Zhu Letao, Li Yanfeng, Zhao Qingshi, Li Changyu, Wu Zongbi, Jiang Youli
Department of Neurology, People's Hospital of Longhua, Shenzhen, 518109, People's Republic of China.
Nursing Department, Shenzhen Traditional Chinese Medicine Hospital, The Fourth Clinical Medical School of Guangzhou University of Chinese Medicine, Shenzhen, Guangdong, 518033, People's Republic of China.
Risk Manag Healthc Policy. 2024 Nov 1;17:2635-2645. doi: 10.2147/RMHP.S476106. eCollection 2024.
Acute ischemic stroke (AIS) is increasingly affecting younger populations, necessitating prompt thrombolytic therapy within a narrow therapeutic window. Pre-hospital delays are prevalent, particularly in China, yet targeted research on the youth population remains scarce.
In this retrospective cohort study, data from AIS patients aged 18-50 admitted to Longhua District People's Hospital, Shenzhen from December 2021 to December 2023 were analyzed using XGBoost and Random Forest machine learning algorithms, coupled with SHAP visualization, to identify factors contributing to pre-hospital delays.
Among 1954 AIS patients, 528 young patients were analyzed. The median time to hospital arrival was 8.34 hours, with 82.0% experiencing delays. Analysis of different age subgroups showed that young patients aged 36-50 years old had a higher delay rate than patients under 36 years old. Machine learning algorithms identified stroke awareness, age, TOAST classification, ambulance arrival, dysarthria, mRS on admission, dizziness, wake-up stroke, etc. as important determinants of delay.
This study highlights the necessity of machine learning in identifying delay risk factors in young stroke patients. Enhanced public education, particularly regarding stroke symptoms and the use of emergency services, is crucial for reducing pre-hospital delays and improving patient outcomes.
急性缺血性卒中（AIS）对年轻人群的影响日益增加，这就需要在狭窄的治疗时间窗内迅速进行溶栓治疗。院前延误情况普遍存在，在中国尤为如此，但针对青年人群的针对性研究仍然匮乏。
在这项回顾性队列研究中，对2021年12月至2023年12月期间入住深圳龙华区人民医院的18至50岁AIS患者的数据，使用XGBoost和随机森林机器学习算法，并结合SHAP可视化分析，以确定导致院前延误的因素。
在1954例AIS患者中，分析了528例年轻患者。到达医院的中位时间为8.34小时，82.0%的患者存在延误。对不同年龄亚组的分析表明，36至50岁的年轻患者延误率高于36岁以下的患者。机器学习算法确定卒中知晓情况、年龄、TOAST分类、救护车到达情况、构音障碍、入院时的改良Rankin量表评分、头晕、醒后卒中等等是延误的重要决定因素。
本研究强调了机器学习在识别年轻卒中患者延误风险因素方面必要性。加强公众教育，特别是关于卒中症状和使用急救服务方面的教育，对于减少院前延误和改善患者预后至关重要。
