Sage Journals: Discover world-class research

Abstract

Objectives: This study proposes a novel architecture for designing digital twins in healthcare units. Methods: A systematic research methodology was employed to develop architecture design patterns. In particular, a systematic literature review was conducted using the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework to answer specific research questions and provide guidelines for designing the architecture. Subsequently, a case study was designed and analyzed at a chemotherapy treatment center for outpatients. Results: System architecture knowledge was distilled from this real-world case study, supplemented by existing software and systems design patterns. A novel five-layer architecture for digital twins in healthcare units was proposed with a focus on the security and privacy of patients’ information. Conclusion: The proposed digital twin architecture for healthcare units offers a comprehensive solution that provides modularity, scalability, security, and interoperability. The architecture provides a robust framework for effectively and efficiently managing healthcare environments.

Keywords

digital twin healthcare hospital architecture and systematic design

Introduction

During the COVID-19 outbreak, novel trends and patterns of digital technology have emerged in the healthcare industry, especially in the ease and accessibility of care. Consequently, worldwide healthcare systems, including Thailand, have recognized the importance of digital transformation to create new inventions or changes in the form of care and services in the digital age to increase the efficiency of the healthcare system.

One of the ideas for digital transformation is digital twin, which means a digital model that represents a situation or real-time of a system or process that collects all the information, and the model can keep up to date when new data is inputted. The digital twin consists of various building blocks. These building blocks, and underlying technologies, need to be implemented as part of digital transformation, combining people, processes, and technology. Any improvements in a healthcare unit center have some limitations in applying to the real service system because patients enter to receive services every day. Consequently, adjusting or improving the real system causes an impact on patient services. Therefore, the digital twin system can visualize patient and staff flow along the care process and evaluate the changes in factors and parameters without interfering with the real world. The benefits of the digital twin are used to test the scenarios of a system or process safely and cost-effectively before applying any changes in real systems or processes in the operating environment. The utilization or application of the digital twin in healthcare can be divided into six areas: diagnosis, monitoring, surgery, medical devices, drug development, and regulatory.

The role of the digital twin, as shown in Figure 1, represents the transformation of a service system in a healthcare unit. Digital twin creates a virtual version of the current healthcare service system through real-time capture. This virtual representation shows the current situation of service system performance. Thus, real-time data type was required for this role. Consequently, we can understand the current situation of the service system from the virtual representation. Subsequently, digital twin could be programmed to simulate the outcomes in various scenarios. This allowed decision-makers to change to future service systems based on the optimal solution obtained from the discrete event simulation model.

Figure 1.

Role of digital twin in transformation process.

In this paper, we focus on the design and development of a digital twin monitoring service system in a healthcare unit, and architecture is one of the important issues that determine the effectiveness and performance of a digital twin, for example, data management, real-time data processing, data security, and data privacy, etc. However, to our best knowledge, there is no study focused on the architecture design of digital twins in a healthcare unit. To prevent the lack of architecture that could be a barrier to successful digital twin implementation, the architecture design of digital twin in a healthcare unit is considered in this study. At the beginning of the study, we identified four specific research questions that direct our entire research work:

Research Question 1

what are the critical factors for developing a healthcare digital twin?

Research Question 2

what types of data should be included in a digital twin model?

Research Question 3

what is the healthcare digital twin architecture?

Research Question 4

How can the identified DT Architecture be connected and merged with current healthcare ICT systems such as EHR (Electronic Health Record), CDR (Clinical Data Repository)?

Hence, this work aims to investigate answers to the aforementioned specific research questions to be a guideline for designing the architecture of a digital twin in monitoring service systems in a healthcare unit. To accomplish the mission, we adopted systematic reviews and meta-analyses as a preliminary study to justify developing other healthcare digital twin projects. This allowed us to present results by combining and analysing research results from various studies conducted on similar research topics.

