Effective Forecasting of Key Features in Hospital Emergency Department

Paithane Pradip M* and Prajwal PS

doi:10.23880/art-16000113

Advances in Robotic Technology Research Article 3 min read

Effective Forecasting of Key Features in Hospital Emergency Department

Paithane Pradip M* and Prajwal PS^*

^* Corresponding author

ISSN: 2997-6197 10.23880/art-16000113 Received: April 11, 2024 Published: April 26, 2024

— views

13 references

1 table

PDF

Keywords

AI Convolutional Neural Networks Deep Reinforcement Learning Deep Belief Network Gated Recurrent Unit

Abstract

Efficiently forecasting the demands within a hospital’s Emergency Department (ED) is critical for optimal resource allocation and patient care management. This study focuses on leveraging deep learning techniques to predict various types of ED patient flows, facilitating informed decision-making by ED managers. The rising success of deep learning networks in modeling timeseries data makes them a compelling choice for patient flow forecasting. In this context, we investigate and compare seven deep learning models-Deep Belief Network (DBN), Restricted Boltzmann Machines (RBM), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), combined GRU and Convolutional Neural Networks (CNN-GRU), LSTM-CNN, and Generative Adversarial Network based on Recurrent Neural Networks (GAN-RNN)to accurately forecast patient flow within a hospital’semergency department. To enable traffic flow forecasting, a forecaster layer is introduced for each model. Real-world patient flow data spanning different ED services (biology, radiology, scanner, and echography) at Lille regional hospital in France serve as a case study to evaluate these models. Four effectiveness metrics are employed to assess and compare the forecasting methods. The outcomes demonstrate the superior performance of deep learning models in predicting ED patient flows compared to conventional shallow approaches like ridge regression and support vector regression. Significantly, the Deep Belief Network (DBN) stands out, achieving an averaged mean absolute percentage error of approximately 4.097.

Introduction

Over the last decades, there has been an expanding demand for emergency department (ED) cares, including medical and surgical treatments worldwide [1]. Efficiently managing healthcare systems can significantly improve resources management in emergency departments (EDs) at a hospital where the number of visits is unpredictable [2]. The successful management of EDs is particularly sensitive because they are expected to provide immediate and often lifesaving of patients. In the US, in the period from 1993 to 2003, the EDs visits considerably raised by an average of 26; however, the number of EDs lowered by approximately 9 [3]. For instance, the demand for EDs services has been doubled within the period 1990 and 2014, and it is still continuously increasing [4].

Related Work

In this study, the focus is on accurately forecasting daily visits in different departments of a Pediatric Emergency Department (PED). Such forecasts are crucial for optimizing appointment scheduling and resource allocation [5]. The study explores the potential of deep learning-driven methods in healthcare, particularly for predicting PED visits [6]. The authors examined seven deep learning models (including DBN, RBM, GRUCNN, LSTM-CNN, GAN-RNN, LSTM, and GRU) and two baseline models (SVR and RR) to forecast patient flow [7]. To prepare the data, multivariate PED visits data from various departments were preprocessed by smoothing and normalization [8]. The deep learning models were then trained on this data to predict daily PED visits at Lille Hospital in France. The results indicate that the DBN model outperformed the others, mainly due to its capacity for learning high-level features through hierarchical training [9]. While the deep learning models showed satisfactory results in patient flow forecasting, there is room for improvement [10]. One direction for enhancement is incorporating external factors like weather data and epidemic events to make the predictions more accurate [11]. Additionally, the study suggests designing an early detection system for overcrowding in Emergency Departments to enable proactive control, reducing the strain on resources and minimizing adverse consequences [12].

