摘要
[目的]通过对医院门诊逐日处方量的监测,建立流感样疾病就诊量的监测模型,从而,间接地监测该病在区域内流行情况,并做出早期预警,以采取快速反应,减少疾病所造成的危害。[方法]搜集了2002年8月1日至2005年12月8日北京某综合医院门急诊就诊人次、门急诊流感样疾病患者人次和治疗流感样疾病的处方量,计算日均该类药物处方量,建立序列图,对该序列进行对数和差分变换后,建立时间序列监测模型。[结果]该院日人均治疗流感样疾病的药物处方量为(0.22±0.21)单位,此序列存在一定的自相关性,但不存在季节效应,最佳拟合模型为Yt=0.14Yt-1+et-0.16et-1。模型的敏感性为0.87,外推预测的平均相对误差为0.307。[结论]对特定治疗药物处方监测是流感样疾病监测的重要手段之一。所建立的时间序列模型可为流感样疾病的暴发做出早期预警。基于症状监测的监测网络的建立将提高传染病早期预警的准确性和灵敏性。
[Objective]To develop the modeling for monitoring the outbreak of flu-like diseases using the prescriptions data for treating the flu-like diseases.[Methods]Time series analysis was conducted through developing an ARIMA（1,1, 1） model using data of prescriptions for treating the flu-like diseases in an general hospital in Beijing from August 1st, 2002 to August 1 st,2005,and the model was tested by the data from August 2nd, 2005 to December 8th, 2005, [Results] The mean of the prescriptions per day was （0.22±0.21） in the hospital. The pattern of it presented an autocorrelation and did not present the season pattern. The model was the best one for surveillance for flu-like diseases based on the prescriptions in the hospital. The sensitivity of the model was 0.87 and the relative error of model for forecasting was percentage of 30.7. [Conclusion]The surveillance of the prescriptions was one of important tool for monitoring the infectious diseases. The time series model can be used to detect the abnormal pattern of flu-like diseases. The veracity and sensitivity of the models for early warning can be improved by constructing surveillance network based on syndrome among all hospitals in a city,as well as all country.
出处
《预防医学论坛》
2007年第10期875-877,共3页
Preventive Medicine Tribune
关键词
处方监测
流感样疾病
预警模型
时间序列分析
Surveillance of the prescriptions
Modeling for early warning
Flu-like diseases
Time series analysis