The manuscript’s organization has followed this structure. After the presentation of the research background in section 1 and the methodology used for the study in section 2, section 3 presents the answer to each research question. Next, a novel digital twin architecture for a healthcare unit is proposed in Section 4. Then, Section 5 presents the implementation of the digital twin architecture in a case study. Finally, Sections 6 and 7 provide the discussion and conclusion of this study, along with some suggestions for further research in this area.

Methodology of designing the architecture

A systematic research methodology^1,2 was employed to develop architecture design patterns for digital twin in monitoring service systems within a healthcare unit. Figure 2 illustrates the key steps involved in this process.

Figure 2.

Adopted research methodology.

Firstly, the research started with a domain analysis to develop key concepts based on a literature review. Since the concept of the digital twin was relatively new and consequently, it was used with different meanings. Therefore, the domain analysis reviewed existing literature on the usage of the digital twin in the context of monitoring service systems in a healthcare unit. Herein, the systematic literature review adopted the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement. PRISMA is a set of guidelines developed to improve the reporting quality of systematic reviews and meta-analyses in healthcare research. It was first published in 2009³ and has since been widely adopted as the standard for reporting these types of studies. The PRISMA guidelines have contributed significantly to improving the reporting quality of systematic reviews and meta-analyses, enhancing the transparency and reproducibility of research in the field of healthcare.

Secondly, after conducting the domain analysis process, we proceeded with architecture designing of digital twin in a healthcare unit. An investigating the digital twin architecture was the studies on existing design patterns. Hence, in this study, the existing design patterns will be explored and adapted to the proposed architecture of a digital twin in monitoring service systems in a healthcare unit.

Thirdly, we determined suitable areas for implementing the digital twin concept. The case study took place at HATYAI – NAMOM Cancer Centre, a cancer centre with the highest potential for treating patients in the lower southern region of Thailand. Consequently, the chemotherapy treatment centre for the outpatient group in HATYAI – NAMOM Cancer Centre was employed to be a case study. Following the case study, a thorough evaluation of outcomes ensues, aimed at gauging the efficacy and feasibility of implementing digital twin architecture in healthcare. This assessment meticulously considers how well the concept aligns with identified needs and challenges, with the overarching goal of validating proposed architecture design patterns.

Digital twins domain analysis

To develop key concepts of the digital twin, a systematic literature review was performed in the context of monitoring service systems in a healthcare unit. This systematic literature review adopted the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement.³ The PRISMA flow diagram of this work was shown in Figure 3. A total of 1109 records were obtained after executing the search strategies, and 8 records from subsequent snowballing. After duplicates were removed, 964 records remained. Of these, 137 were excluded because they might be not research articles since the words “Review”, “Survey” or “State of the art” existed in the title, abstract, or keywords. Since we focus on the application Digital Twin in a monitoring system in hospital, after reviewing the abstracts of, 683 were excluded from the remaining 827 records because they did not mention monitoring hospital systems in the abstract, leaving 144 articles for the final screening. 121 of these were excluded because they did not relate to our specific question in this study. In total, 23 articles were included in the analysis.

Figure 3.

PRISMA flow diagram.

The number of articles published on digital twin in healthcare has increased over time. Out of 23 articles analysed, 12 (52%) were published between 2019 and 2021. The analysis revealed that digital twin was used in 7 studies in the US and in collaborative research between the US and other countries in Europe and Asia, such as China, Germany, and Spain. There were 6 studies in Europe, 2 studies in Africa, specifically from Nigeria, and 8 studies in Asia, with collaborative research between Asia and other countries such as France, Australia, and Sweden. Among these studies, 14 were conducted in architecture designing, while the remaining studies focused on digital twin factors, data, and Healthcare Information and Communications Technology (ICT) systems.

Critical factors

To respond to Research Question 1, the critical factors for developing digital twin in a healthcare unit were analysed.⁴ developed an API model for digital twin and defined that the physical devices which monitored, engaged with, and possibly controlled were critical factors for digital twin. Hence, in the case of a healthcare unit, the physical entities involved in the service operation environment such as patients, staff, equipment, etc. In addition, several pieces of information existed in an operational environment of a healthcare unit, for example, patient’s symptoms, treatment time, equipment status, etc. Therefore, the critical factors for developing digital twin in a healthcare unit were physical entities and information existing within an operational environment.

