ARIMA模型在陕西省梅毒发病率预测中的应用效果

Application effect of ARIMA model in prediction of syphilis incidence in Shaanx province

目的分析1999—2016年陕西省梅毒发病情况变化趋势,通过时间序列模型预测我省梅毒未来的发病情况,为制定梅毒防控策略提供科学依据。方法收集1999—2018年陕西省梅毒发病率资料,建立ARIMA时间序列模型并进行发病率拟合预测。结果1999—2003年陕西省梅毒发病率较低且相对稳定,在0.50/10万~0.94/10万;2004—2006年发病率呈明显上升趋势,2007年稍有下降,随后发病率急剧上升,截至2016年末,陕西省梅毒累计发病率为27.13/10万。利用ARIMA模型拟合陕西省梅毒发病趋势效果较好,预测2017、2018年梅毒发病率为28.51/10万和29.92/10万,均在95%CI内。2017、2018年梅毒预测发病率与实际发病率一致,差异无统计学意义(P>0.05)。结论运用ARIMA时间序列模型对我省梅毒发病率进行拟合及预测是可行的,可以为陕西省梅毒防控策略的制定提供理论依据。

Objective To analyze the trend of syphilis incidence in Shaanxi province from 1999 to 2016,forecasts the future incidence of syphilis in Shaanxi Province through time series model,and provides scientific basis for the formulation of syphilis prevention and control strategies.Methods The incidence of syphilis in Shaanxi province was collected from 1999 to 2018.The ARIMA time series model was established and the incidence fitting prediction was implemented.Results From 1999 to 2003,the incidence of syphilis in Shaanxi province was low and relatively stable,which was 0.50/10 million-0.94/10 million.The incidence of syphilis increased obviously from 2004 to 2006,and decreased slightly in 2007,then increased sharply.By the end of 2016,the incidence of syphilis in Shaanxi was 27.13/10 million.ARIMA model has a good effect in the fitting of syphilis incidence trend in Shaanxi province.The predicted incidence of syphilis in 2017 and 2018 was 28.51/10 million and 29.92/10 million respectively,all within 95%CI.The predicted incidences and actual incidences of syphilis in 2017 and 2018 were identical,the difference was not statistically significant(P>0.05).Conclusion It is feasible to fit and predict the incidence of syphilis by using ARIMA time series model,which can provide a theoretical basis for the formulation of syphilis prevention and control strategy in Shaanxi province.