出血性脑卒中发病的日节律及SARIMA模型的定量预测

Circadian rhythm and quantitave predication based on SARIMA model of hemorrhagic stroke

目的了解出血性脑卒中（HS）发病的日分布规律，构建时序预测模型，为HS的一级预防提供参考。方法人选3个研究单位2017年6月1至30日诊治的895例HS患者，统计昼夜发病的时间点，应用季节乘积模型（SARIMA）建模。结果各时段的发病例数不等或不全等，P〈0．05，北京时间8：00-11：00时段发病243例（占27．15％），18：00-21：00时段发病297例（占33．18％）。通过识别与检验，逐一尝试法建立了SARIMA（1，0，1）×（1，1，1）。模型，该模型较准确地预测了历史数据。结论北京时间8：00—11：00时和18：00一21：00时为HS的高发时段，SARIMA（1，0，1）×（1，1，1）24模型为HS发病日分布趋势提供了较好的预测，可作为HS防治措施的重要参考依据。

Objective To understand the distribution characteristics of circadian rhythm of hemorrhagic stroke and to develop a time series predication model, so as to provide some references for the primary prevention of hemorrhagic stroke. Methods A total of 895 patients with hemorrhagic stroke from 3 hospital were enrolled from June 1 to 30 in 2017. The circadian occurrence times were collected. The time series analysis was explored using seasonal autoregressive integrated moving average （ SARIMA ） . Results There were statistical differences in occurrence cases between different periods （P 〈 0.05）. 243 cases （ 27.15% ） occurred hemorrhagic stroke from Beijing time 8：00 to 11：00 and 297 cases （ 33.18% ） 18：00 to 21： 00. By identifying and testing, a SARIMA （ 1, 0, 1 ） × （ 1, 1, 1 ） 24 model was developed and it was accurate for the prediction of the history data. Conclusions The peak periods of hemorrhagic stroke occurrence are Beijing time 8：00 to 11： 00 and 18：00 to 21： 00. SARIMA （ 1, 0, 1 ） × （ 1, 1, 1 ） 24 model is accurate for prediction of circadian occurrence tendency of hemorrhagic stroke and can provide more reference information for taking active prevention and treatment measures.