Data

To respond to Research Question 2, the types of data which should be included in a digital twin model are discussed in this section. According to the complexity of healthcare related issues, a particularly large and multidimensional amount of data must be collected and organized for creating digital twin and keep them updated.⁵ Data related to Physical entities, time, treatment protocol, and the patient’s physiological status data should be collected in a database for the construction of digital twin in a healthcare unit, as shown in Figure 4.

Figure 4.

Data type of digital twin model in a healthcare unit.

Architecture

This section presents the architecture of digital twin in the healthcare unit centre to respond to Research Question 3. Since the digital twin is a key components in an Industrial IoT (Internet of Things) ecosystem, owned and managed by business stakeholders to provide secure storage, processing, and data sharing within an architectural tier.⁶ To our best knowledge, there was no architecture of a digital twin in monitoring service systems in a healthcare unit recently. Therefore, the requirements for designing the digital twin architecture were analyzed based on the works of literature on IoT architectures.^7–10

In the early stages of the research about IoT architecture design, the three-layer architecture was introduced. This architecture defined the basic minimal things needed for the devices to be connected to the internet. However, research always will be in search of finer things in architecture. Consequently, ecosystem tiers were increased to five-layer and then more than five layers.⁹ Figure 5 shows the architecture of IoT with different numbers of layers.

Figure 5.

Architecture of IoT with different number of layers.

According to Figure 5(a), the architecture consists of three layers: the perception layer, the network layer, and the application layer.^11–13 The perception layer was the physical layer, which had sensors for collecting data from devices about the operating environment and relaying it to the network layer. The network layer is in charge of connecting the application layer and perception layer together. This layer must support secure wired or wireless connections as well as other connectivity technologies. The primary role of the application layer was to provide specialized application services to users. It encompassed a wide range of applications in which the Internet of Things could be utilized, such as smart homes, smart cities, and smart health.

Figure 5(b) shows the five-layer architecture, which is the same as the three-layer with two additional layers. However, the additional layers would be different depending on the aims of the architectural design. In Figure 5(b1), the access gateway layer and middleware layer were added in classical three-layer architecture. The management of IoT communication in the specific environment and the transmission of messages between systems and objects fell under the purview of the access gateway layer. The relationship between physical devices and applications was more flexible because to the middleware layer.⁷ In Figure 5(b2),¹⁰ added remote servers layer and knowledge layer to IoT architecture for the smart hospital. The remote servers layer represented the remote computational technology of the IoT system. Besides, the knowledge layer processing was done. In Figure 5(b3), the perception layer (from the device) sent data to the application layer through the transport layer., while the processing layer stores, analyses, and processes for some definite pattern that can be made useful for business layer.^14,15 However, IoT has recently changed from several layers that were added to the architecture to presenting some specific designs. For example, a fog architecture⁸ inserted monitoring, pre-processing, storage, and security layers between the physical and transport layers as illustrate in Figure 5(c).

In addition, the studies related to general digital twin architecture were reviewed herein. In general, the fundamental principle of constructing a digital twin shares similarities with the foundational three-layer architecture of the Internet of Things (IoT). These aspects include the physical domain, the virtual domain, and the interconnection that facilitates the exchange of data and information between them.¹⁶ Similarly, within the industrial domain, there exists a comparable notion known as a Cyber-Physical System (CPS) or, more precisely, a Cyber-Physical Production System (CPPS).¹⁷ proposed a five-layer architecture for CPSs, defining a “Smart Connection Level”, “Data-to-Information Conversion Level”, “Cyber Level”, “Cognition Level”, and “Configuration Level”. At the Smart Connection Level, the focus lies on obtaining precise and dependable data from the physical entity. Subsequently, at the Data-to-Information Conversion Level, information is derived from the collected data. The Cyber Level acts as an information hub and the Cognition Level creates in-depth knowledge of the monitored system. Thereafter,¹⁸ proposed generic digital twin architecture (GDTA) for industrial energy systems. The five-layer architecture for CPSs inspired the GDTA and it was aligned with the information technology layers of the Reference Architecture Model Industry 4.0 (RAMI4.0)¹⁹ based on six IT layers; Business, Functional, Information, Communication, Integration, and Asset Level.

