目的 探讨ARIMA乘积季节模型在医院感染患病率预测中的应用,为前瞻性了解医院感染患病率变化趋势和季节性规律提供参考。方法 利用安徽省某三甲医院2013年1月-2017年12月的医院感染患病率数据作为训练数据集,建立ARIMA乘积季节预测模型,选取2018年1-12月的医院感染患病率数据作为验证数据集,评价该模型的预测效能。结果 本次研究最终确定最优模型为ARIMA(0,1,1)×(2,1,0)12,BIC=-2.595,MAPE=23.822,残差序列Ljung-Box Q检验为白噪声(Q=20.258,P=0.209),ARIMA乘积季节模型预测精度良好,绝对偏差在0.01~0.20,相对偏差在1.43%~33.33%,2013年1月-2017年12月实际值与预测值的平均绝对误差百分比为23.78%,2018年1-12月实际值和预测值平均绝对误差百分比为13.60%,实际值均位于95%CI内,医院感染患病率高发于春、冬两季。结论 ARIMA乘积季节模型能有效模拟医院感染患病率时间序列上的变化趋势,对预防和控制医院感染在高发季节的发生提供早期预警及短期预测具有重要的应用价值,可为医院感染的防控提供指导。
Abstract
Objective To explore the application of ARIMA product seasonal model in the prediction of hospital infection prevalence,and provide a reference for the prospective understanding of the trend and seasonality of hospital infection prevalence. Methods Using the data of nosocomial infection rate from January 2013 to December 2017 in a top three hospital in Anhui Province as the training data set,the ARIMA product seasonal prediction model was established,and the data of nosocomial infection rate from January to December 2018 was selected as the validation data set to evaluate the prediction effect of the model. Results The best model was ARIMA (0,1,1)×(0,1,1)12,BIC=-2.595,MAPE=23.822,the residual sequence Ljung-Box Q test was white noise (Q=20.258,P=0.209),the seasonal model of ARIMA product had good prediction accuracy,the absolute deviation was 0.01-0.20,the relative deviation was 1.43%-33.33%,from January 2013 to December 2017,the average absolute error percent age of the predicted and actual values was 23.78%,and from January to December 2018,the average absolute error percent age of the predicted and actual values was 13.75%,the actual value was in 95% CI,the prevalence of nosocomial infection was high in Spring and winter. Conclusion ARIMA product seasonal model can effectively simulate the trend of nosocomial infection rate in time series,which has important application value in early warning and short-term prediction for prevention and control of nosocomial infection in high incidence season,and can provide guidance for prevention and control of nosocomial infection.
关键词
ARIMA模型 /
医院感染 /
季节 /
预测 /
患病率
Key words
ARIMA model /
Nosocomial infection /
Season /
Prediction /
Prevalence
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] KOSTAKOĞ LU U,Saylan S,KARATAŞ M,et al.Cost analysis and evaluation of nosocomial infections in intensive care units[J].Turkish journal of medical sciences,2016,46(5): 1385-1392.
[2] Dawczynski K,Schleussner E,Dobermann H,et al.Infection Prevention in Premature Infants and Newborns in Thuringia: Implementation of Recommendation of the Commission for Hospital Hygiene and Infection Prevention (KRINKO)[J].Zeitschrift fur Geburtshilfe und Neonatologie,2017,221(1): 30-38.
[3] Peterson AMB,Walker PH.Hospital-acquired infections as patient safety indicators[J].Annual review of nursing research,2006,24(1): 75-99.
[4] 刘小燕,冼翠尧,王法霞,等.基层三甲医院2012-2015年医院感染现患率及危险因素[J].中国感染控制杂志,2017,16(11): 1026-1029.
[5] 李金梅,李家斌,王进.综合医院医院感染横断面调查分析[J].中华医院感染学杂志,2015,25(1): 103-105.
[6] 李红,潘东峰,郭忠琴,等.时间序列模型在医院感染发生率拟合预测中的比较研究[J].中国卫生统计,2013(1):87-89.
[7] 刘海鹏,金玉莲,刘光辉,等.住院患儿医院感染发生率ARIMA时间序列模型[J].中国感染控制杂志,2017,16(3):243-246.
[8] 李亚伟,刘玲,宋士勋,等.ARIMA乘法季节模型的R软件实现[J].环境卫生学杂志,2018,8(4): 69-73.
[9] Lin Y,Chen M,Chen G,et al.Application of an autoregressive integrated moving average model for predicting injury mortality in Xiamen,China[J].BMJ Open,2015,5(12):e008491.
[10] Lida B,Rodríguez N,Cecilia M.Smoothing Strategies Combined with ARIMA and Neural Networks to Improve the Forecasting of Traffic Accidents[J].The Scientific World Journal,2014,2014:1-12.
[11] 彭振仁,杨莉,刘勇,等.南宁市2000~2009年道路交通伤害时间序列分析[J].中国公共卫生,2012,28(5): 574-575.
[12] 黄文辉,邹林南.疏系数ARIMA模型预测江西省肺结核发病趋势[J].安徽预防医学杂志,2016,22(3):145-148,179.
[13] 翟和亮,秦伟,杨涛,等.六安市2005~2013年甲型病毒性肝炎流行特征分析及发病趋势预测[J].安徽预防医学杂志,2014,20(6):458-460.
[14] 黄国,朱宇平,黄焕莺.季节性ARIMA模型在江门市手足口病疫情预测中的应用[J].中国卫生统计,2019,36(1):65-67.
[15] 张进,陈国平,曹明华,等.乘法季节ARIMA模型在安徽省细菌性痢疾预测中的应用研究[J].安徽预防医学杂志,2018,24(5):343-345,353.
[16] 中华人民共和国卫生部.医院感染诊断标准(试行)[J].中华医学杂志,2001,81(5): 314-320.
[17] Li Q,Guo NN,Han ZY,et al.Application of an Autoregressive Integrated Moving Average Model for Predicting the Incidence of Hemorrhagic Fever with Renal Syndrome[J].American Journal of Tropical Medicine and Hygiene,2012,87(2):364-370.
[18] Wulff SS.Time Series Analysis: Forecasting and Control,5th edition[J].Journal of Quality Technology,2017,49(4):418-419.
[19] 李红,梁沛枫,潘东峰,等.自回归滑动平均混合模型在医院感染发病率预测中的应用研究[J].中华医院感染学杂志,2013,23(11): 2693-2695.
[20] 王高帅,陈晓娟,梁进娟.基于ARIMA-BPNN模型对医院感染患病率的预测研究[J].中华医院感染学杂志,2017,27(2): 448-451.
[21] 张红秋,王晓丽,刘平,等.季节性医院感染发病的调查分析[J].慢性病学杂志,2006,(9):35-36.
[22] 管利华.ARIMA模型预测医院感染发病状况研究[J].实用预防医学,2013,20(10):1247-1249.
PDF(3412 KB)
