Enhancing Patient Experience in Radiology: Predictive Modeling of Wait Times using Feature Selection Techniques

Abstract

The increasing patient flow and overcrowding in critical hospital departments has prompted the need for effective strategies to enhance patient satisfaction. This study focuses on machine learning algorithms to predict patient waiting times for X-ray services using the dataset from a high-volume radiology department.  Three regression models such as Linear Regression (LR), K-Nearest Neighbor (KNN), and Random Forest (RF) were proposed and integrated with the recursive feature elimination (RFE) algorithm to reduce the dimension of the dataset and to enhance the model’s efficiency by selecting optimal features.

The findings indicate that LR-RFE model with 30 features predicted waiting time with mean absolute error 3.63 minutes as compared to standard LR model with 63 features. Comparable results were observed with the RF and KNN models, which demonstrated mean absolute errors of 3.77 minutes and 3.81 minutes respectively. Furthermore, the feature revealed key contributors to waiting times, such as the sum of patient queue wait times, the number of patients waiting in line, and wait time for the most recent patient This study underscores the potential of machine learning techniques combined with feature selection to offer actionable insights for better patient queue management.