Digital twin and healthcare ICT systems

To respond to Research Question 4, this section explains how digital twins can work together with existing healthcare technology systems. Healthcare has already benefited greatly from using computers and technology, like electronic health records (EHRs) and clinical data repositories (CDRs). These tools have helped improve efficiency, reach, and the quality of care, while also reducing costs.²⁰ In healthcare, ICT (Information and Communication Technology) includes tools like e-health and health information technology. This covers areas such as telemedicine (remote healthcare services) and medical informatics (the management of medical data). Using the internet, ICT makes it possible for doctors to monitor patients remotely, track their vital signs in real-time, and offer treatment advice. It also allows easy sharing of information between healthcare professionals. Technologies like endoscopes and scanned images of tumours help surgeons by giving them clear views during surgery, making procedures more efficient and reducing complications.²¹

Connecting and merging digital twin architecture with current healthcare ICT systems involves a complex process that requires careful consideration which can be considered into two aspects.

Data integration aspect

Application programming interfaces (APIs)

To ensure sure the digital twin always has the latest patient information, we need to create connections between the digital twin platform and the electronic health records (EHRs) and clinical data repositories (CDRs). These connections are called application programming interfaces (APIs).

APIs are like bridges that let the digital twin communicate with the real-world healthcare systems. They help the digital twin get information from these systems and even control some things. A strong API is important to make sure the digital twin works well and is reliable.⁴ API development facilitates communication between systems, supporting data retrieval and push-back mechanisms.

Standardized data formats

To ensure the digital twin can work well with other healthcare systems, standard healthcare data formats like FHIR (Fast Healthcare Interoperability Resources) can be utilized.

FHIR is like a building block set that lets us create different healthcare systems. It uses “resources” to build these systems, which can be used for both patient care and administrative tasks. A digital twin that follows FHIR can use the resources provided by FHIR. For example, a digital twin of a patient can use the “Patient” resource to store information about the patient, like their name and address. It can also use the “Observation” resource to store information about the patient’s health, like their blood pressure or temperature. Similarly, a digital twin of a medical device can use the “Device” resource to track where the device is and what it's doing. It can also use the “Observation” resource to store information about the measurements the device has taken.

A digital twin can be thought of as a tool that combines different types of information from FHIR to show the current state of the real-world object it represents. This is like a shadow that follows the object. As shown in Figure 6, the digital twin layer sits above the healthcare system layer. It uses FHIR and other standards to work together. Not every part of a digital twin system in healthcare needs to use FHIR. Whether or not to use FHIR depends on if there are already standards for describing the information about the real-world object. These standards help build the information model for the digital twin.

Figure 6.

Layered representation integrating DTs, adopted from.²²

Data warehouse and database

In healthcare, a data warehouse is like a central storage place for all kinds of data, which helps create digital twins. Digital twins are virtual copies of real-world entities, such as patients, medical devices, or entire healthcare systems. They are used to monitor, analyze, predict, and improve healthcare outcomes. A database also stores data but focuses on specific aspects like patient health, medical records, devices, and other operational information. For example, an electronic health record (EHR) system is a database that contains patient information, medical history, diagnoses, medications, and treatment plans.

Databases are designed to handle individual records quickly and efficiently, making them ideal for real-time or near-real-time updates. There are two main types of databases for digital twins: standalone databases and organized databases. Standalone databases store and access data needed to build digital twins, while organized databases hold various types of information, such as patient data, treatment plans, medication details, and financial records. To ensure patient privacy, organized databases often require information management systems or middleware. Standalone databases may offer quicker access to data but may involve less emphasis on data security.

