基于ARIMA模型的医疗机构药品需求量预测研究——以他汀类药物为例

吴义来; 卢猛猛; 张烨雯; 周洲; 鲍泽响; 徐青文; 吴敏; 刘孟雪; 陶云松; 陈文刚; 张文; 邹建军; 栾家杰

doi:10.13748/j.cnki.issn1007-7693.20250068

基于ARIMA模型的医疗机构药品需求量预测研究——以他汀类药物为例

Research on Drug Demands Forecasting in Medical Institutions Based on an ARIMA Model: A Case Study of Statins

摘要

摘要:
目的构建自回归移动平均(autoregressive integrated moving average，ARIMA)模型预测他汀类药物需求量，为医疗机构制定药品采购计划，尤其是集采药品的科学报量提供可靠工具。
方法收集某三甲医院2022年1月—2024年6月每月他汀类药物的使用量数据，基于其使用频度(defined daily dose frequency，DDDs)建立时间序列，构建ARIMA模型。以该院2024年7月—12月的他汀类药物使用强度作为内部验证数据评估模型预测效果。收集区域内2所同级别医疗机构2022年1月—2024年12月他汀类药物使用量数据对模型进行外部验证。
结果医疗机构他汀类药物月度DDDs数据波动具有季节周期性和不规则变动等特征，构建乘积季节性ARIMA(1，1，0)(0，1，0)₁₂模型拟合序列，最小拟合BIC=439.1，残差序列通过Ljung-Box Q检验。内部验证和外部验证均提示模型预测值与实际DDDs相对误差较小，模型预测精度较高。
结论本研究构建的ARIMA模型对他汀类药物使用量预测效果良好，可为医疗机构药品采购及科学报量提供决策参考。

Abstract:
OBJECTIVE To construct an autoregressive integrated moving average(ARIMA) model for predicting statins demands, to provide medical institutions with a reliable tool to formulate drug procurement plans, especially for the scientific reporting of drug demands in centralized procurement.
METHODS Monthly data on statins usage from January 2022 to June 2024 were collected from a tertiary hospital. The defined daily dose frequency(DDDs) was calculated, and an ARIMA model was constructed. The statins usage intensity data of the hospital from July to December 2024 were used for internal validation to assess the model’s predictive performance. Additionally, data from two other tertiary hospitals from January 2022 to December 2024 were collected for external validation of the model.
RESULTS The monthly DDDs of statins in medical institutions exhibited seasonal periodicity and irregular fluctuations. A multiplicative seasonal ARIMA(1, 1, 0)(0, 1, 0)₁₂ model was constructed to fit the series, with a BIC value of 439.1. The residual series passed the Ljung-Box Q-test. Both internal and external validations indicated that the relative errors between the model's predicted values and the actual DDDs were acceptable, and the model had high prediction accuracy.
CONCLUSION The ARIMA model constructed in this study demonstrated good performance in predicting statin usage and can provide a reference for drug procurement and scientific reporting of drug demands in medical institutions.

HTML全文

参考文献(18)

施引文献

资源附件(0)