基于气象及流感病毒活动度的流感流行预测研究

目的 探索基于气象及流感病毒活动度建立北京市流感预测模型的可行性.方法 利用北京市2012—2014年流感样病例监测数据、流感病原学监测数据以及气象数据,采用Poisson回归模型,拟合流感病例数,并对流感流行趋势进行预测.结果 2012—2014年,北京市流感病毒活动度及流感发病水平呈现季节性,每周平均气温、气压、降雨量、相对湿度、最大温差和日照时数与每周流感病例数呈负相关.基于两周前流感病毒各亚型的阳性率及两周前的气象因素拟合北京市流感病例数的效果最好,且预测效果最佳.结论 基于气象及流感病毒活动度数据预测流感发病水平是可行的.

A study on the prediction of influenza based on the climate factors and influenza viral activity

Objective To explore the practicability of the prediction of influenza based on the climate factors and influenza viral activity. Methods Based on the data of influenza-like illness (ILI) and virological surveillance, and the climate data during the period of 2012 to 2014, Poisson model was used to fit and to predict the case counts of influenza. Results During the period of 2012 to 2014, the activity of influenza virus and the case number of influenza showed seasonality obviously. Weekly average temperature, atmospheric pressure, precipitation, relative humidity, maximum temperature difference and sunshine duration showed negative correlation with weekly case counts of influenza. The best fit and prediction model of the case number was established base on previous 2-week positive rate by subtype and previous 2-week climate factors. Conclusions Climate factors and viral activity of influenza are suitable for the prediction of the case number of influenza.