Technical aspect

Security and privacy

To ensure the digital twin works well with electronic health records (EHRs) and clinical data repositories (CDRs), we need to adjust the data, so they are compatible. Strong security measures are also necessary to protect patient information and comply with regulations like HIPAA (Health Insurance Portability and Accountability Act). The digital twin should have access to the latest patient information in real time. Careful testing, proper deployment, and continuous monitoring are needed to ensure accuracy, consistency, and security.

Scalability

Since, a digital twin was built on the physical world, scalability can be categorized into three categories: units, systems, and systems of systems (SoS)²³ as shown in Figure 7. The “unit level” refers to the smallest part involved in healthcare activities, such as a single piece of equipment. The “system level” connects multiple unit levels, allowing data and resources to be shared. At the “System of Systems (SoS) level,” different systems collaborate on a smart service platform, enabling them to connect, share information, and optimize their collective performance.²⁴

Figure 7.

Hierarchical levels of physical entities in healthcare system (adopted form²⁵).

Interoperability

To make digital twins work well with other healthcare systems, we need to ensure they can communicate with each other, a process called interoperability. This can be achieved by using standards, such as Health Level Seven (HL7), International Organization for Standardization (ISO), Institute of Electrical and Electronics Engineers (IEEE), Logical Observation Identifiers Names and Codes (LOINC), as well as entities like the World Wide Web Consortium (W3C), Internet Engineering Task Force (IETF), and Organization for the Advancement of Structured Information Standards (OASIS).

Digital twins for personalized medicine highlight the importance of connecting digital twins with existing healthcare ICT (Information and Communication Technology) systems. Personalized medicine requires a lot of detailed data about individual patients, which is analyzed with various drugs to find the best treatment. For example, the Swedish Digital Twin Consortium (SDTC) is focused on personalized medicine by creating digital twins of patients. These digital twins represent the patient’s molecular, physical, and environmental factors related to their illness. Thousands of drugs are then tested on the digital twin to find the most effective treatment.

In medicine and public health, digital twin technology has the potential to revolutionize traditional electronic health records (EHRs), which mainly focus on individual cases and large population groups. This transformation aims to prepare healthcare systems for a new era of precision medicine. However, there are challenges with using digital twins in healthcare, such as designing clear data displays, making data easy to access, and smoothly integrating them into clinical workflows.²⁶ Overall, to successfully connect digital twin technology with healthcare systems like Electronic Health Records (EHR) and Clinical Data Repositories (CDR), careful planning is needed to make sure everything works together smoothly.The process begins with identifying integration points and standardizing data formats using industry standards like HL7.

Design of digital twin architecture

In this section, a novel architecture of digital twin for a healthcare unit was proposed by building on or resembling existing design patterns. Rather than industrial IoT architectures, patients’ right was the important issue that we needed to consider in developing the digital twin architecture for the healthcare unit. Therefore, to ensure the security and privacy of patient information, a five-layer architecture is proposed, with each layer requiring specific security measures, encryption protocols, and access control policies. The proposed architecture was adopted from Generic Digital Twin Architecture (GDTA) model by Steindl et al in 2020¹⁸ as shown in Figure 8.

Figure 8.

Generic digital twin architecture in a healthcare unit.

The details in each layer of the proposed architecture were described as following.

(i) The physical layer: the physical representation of the digital twin was located at the physical layer and corresponded to the physical entity in a healthcare unit. It holed the physical parts of the digital twin in a healthcare unit including all physical data such as patients, staff, equipment, building, asset, etc. At the physical layer, security begins with controlling access to the physical components of the digital twin, such as patients, staff, and equipment. Biometric authentication, such as fingerprint or facial recognition, ensures that only authorized personnel can access sensitive systems and areas. This method enhances security by uniquely identifying individuals, thereby reducing the risk of unauthorized access. In addition to biometric systems, Radio-Frequency Identification (RFID) access cards provide a supplementary layer of security. These cards regulate physical access to critical infrastructure and equipment, helping to safeguard against unauthorized entry. Surveillance is another critical component of security within the physical layer. The strategic deployment of cameras and motion sensors allows for continuous monitoring of key areas, acting as both a deterrent to unauthorized access and a means of collecting evidence in case of security incidents. This real-time monitoring capability is indispensable in maintaining a secure environment.

