基于传染病自动预警信息系统的“流行标准”最优化选择分析
|
摘要
目的通过纳入"流行标准"备选模型,探讨各模型对不同传染病类型预警阈值设定的适用性,进而优选出各传染病的适宜预警阈值,改善预警效果。方法按照控制图预警模型原理,分别计算各重点传染病2014年周病例数指定的12个百分位数,然后分别应用备选"流行标准"对各重点传染病2014年相应周的疫情进行预警,通过比较备选模型和控制图预警模型预警结果,优选出预警阈值,然后依据2015年传染病聚集性疫情的实际发生情况验证预警界值预警效果。结果纳入松江区3种重点传染病,流行性腮腺炎整体疫情呈下降趋势,定为"TYPE A",C2、累积和控制图(CUSUM)和季节趋势模型(SM)推荐P_(50),"μ+2σ"推荐P_(80);流行性感冒整体疫情平稳,定为"TYPE B",C2、CUSUM和SM推荐P_(40),"μ+2σ"推荐P_(75);猩红热整体疫情呈上升趋势,为"TYPE C",C2和SM推荐P_(90),CUSUM推荐P75,"μ+2σ"推荐P_(80)。结论 C2、CUSUM和SM适合"TYPE A"型传染病,推荐预警阈值低,结果保守;4种模型均适合"TYPE B"型传染病,但μ+2σ的预警的成本效益好;4种模型也均适合"TYPE C"型传染病,但倾向于推荐大的预警阈值,建议根据传染病社会影响和现有防治水平对预警阈值进行适当调整。
Objective To explore the adaptability of 4 outbreak detection algorithms to provide optimized early warning thresholds( OEWT) for different infectious diseases,and then recommend proper OEWT for each infectious disease to improve early warning effect. Methods According to principle of early warning control graph model( EWCGM),outbreak signals of the 12 alternative Px were calculated in 2014,and ‘μ + 2σ',C2,seasonal model( SM),and cumulative sum( CUSUM) were applied. When outbreak signals generated by algorithm were consistent with a Px,this Px was then ascertained as the optimized threshold by this algorithm,finally all ascertained Px of different infectious diseases were verified in CIDARS by real outbreak events in 2015. Results Three key infectious diseases were finally ascertained,mumps showed a declining trend which was set as TYPE A,C2,CUSUM and SM recommended P_(50) for mumps,and ‘μ + 2σ'recommended P_(80); influenza showed no increasing or decreasing trend which was set as TYPE B,C2,CUSUM and SM recommended P_(40) for mumps,and‘μ + 2σ'recommended P_(75); scarlet fever showed an slightly ascending trend which was set as TYPE C,C2,SM recommened P_(90),CUSUM recommened P_(75),and‘μ + 2σ'recommended P_(80). Conclusion C2,CUSUM,and SM were suitable for TYPE A with lower thresholds,all 4algorithms( OGS) were suitable for TYPE B,and were all also suitable for TYPE C but with higher thresholds. The selection of optimized thresholds should also consider the social and economical influence of infectious diseases as well as the response capacity of local CDCs.
Objective To explore the adaptability of 4 outbreak detection algorithms to provide optimized early warning thresholds( OEWT) for different infectious diseases,and then recommend proper OEWT for each infectious disease to improve early warning effect. Methods According to principle of early warning control graph model( EWCGM),outbreak signals of the 12 alternative Px were calculated in 2014,and ‘μ + 2σ',C2,seasonal model( SM),and cumulative sum( CUSUM) were applied. When outbreak signals generated by algorithm were consistent with a Px,this Px was then ascertained as the optimized threshold by this algorithm,finally all ascertained Px of different infectious diseases were verified in CIDARS by real outbreak events in 2015. Results Three key infectious diseases were finally ascertained,mumps showed a declining trend which was set as TYPE A,C2,CUSUM and SM recommended P_(50) for mumps,and ‘μ + 2σ'recommended P_(80); influenza showed no increasing or decreasing trend which was set as TYPE B,C2,CUSUM and SM recommended P_(40) for mumps,and‘μ + 2σ'recommended P_(75); scarlet fever showed an slightly ascending trend which was set as TYPE C,C2,SM recommened P_(90),CUSUM recommened P_(75),and‘μ + 2σ'recommended P_(80). Conclusion C2,CUSUM,and SM were suitable for TYPE A with lower thresholds,all 4algorithms( OGS) were suitable for TYPE B,and were all also suitable for TYPE C but with higher thresholds. The selection of optimized thresholds should also consider the social and economical influence of infectious diseases as well as the response capacity of local CDCs.
