An optimization protocol for MRI examination resource allocation based on demand forecasting and linear programming

被引：0

作者：

Zhongbin Zhou ^{[1
]}

Hanyu Zhou ^{[2
]}

Yuanyuan Qiao ^{[1
]}

Zihan Gao ^{[1
]}

Ying Yang ^{[1
]}

机构：

[1] The 6th Medical Center of PLA General Hospital, Beijing

[2] China Unicom Beijing Branch, Beijing

来源：

Scientific Reports | / 15卷 / 1期

关键词：

ARIMA model; Integer linear programming; MRI; NAR model; Optimization protocol;

D O I：

10.1038/s41598-025-98817-z

中图分类号：

学科分类号：

摘要：

The accessibility of medical services in Mainland China had been on the rise, leading to a surge in the number of Magnetic Resonance Imaging (MRI) scans. This increase had caused substantial delays in MRI examination queues at large hospitals. With MRI equipment and exams being costly, over-purchasing machines could lead to underutilization of resources. It was, therefore, crucial to devise a comprehensive method that could shorten patient wait times and optimize the use of medical resources within hospitals. The research had utilized daily MRI examination application data from a hospital covering the period from July 1, 2017, to November 30, 2022. The Autoregressive Integrated Moving Average (ARIMA) model and the AutoRegressive Integrated Moving Average with exogenous (ARIMAX) model were developed using SAS (version 9.3) software. Moreover, Non-AutoRegressive (NAR) and Non-AutoRegressive with exogenous (NARX) models were built using MATLAB (version R2015b) to forecast future MRI examination demands. Integrating the ARIMAX model with the NARX model, an ARIMAX-NARX model had been constructed.The predictive accuracy of these models was then assessed and compared. Based on the prediction outcomes, an Integer Linear Programming model was employed to calculate the optimal number of MRI examinations per machine per day, targeting cost reduction. An optimization flowchart for MRI examination resource allocation was developed by integrating critical process components, thus streamlining and systematizing the optimization process to improve efficiency. Analysis of the data revealed a weekly cyclical trend in MRI examination applications. Among the ARIMA, ARIMAX, NAR, NARX, ARIMAX-NARX models evaluated for their predictive skills, the NARX model emerged as the most accurate for forecasting. An Integer Linear Programming (ILP) model was utilized to plan the number of examinations for each MRI machine, effectively reducing costs. An optimization flowchart was developed to integrate key factors in MRI examination resource allocation, streamlining and systematizing the optimization process to enhance work efficiency. This study offers a comprehensive protocol for optimizing MRI examination resource allocation, combining the predictive power of the NARX model, the planning capabilities of the Integer Linear Programming model, and the integration of other relevant factors via an optimization flowchart. © The Author(s) 2025.

引用

共 28 条

[1]

Duwalage K.I., Burkett E., White G., Wong A., Thompson M.H., Forecasting daily counts of patient presentations in Australian emergency departments using statistical models with time-varying predictors, Emerg. Med. Australas, 32, pp. 618-625, (2020)

[2]

Jilani T., Et al., Short and long term predictions of hospital emergency department attendances, Int. J. Med. Inf, 129, pp. 167-174, (2019)

[3]

McAllan F.J., Egerton-Warburton D., O'Reilly G., Weiland T.J., Jelinek G.A., Planning for the future: modelling daily emergency departmentpresentations in an Australian capital City, Emerg. Med. Australas, 31, 5, pp. 750-755, (2019)

[4]

Marcilio I., Hajat S., Gouveia N., Forecasting daily emergency department visits using calendar variables and ambient temperature readings, Acad. Emerg. Med, 20, 8, pp. 769-777, (2013)

[5]

Boyle J., Et al., Predicting emergency department admissions, Emerg. Med. J, 29, 5, pp. 358-365, (2012)

[6]

Calegari R., Et al., Forecasting daily volume and acuity of patients in the emergency department, Comput. Math. Methods Med, (2016)

[7]

Zha W.T., Et al., Effects of meteorological factors on the incidence of mumps and models for prediction, BMC Infect. Dis.

[8]

Yang J., Et al., Predicting pulmonary tuberculosis incidence in China using Baidu search index: an ARIMAX model approach, Environ. Health Prev, (2023)

[9]

Zhang G.P., Time series forecasting using a hybrid ARIMA and neural network model, J. Neurocomputing, 50, pp. 159-175, (2003)

[10]

Fei Y., Li W.Q., Improve artificial neural network for medical analysis, diagnosis and prediction, J. Crit. Care, 40, (2017)

← 1 2 3 →