(ii) The data conversion layer: At the integration level, there’s a distinction between Run-time Data and Engineering Data. Run-time Data are produced by devices to gather information about physical entities in the operational environment or Internet of Things (IoT). They reflect the current status of these entities and typically consist of time-series data. These data are highly dynamic, akin to big data processing or real-time responses. Consequently, inputting such data into systems can be done manually or automatically, depending on the device type. For instance, data can be manually entered into a computer or automatically captured by sensors or barcode readers. In contrast, Engineering Data are typically static, meaning they don’t change frequently over time. Examples include plant layouts, details about specific components in a hospital, treatment protocol, etc. This information often exists in Analog formats such as floor plans and diagrams, requiring digitization. Even if this information is already in a machine-readable format, it needs to be converted and integrated into the digital twin as contextual data. Therefore, a network data conversion layer is utilized to transmit data from devices to the hospital’s database. The selection of connection technologies for implementation at this layer, such as wired or wireless secure connections, must be carefully considered. The data conversion layer is critical for transmitting real-time data from devices to the hospital’s database. To protect this data, encryption plays a vital role in safeguarding sensitive information as it moves through the network. Advanced Encryption Standard (AES-256) is widely used due to its robustness, effectively protecting data from unauthorized access. This encryption standard ensures that even if data is intercepted, it remains indecipherable without the appropriate decryption keys. Additionally, secure communication protocols such as Hypertext Transfer Protocol Secure (HTTPS) and Message Queuing Telemetry Transport (MQTT) over Transport Layer Security (TLS) are employed. These protocols encrypt data transmissions between Internet of Things (IoT) devices and central servers, ensuring both confidentiality and authenticity, thereby preventing data breaches and unauthorized data manipulation.

(iii) The information layer: In the proposed architecture’s information layer, data is gathered and enhanced with semantics, establishing connections with other contextual information. A crucial component in this layer is the database, responsible for storing contextual information regarding resources and services. Since unrelated data or confidential data should be protected. Herein, 2 types of databases were considered as the standalone unit of database and organized database. The database’s information access is facilitated through an information query service, which is responsible for retrieving data from the database and delivering it to other services. On the other hand, the information insertion service is utilized to add or modify information within the upper layers of the proposed architecture. Additionally, beyond the functional layer, information pertaining to these services becomes part of the Shared Knowledge in the Information Layer. This information will be inserted and queried through the offered Smart Data Service interface. Integration with existing healthcare ICT systems occurs within this layer, necessitating data integration and standardization as crucial steps. Database security is paramount within the information layer, where data is gathered and contextualized. Access control policies are essential for maintaining data security within the information layer of digital twin architectures. Role-Based Access Control (RBAC) is a key policy that restricts data access based on user roles within the organization. This approach minimizes the risk of data exposure by ensuring that users only access the information necessary for their specific job functions, effectively mitigating the potential impact of security breaches. Multi-Factor Authentication (MFA) is another critical component of access control, providing an additional layer of security. MFA requires users to provide multiple forms of verification before accessing sensitive systems, such as a combination of something they know (password), something they have (security token), or something they are (biometric data). This layered approach significantly enhances security by making unauthorized access considerably more difficult. Continuous monitoring and regular audits are crucial for maintaining the integrity of digital twin systems. These processes promptly enable healthcare organizations to detect unauthorized activities and identify potential vulnerabilities. By continuously evaluating security measures and system access, organizations can respond swiftly to emerging threats and maintain compliance with regulatory standards. Compliance with healthcare standards such as the Health Insurance Portability and Accountability Act (HIPAA), Health Level Seven (HL7), and the International Organization for Standardization (ISO) is paramount in safeguarding patient information and ensuring data handling practices meet stringent regulatory requirements. Adhering to these standards not only protects sensitive data but also fosters trust among patients and stakeholders.

