基于机器学习的老年创伤性颅脑损伤预后研究

Prognosis on prognostic models of elderly traumatic brain injury based on machine learning

目的建立老年创伤性颅脑损伤预后模型,并分析预后的影响因素。方法收集2009年1月-2019年1月颅脑外伤患者2272例资料,其中老年组患者680例(年龄≥65岁),非老年组1592例(年龄<65岁)。将伤后第3个月格拉斯哥结局评分、住院天数、并发症次数作为终点指标,利用多种机器学习的算法进行两组间预后因素差异的分析。结果老年组患者与非老年组患者相比,预后更差,住院时间更长,两组间差异有统计学意义(P=0.024,P<0.001)。老年组经过筛选,老年组患者的3个终点指标使用多层感知器模型,非老年组中格拉斯哥结局评分使用多层感知器模型,住院天数、并发症次数的预测采用C5.0决策树模型。急诊GCS、具体年龄对老年组患者的预后有更大的影响。结论老年颅脑外伤患者与年轻人所适用的机器学习模型不尽相同,老年人预后更差,年龄和急诊GCS对预后的影响可能更大。

Objective To establish prognostic models of senile traumatic brain injury based on machine learning. Methods A total of 2272 medical records were collected from patients with brain trauma from January 2009 to January 2019 in our hospital, including 680 patients aged ≥65 year(elderly group) and 1 592 patients <65 years(non-elderly group). The Glasgow Outcome Scale(GOS), length of hospital stay, and number of complications in the third month after injury were used as endpoints. A variety of machine learning algorithms were used to analyze the differences in prognostic factors between the two groups. Results Preliminary analysis showed that the elderly patients had a worse prognosis and longer hospital stay than the non-elderly patients(P=0.024, P<0.001). After models screening, the multi-layer perceptron model was used for three endpoints in the elderly group;while the multi-layer perceptron model was used for the GOS in the non-elderly group. The length of hospital stay and the number of complications were predicted using the C5.0 decision tree model. Emergency Glasgow Coma Score(GCS) and age of patients had a greater impact on the prognosis of the elderly group. Conclusion The machine learning models used in elderly patients with traumatic brain injury are different from those in younger patients. The prognosis of the elderly is worse, and the age and emergency GCS are associated with the prognosis of patients.