Sr No.	Authors	Title	Methods	Limitations
1	Yasser Zeinali and Seyed Taghi Akhavan Niaki [7]	Heart sound classification using signal processing and machine learning algorithms	TD3,FinBer	Complex Implementation, Data Dependency, Resource-Intensive
2	Yang, Liu, Zhong, and Walid [13]	Deep Reinforcement Learning for Automated Stock Trading: An Ensemble Strategy	Actor critic-based algorithms, Proximal Policy Optimization (PPO), Advantage Actor Critic (A2C),DDPG.	Limited period Evaluation, High Complexity, Uses limited data.
3	Vargas dos Anjos, Bichara, Evsukoff	Deep Learning for Stock Market Prediction Using Technical Indicators and Financial News Articles	Deep learning models, CNN, RNN, Technical Indicators	Involves significant data volume; however, historical stock market trends may not always correlate with future behavior.
4	Usha Manjunath,Mudunuri	Commodity and Forex trade automation using Deep Reinforcement Learning	Data collection, Data Pre-processing and Formatting, Training and testing of Agent.	Produce profits in short term only, Less effective in long trends.

Table 1: Comparison with Literature Data.

Declarations

Author Contributions

PMP: Conceptualization, Investigation, Writing-review & editing, Writing-review & editing, Supervision, Validation.

PPS: Conceptualization, Investigation, Writing-original draft, Writing-review & editing, Writing-original draft.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Ethical approval

Not applicable.

Availability of data and materials

Not applicable.

Funding

Not applicable.

Copyright

The Author(s) 2024.

References

Metev SM, Veiko VP (1998) Laser Assisted Microtechnology. 2nd(Edn.), In: Osgood (Ed.), Springer- Verlag, Berlin, Germany.
Breckling J (1989) The Analysis of Directional Time Series: Applications to Wind Speed and Direction, ser. Lecture Notes in Statistics. Berlin, Germany 61.
Zhang S, Zhu C, Sin JK, Mok PKT (1999) A novel ultrathin elevated channel low-temperature poly-Si TFT. IEEE Electron Device Lett 20: 569-571.
Wegmuller M, Weid JP, Oberson P, Gisin N (2000) High resolution fiber distributed measurements with coherent OFDR. Proc ECOC 11(3-4): 109.
Sorace RP, Reinhardt VS, Vaughn SA (1997) High-speed digitalto-RF converter. USA, 5: 668-842.
Zeinali, Yasser, Seyed TN (2022) Heart sound classification using signal processing and machine learning algorithms. Machine Learning with Applications 7(2022): 100206.
Kakarwal S, Paithane P (2022) Automatic pancreas segmentation using ResNet-18 deep learning approach. System research and information technologies 30(2): 1-13.
Paithane P, Kakarwal S (2023) LMNS-Net: Lightweight Multiscale Novel Semantic-Net deep learning approach used for automatic pancreas image segmentation in CT scan images. Expert Systems with Applications 30(234): 121064.
Wagh SJ, Paithane PM, Patil SN (2021) Applications of Fuzzy Logic in Assessment of Groundwater Quality Index from Jafrabad Taluka of Marathawada Region of Maharashtra State: A GIS Based Approach. InInternational Conference on Hybrid Intelligent Systems, pp: 354-364.
Paithane PM (2022) Yoga Posture Detection Using Machine Learning: Artificial Intelligence in Information and Communication Technologies. Healthcare and Education: A Roadmap Ahead, pp: 7.
Paithane P, Wagh SJ, Kakarwal S (2023) Optimization of route distance using k-NN algorithm for on-demand food delivery. System research and information technologies 30(1): 85-101.
Paithane PM (2023) Random forest algorithm use for crop recommendation. ITEGAM-JETIA 9(43): 34-41.
Hongyang Y, Liu XY, Zhong S, Walid A (2020) Deep reinforcement learning for automated stock trading: An ensemble strategy. Proceedings of the first ACM international conference on AI in finance.

← Previous Article Survey on Virtualization Technique and Its Role in Cloud Computing Environment Next Article → Heart Disease Prediction Using Multiple Machine Learning Algorithms

Effective Forecasting of Key Features in Hospital Emergency Department

Introduction

Related Work

Declarations

Author Contributions

Conflicts of Interest

Ethical approval

Consent to participate

Consent to publication

Availability of data and materials

Funding

References

Cite this article