青岛市某三甲医院痛风患者就诊时间序列分析

A TIME SERIES ANALYSIS OF GOUT PATIENTS IN A GRADE A TERTIARY HOSPITAL IN QINGDAO,CHINA

目的通过分析痛风患者就诊的时间序列资料,探讨患者的就诊规律,为医疗卫生部门的痛风防治工作提供参考。方法收集青岛市某三甲医院2013—2018年痛风患者就诊的时间序列资料,对所有患者的就诊时间、性别、年龄等资料进行描述性分析。利用2013—2018年痛风患者时间序列资料建立差分自回归移动平均(ARIMA)模型,再以2019年1月—10月痛风患者就诊例次数据进行模型预测效果的外推验证,并采用X-11法分析患者时间序列的季节因子、长期趋势和随机波动。结果2013—2018年就诊的痛风患者中,男性占94.68%,女性占5.32%;年龄构成中,30~39岁占21.50%,40~49岁占21.67%,50~59岁占19.74%,60岁以上占23.32%。经2013—2018年痛风患者时间序列资料建立的ARIMA模型为ARIMA(0,1,1)×(0,1,1)12,AIC值为674.89,SBC值为679.05,用于模型预测效果的外推验证的MAE值为86.28,MAPE值为7.64%。经X-11法进行稳定性季节检验(F=27.81,P<0.05)及移动性季节检验(F=1.06,P>0.05),显示痛风患者就诊时间序列具有稳定识别的季节性且不受时间的影响,每年的7、8月患者的就诊例次高于平均值,每年的2月低于平均值,其他月份较为平稳。结论ARIMA模型有效且预测结果较稳定,结合X-11法提取的季节因子与长期趋势,能够较好地解释痛风患者的就诊规律,可为卫生主管部门、医疗机构制定痛风预防控制政策、开展健康宣教与进行人力资源配置提供数据参考。

Objective To investigate the rules of gout patients attending the hospital by analyzing their time series data,and to provide a reference for the prevention and treatment of gout in medical and health departments.Methods Time series data were collected from the gout patients who attended a grade A tertiary hospital in Qingdao from 2013 to 2018,and a descriptive analysis was performed for the data including the time of patients attending the hospital,age,and sex.An Autoregressive Integra-ted Moving Average(ARIMA)model was established based on the time series data of gout patients in 2013—2018,and the data of gout patients from January to October,2019,were used for the extrapolation validation of this model.The X-11 method was used to analyze seasonal factors,long-term trends,and random fluctuations.Results Among the gout patients attending the hospital in 2013—2018,male patients accounted for 94.68%and female patients accounted for 5.32%;in terms of age composition,the patients aged 30-39 years,40-49 years,50-59 years,and>60 years accounted for 21.50%,21.67%,19.74%,and 23.32%,respectively.The ARIMA model established based on time series data of the gout patients in 2013—2018 was ARIMA(0,1,1)×(0,1,1)12,with an AIC value of 674.89 and an SBC value of 679.05,and the extrapolation verification of the predictive effect of this model showed an MAE value of 86.28 and an MAPE value of 7.64%.The X-11 method for seasonal stability(F=27.81,P<0.05)and seasonal mobility(F=1.06,P>0.05)showed that the time series of gout patients had stable seasonality and was not affected by time.The number of patients attending the hospital in July and August each year was higher than the mean value,and that in February was lower than the mean value;the number of patients was stable in the other months.Conclusion The ARIMA model is effective with stable prediction results,and combined with the seasonal factors and long-term trend extracted by the X-11 method,it can better explain the rules of gout patients attending the hospital and thus provide a reference for health authorities and medical institutions to formulate the prevention and control policies for gout,conduct health education,and allocate human resources.