引文
[1]马芬,王丽,李辉.传染病预警方法研究进展.卫生研究,2008,37(2):249-251.
[2]杨维中,邢慧娴,王汉章,等.七种传染病控制图法预警技术研究.中华流行病学杂志,2004,25(12):1039-1041.
[3]孟军.突发公共卫生事件预警现状与模式探讨.广西医学,2006,28(11):1742-1743.
[4]葛利辉,徐斌,梁红慧.利用电子地图实行传染病预警预测的探讨.实用预防医学,2006,13(1):37-38.
[5]Najera J,Kouznetasov R,Dilacollette C.Malaria epidemics.Detection and control.Forecasting and prevention.Geneva,World Health Organization,1998.
[6]Early warning syetems.Communicable disease surveillance and response(CSR).2004.http://w w w.w ho.int/csr/labepidemiology/projects/earlyw arnsystem.
[7]杨维中,兰亚佳,李中杰,等.国家传染病自动预警系统的设计与应用.中华流行病学杂志,2010,31(11):1240-1244.
[8]王瑞平,陈立凌,刘成,等.苏州市重点传染病移动百分位数法预警界值优选分析.中国卫生统计,2014,31(4):598-601.
[9]Yang W,Li Z,Lan Y,et al.A nationwide web based automated system for early outbreak detection and rapid response in China.WPSAR,2011,2(1):1-6.
[10]杨维中,李中杰,赖圣杰,等.国家传染病自动预警系统运行状况分析.中华流行病学杂志,2011,32(5):431-435.
[11]梁万年.加强我国传染病预警的研究和应用.中华流行病学杂志,2011,32(5):429-430.
[12]刘文东,胡建利,艾静,等.CUSUM模型在流行性腮腺炎早期预警中的应用研究.中国卫生统计,2014,31(4):563-565.
[13]徐旭卿,鲁琴宝,王臻,等.浙江省传染病自动预警系统暴发预警效果评价.中华流行病学杂志,2011,32(5):442-445.
[2]杨维中,邢慧娴,王汉章,等.七种传染病控制图法预警技术研究.中华流行病学杂志,2004,25(12):1039-1041.
[3]孟军.突发公共卫生事件预警现状与模式探讨.广西医学,2006,28(11):1742-1743.
[4]葛利辉,徐斌,梁红慧.利用电子地图实行传染病预警预测的探讨.实用预防医学,2006,13(1):37-38.
[5]Najera J,Kouznetasov R,Dilacollette C.Malaria epidemics.Detection and control.Forecasting and prevention.Geneva,World Health Organization,1998.
[6]Early warning syetems.Communicable disease surveillance and response(CSR).2004.http://w w w.w ho.int/csr/labepidemiology/projects/earlyw arnsystem.
[7]杨维中,兰亚佳,李中杰,等.国家传染病自动预警系统的设计与应用.中华流行病学杂志,2010,31(11):1240-1244.
[8]王瑞平,陈立凌,刘成,等.苏州市重点传染病移动百分位数法预警界值优选分析.中国卫生统计,2014,31(4):598-601.
[9]Yang W,Li Z,Lan Y,et al.A nationwide web based automated system for early outbreak detection and rapid response in China.WPSAR,2011,2(1):1-6.
[10]杨维中,李中杰,赖圣杰,等.国家传染病自动预警系统运行状况分析.中华流行病学杂志,2011,32(5):431-435.
[11]梁万年.加强我国传染病预警的研究和应用.中华流行病学杂志,2011,32(5):429-430.
[12]刘文东,胡建利,艾静,等.CUSUM模型在流行性腮腺炎早期预警中的应用研究.中国卫生统计,2014,31(4):563-565.
[13]徐旭卿,鲁琴宝,王臻,等.浙江省传染病自动预警系统暴发预警效果评价.中华流行病学杂志,2011,32(5):442-445.