Making the hospital smart: using a deep long short-term memory model to predict hospital performance metrics
| DOI | https://doi.org/10.1108/IMDS-12-2021-0769 |
| Published date | 01 April 2022 |
| Date | 01 April 2022 |
| Pages | 2151-2174 |
| Subject Matter | Information & knowledge management,Information systems,Data management systems,Knowledge management,Knowledge sharing,Management science & operations,Supply chain management,Supply chain information systems,Logistics,Quality management/systems |
| Author | Qiong Jia,Ying Zhu,Rui Xu,Yubin Zhang,Yihua Zhao |
Making the hospital smart: using a
deep long short-term memory
model to predict hospital
performance metrics
Qiong Jia
Department of Management, Hohai Business School, Hohai University,
Nanjing, China
Ying Zhu
Faculty of Management, The University of British Columbia - Okanagan Campus,
Kelowna, Canada
Rui Xu
Department of Management Science and Engineering, Hohai Business School,
Hohai University, Nanjing, China
Yubin Zhang
Infrastructure Construction Office,
Jiangsu Provincial Maternal and Child Health Hospital,
Jiangsu Province Hospital and Nanjing Medical University First Affiliated Hospital,
Nanjing, China, and
Yihua Zhao
Administration Office, Jiangsu Provincial Maternal and Child Health Hospital,
Jiangsu Province Hospital and Nanjing Medical University First Affiliated Hospital,
Nanjing, China
Abstract
Purpose –Abundant studies of outpatient visits apply traditional recurrent neural network (RNN)
approaches; more recent methods, such as the deep long short-term memory (DLSTM) model, have yet to be
implemented in efforts to forecast key hospital data. Therefore, the current study aims to reports on an
application of the DLSTM model to forecast multiple streams of healthcare data.
Design/methodology/approach –As the most advanced machine learning (ML) method, static and
dynamic DLSTM models aim to forecast time-series data, such as daily patient visits. With a comparative
analysis conducted in a high-level, urban Chinese hospital, this study tests the proposed DLSTM model against
several widely used time-series analyses as reference models.
Findings –The empiricalresults show that the static DLSTM approachoutperforms seasonal autoregressive
integratedmoving averages (SARIMA),single and multipleRNN, deep gated recurrentunits (DGRU), traditional
long short-term memory (LSTM) and dynamic DLSTM, with smallermean absolute, root mean square, mean
absolutepercentage and root mean squarepercentage errors (RMSPE).In particular, static DLSTMoutperforms
all other modelsfor predicting daily patientvisits, the number of dailymedical examinations andprescriptions.
Practical implications –With these results, hospitals can achieve more precise predictions of outpatient
visits, medical examinations and prescriptions, which can inform hospitals’construction plans and increase the
efficiency with which the hospitals manage relevant information.
Application of
DLSTM model
with hospital
data
2151
Qiong Jia and Ying Zhu are contributed equally to this paper and are equal first authors. Their names are
listed in alphabetical order.
Funding: This research is supported by the Nanjing Soft Science Research Foundation under Grant
No. 2021SR00400030 and the National Natural Science Foundation of China under Grant No. 71702045.
The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/0263-5577.htm
Received 17 December 2021
Revised 15 February 2022
Accepted 1 March 2022
Industrial Management & Data
Systems
Vol. 122 No. 10, 2022
pp. 2151-2174
© Emerald Publishing Limited
0263-5577
DOI 10.1108/IMDS-12-2021-0769
Originality/value –To address a persistent gap in smart hospital and ML literature, this study offers
evidence of the best forecasting models with a comparative analysis. The study extends predictive methods for
forecasting patient visits, medical examinations and prescriptions and advances insights into smart hospitals
by testing a state-of-the-art, deep learning neural network method.
Keywords Smart hospital, Machine learning, Artificial intelligence, Deep learning, Hospital performance,
Neural network
Paper type Research paper
1. Introduction
Advances in technology, data science and ML techniques offer increasing options for smart city
initiatives (Joss et al., 2019;Yigitcanlar et al., 2019). In particular, smart hospitals promise to
improve, redesign and rebuild hospitals’clinical processes, infrastructure and management
systems by applying digital information and advanced data science techniques (Gomez-Sacristan
et al., 2015;Moro Viscontiet al., 2019). These complicated structures feature multiple smart systems
that can optimize and balance clinical processes and management systems, with the ultimate goal
of providing enhanced care to patients (Tripathi et al., 2020). As governments around the world
increase their expenditures on health care infrastructures (Iyengar et al., 2020;Jakovljevic et al.,
2019;Li et al., 2017)–notably, the smart hospital market is expected to grow to US$77.80bn by
2026, with a compound annual growth rate of 23.5% (Engine, 2021)–the opportunities for building
smart hospitals or transforming traditional hospitals into smart hospitals are vast.
Prior research into smart hospitals takes various perspectives on hospital operations (Al-Refaie
et al., 2017), information systems (Thakare and Khire, 2014), sustainability (Moro Visconti et al.,
2019), security and privacy (Chou et al., 2018;Sarosh et al., 2021) and patient routing decisions in
emergency departments (Ding et al., 2019). In addition, by leveraging expanded computing power,
scholars have applied different technologies to analyze health care data, including advanced data
analytics and artificial intelligence (Kim et al., 2021;Lam et al., 2021;Li et al., 2020;Rathi et al., 2021;
Wang and Fung, 2015), ML (Abdulkareem et al., 2021;Liu and Pu, 2019;Nasir et al., 2019)andbig
data analytics (Yang et al., 2021) and thus identify ways to improve the effectiveness of hospital
operations. Abdulkareem et al. (2021) examinehowthreeMLmodels[naiveBayes,randomforest
(RF) and support vector machine (SVM)] function in the Internet of Things, revealingh ow SVM can
improve the accuracy of coronavirus disease 2019 (COVID-19) diagnoses. In turn, the digital
solutions suggested by recent research provide options for enhancing clinical processes,
management systems and medical supplies, whether by leveraging remote health care (Taiwo and
Ezugwu, 2020), mobile health solutions through wearable devices (Kumar and Gandhi, 2018)or
radio-frequency identification for medical equipment (Nabelsi and Gagnon, 2017).
Building on this stream of cutting-edge research, the current study representsan effort to
help hospitals become smarter by improving their forecasting capability through advanced
deep ML algorithms.The overall goal of smart hospitalsis to improve performancemetrics, so
we seek a way to obtain better estimations of relevant health care metrics by applying
advanceddata analytics. For example,patient volume is a criticalperformance metric; it affects
patients’experience, hospitals’service qualityand the satisfaction of doctors,staff and visitors
(Bao et al.,2017;Naidu, 2009). Too many patients can create serious problems for hospital
managementand patients, such as delays in providingneeded care or receiving criticalcare, as
well as reduced quality. Even if the problems are less intense, high volumes impose time,
economic and psychological costs onpatients and family members who haveto wait for care
and perhapsmiss work, leading to more stress. Therefore,more accurate predictionsof patient
flow are essentialto hospitals’day-to-dayoperations and offer promise foroptimizing hospital
resource allocations, increasingoperational efficiency and improving service quality.
Substantial research has tried to predict patient volume, using methods that include generalized
linear models (Marcilio et al.,2013), time-series models such as autoregressive integrated moving
average (ARIMA) (Luo et al.,2017), SARIMA (ArunkKumar et al.,2021) and traditional neural
IMDS
122,10
2152
To continue reading
Request your trial