基于ARIMA模型的中国青光眼疾病负担预测研究

Research on Predicting the Burden of Glaucoma in China Basedon ARIMA Model

目的描述和分析1990—2019年中国青光眼疾病负担状况及其变化趋势,并预测2020—2050年中国青光眼疾病负担。方法从2019年全球疾病负担数据库中提取1990—2019年我国青光眼患病率、YLD率和DALYs率等疾病负担指标的数据,基于该数据建立自回归滑动平均模型(ARIMA模型)并预测2020—2050年我国青光眼疾病负担。结果1990—2019年,我国全年龄段青光眼的患病率从49.26/105增长至94.08/105,平均每年增长3.03%,且男性和女性青光眼患病率年均增长分别为2.63%和3.47%;我国全年龄段青光眼伤残所致生命年损失(YLD)率及伤残调整寿命年(DALY)率均由5.68/105提高至7.91/105,年均变化率为1.31%,男性和女性的青光眼YLD率及DALYs率均呈升高趋势,平均每年增长1.01%和1.69%。由ARIMA模型预测的青光眼疾病负担结果显示动态趋势与实际情况基本相同,且用于评估ARIMA模型的预测值与实际值的预测准确性指标:相对误差、平均绝对百分比误差(MAPE)、平均绝对误差(MAE)和均方根误差(RMSE)均很小,预测结果精确。经ARIMA模型预测得到2020年我国青光眼患病率为128.70/105,YLD率为10.63/105,DALYs率为10.63/105,到2050年,预计我国青光眼患病率为942.17/105,YLD率为10.87/105,DALYs率为10.87/105。结论ARIMA模型在拟合我国青光眼疾病负担应用中拟合效果和预测精度良好,可为青光眼疾病负担预测提供借鉴和参考,预计2020—2050年我国青光眼疾病负担呈现上升趋势。

Objective To describe and analyze the disease burden of glaucoma in China from 1990 to 2019 and its changing trend,and to predict the disease burden of glaucoma in China from 2020 to 2050.Methods In the data on the disease burden indicators of glaucoma in China from 1990 to 2019,including prevalence,YLD rate and DALYs rate,were extracted from the Global Burden of Disease Study 2019,and an autoregressive integrated moving average(ARIMA)model was established.Subsequently,the disease burden of glaucoma in China from 2020 to 2050 was predicted by using the ARIMA model.Results From 1990 to 2019,the prevalence of glaucoma in China across all age groups increased from 49.26/105 to 94.08/105,with an average annual growth rate of 3.03%.The annual growth rates of glaucoma prevalence in males and females were 2.63%and 3.47%,respectively.The years lost due to disability(YLD)rate and disability adjusted life year(DALYs)rate of glaucoma in China across all age groups increased from 5.68/105 to 7.91/105,with an average annual change rate of 1.31%.The YLD rate and DALYs rate of glaucoma in males and females showed an increasing trend,increasing by 1.01%and 1.69%per year on average,respectively.The results predicted by the ARIMA model showed that the dynamic trend was basically the same as the actual burden of glaucoma,and the relative error and mean absolute percentage error(MAPE),decomposition of the mean absolute error(MAE)and root mean square error(RMSE)were all very small,and the prediction results were accurate.The results of the study showed that after being predicted by the ARIMA model,the prevalence of glaucoma in China in 2020 was 128.70/105,the YLD rate was 10.63/105,and the DALYs rate was 10.63/105.Meanwhile,the prevalence of glaucoma in China in 2050 was predicted to be 942.17/105,with a YLD rate of 10.87/105 and a DALYs rate of 10.87/105.Conclusion This study shows that the ARIMA model has good fitting effect and accuracy in the application of simulating the disease burden of glaucoma in China,which can provide reference for the prediction of glaucoma disease burden.The results of this model show that the disease burden of glaucoma in China is on the rise from 2020 to 2050.