息和患者的SM状况。本研究结果可应用于更好地实现症状管理个性化，以支持心力衰竭患者的生活质量。
