Predictive Modeling for Patient Queue Length in Blood Collection Centers during Peak Hours using Multi-step-ahead Forecasting and Machine Learning Models

機構典藏 > College of Business and Management > Graduate Institute & Department of Business Administration > Proceeding > Item 987654321/125903

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/125903

Title:	Predictive Modeling for Patient Queue Length in Blood Collection Centers during Peak Hours using Multi-step-ahead Forecasting and Machine Learning Models
Authors:	Chen, Ming-shu;Liu, Tzu-chi;Kao, Kuo-ching;Yang, Chih-te;Lu, Chi-jie
Keywords:	Blood collection center;Patient Queue Length;Multi-step ahead forecasting;Machine learning;Outpatient Phlebotom
Date:	2024-07-09
Issue Date:	2024-08-08 12:05:46 (UTC+8)
Abstract:	Blood collection centers in hospitals experience congestion during peak hours, leading to long waiting times for patients. This study investigates the application of machine learning to predict patient queue lengths in blood collection centers, aiming to minimize wait times and improve patient satisfaction. Existing literature explores various approaches to address congestion, including call systems, quality improvement initiatives, and phlebotomy assistant systems. However, these methods primarily focus on improving service efficiency, neglecting the challenge of predicting patient arrival patterns. Traditional queue length forecasting methods like simple moving average (SMA) have limitations. This study proposes a multi-step forecasting approach using machine learning techniques to predict patient queue lengths during peak times. The research employs two frameworks, Direct and Hybrid, incorporating six machine learning algorithms: Random Forest (RF), Extreme Gradient Boosting (XGBoost), Lasso Multiple Linear Regression (LaMLR), Multivariate Adaptive Regression Splines (MARS), Light Gradient Boosting Machine (LightGBM), and CatBoost. The study utilizes data from a medical center in Taiwan, covering a period of three years. Empirical results demonstrate that the Random Forest technique with the Direct framework achieves the most accurate predictions for one to four time steps ahead. For four-step-ahead forecasting, CatBoost with the Hybrid framework proves most effective. These findings suggest that machine learning offers a promising approach for predicting patient queue lengths in blood collection centers. This information can be valuable for staff scheduling, resource allocation, and implementing early congestion mitigation strategies, ultimately enhancing patient experience and healthcare service quality.
Appears in Collections:	[Graduate Institute & Department of Business Administration] Proceeding

Files in This Item:

There are no files associated with this item.

Loading...