季节性时间序列分析在医院感染中的应用

Seasonal Time Sequence Analysis Applying for Nosocomial Infection

目的为了研究医院感染人数与不同季节的关系.方法利用季节性时间序列的线性模型对1996～2002年的数据建模,并比较各个模型对2003年的数据预测能力,从而总结出最佳模型.结果最后得到4个模型,各模型的P值均＜0.0001,说明新建立的模型是可靠的;预测的结果表明最后一个模型的效果最好.结论季节性时间序列的线性模型能较好的模拟医院感染人数同季节的关系,并有较强的预测能力,从而为医院感染人数的预测提供了有效的工具.

OBJECTIVE To investigate the relationship between bacterial infection number in hospitals and different seasons. METHODS A seasonal time sequence linear model was applied to establish a model based on the data obtained from the years 1996 to 2002, then compare the ability of every model to predict data on 2003 and pick out the best model. RESULTS We got four models at last with the P value <0.0001 each, which showed the new constructed model was confidential. From predicted results, the last model was the best one. CONCLUSIONS Seasonal time sequence analysis can well simulate the relation of infection numbers and seasons. Its strong predictable ability can afford an effective tool to forecast the numbers of nosocomial infection.