武冈市农村地区心脑血管住院病例的时间序列预测分析

目的  建立武冈市农村地区心脑血管疾病（cardio-cerebrovascular disease，CVD）住院病例的预测模型，并对CVD住院病例的变化趋势进行预测分析，为医院合理配置CVD科室医疗资源提供参考依据。 方法  利用Stata 14.0软件对武冈市2013年1月~2016年12月农村地区CVD住院人次月度数据构建季节性自回归移动平均混合模型（seasonal autoregressive integrated moving average model，SARIMA），并对2017年武冈市农村地区CVD住院病例进行预测分析。 结果  通过模型构建最终拟合的CVD住院病例预测模型为SARIMA（2，1，1）x（0，1，0）₁₂。Ljung-Box Q检验结果显示残差序列为白噪音序列（Q=11.12，P=0.680），说明所建模型拟合度较好，且2017年的预测结果与观测结果基本一致，总体相对误差在-1.2%左右。预测结果显示，夏季为每年CVD住院高峰期。 结论  SARIMA模型可以对武冈市CVD住院病例进行较准确的短期预测，医院可以根据不同月份CVD就医需求合理配置院内CVD科室医疗资源。

A time-series prediction and analysis on rural inpatient with cardio-cerebrovascular disease in Wugang

  Objective  To establish a predictive model for inpatients of cardio-cerebrovascular disease in rural areas of Wugang through time series analysis, and predict the changing trend of cardio-cerebrovascular disease, so as to offer guidance for the health care resources allocation and prevention and control of cardio-cerebrovascular disease.  Methods  The seasonal autoregressive integrated moving average model (SARIMA) was constructed based on the monthly number of cases of cardio-cerebrovascular disease in rural areas from January 2013 to December 2016 by Stata 14.0 software, and the predictive effect of the model was verified with the monthly number of inpatients of cardio-cerebrovascular disease in 2017.  Results  The final fitting model of inpatients of cardio-cerebrovascular disease was SARIMA (2, 1, 1)×(0, 1, 0)₁₂. The residual sequence of the model was diagnosed. Results of Ljung-Box Q test showed that the residual sequence was white noise sequence (Q=11.12, P=0.68). In addition, the 2017 forecast was basically consistent with the observations, the overall relative error was around -1.2%. The results showed that the summer was the peak period of cardiovascular and cerebrovascular hospitalization.  Conclusion  SARIMA model can accurately predict the number of inpatients of cardio-cerebrovascular disease in Wugang, which can provide data support for the hospital administrator to rationally allocate medical resources in the cardiovascular according to the needs of cardio-cerebrovascular treatment in different months.