(iv) The functional layer: as the virtual essence of the digital twin concept, in the functional layer, simulation and decision support services played a vital role and served as the core services. Simulation was responsible for delivering a digital twin in monitoring decision support service in a healthcare unit to represent its virtual world with both physical entities and critical information. While decision support can be categorized based on their functionalities, which often complement each other. These categories include monitoring services, diagnosis services, prediction services, management services, and reconfiguration services.

Following are examples of decision support services in the literature. In monitoring and refiguration, there were studies that used simulation software named FlexSim HealthCare to build a digital twin for monitoring patients’ pathways in hospitals.^27–29 It models rich solutions for hospitals, clinics, and care facilities and their care processes. The benefits of a digital twin in monitoring a healthcare service system emerged in two aspects both near real-time monitoring system and a decision support tool.^29,30 present these two aspects in their research on the digital hospital. Commonly, near real-time monitoring was used to detect unexpected events from the current state of the real world. Besides, in a refiguration service, the discrete event simulation model was used to determine the acceptable scenario to adopt in the real healthcare units with various details such as manpower, risks, finances, etc. This enhanced the visibility of implications before decision-makers put plans into action. Noteworthy, the discrete event simulation model runs based on historical data gathered from the real-time monitoring system.

In diagnostic and prediction, the application of digital twin technology extends to personalized medicine, exemplified by the concept’s role in simulation. For instance, the Swedish Digital Twin Consortium (SDTC) is actively pursuing the development of a personalized medicine strategy.³¹ In this context, consider an individual patient presenting with localized disease manifestations. Through computational modeling, a digital twin representing this patient is generated, replicable in countless iterations, utilizing data from thousands of disease-relevant variables. Each digital twin undergoes computational treatment with various drugs from a vast selection. This process yields a digital prognosis tailored to the specific patient. Subsequently, the drug demonstrating the most favorable outcome on the digital twin is identified for the patient’s treatment regimen. Furthermore, a diverse array of digital twin types can be envisioned.²⁶ These range from comprehensive models of the entire human body to more focused representations, such as individual body systems or functions like digestion. Digital twins can also be finely tuned to specific organs like the liver or delve into even finer levels of biological complexity, encompassing cellular, subcellular (organelle/sub-organelle), or molecular levels. Additionally, digital twins are versatile tools that can be customized to mirror specific diseases or disorders. This includes instances such as simulating a diseased organ like a liver affected by non-alcoholic fatty liver disease. Furthermore, the scope of digital twin applications extends beyond human physiology to include relevant organisms, such as viruses, interacting with the aforementioned human digital twin models.

In monitoring, prediction, and management,^22,32 introduced the concept of a trauma digital twin, which offers substantial advancements in assisting medical professionals in trauma care. Their research proposes several key elements. Initially, they propose a platform for real-time monitoring of trauma management procedures. Following this, they suggest consolidating data from diverse sources related to specific traumas into a unified dataset. This facilitates more comprehensive data analysis. Lastly, they present a framework for conducting and overseeing trauma management simulations. In addition,³³ introduced digital twins for operating room management. A prototype of a digital twin-based solution to monitor and support operating room management by integrating it with the current legacy system was implemented. This solution offers a new dashboard that enables visualization of data about operating rooms and their planned, past, and ongoing surgical procedures. Users can inspect the timestamps of each step, correlated with warnings about data inconsistency or possible mistakes generated by the surgery digital twins. Additionally, notes about these warnings are included to justify any identified mistakes.

Recently, significant research has focused on the use of AI and machine learning in Digital Twins within healthcare. For example,³⁴ proposed a deep learning-powered clinical big data analytics approach for healthcare digital twins. In their digital twin model, data is collected both at the client and server levels. Using an improved Random Forest algorithm, they performed demand analysis and designed target functions for the medical and healthcare system. In terms of diagnosis,³⁵ introduced a framework on digital twinning for cancer patients, leveraging machine learning. As the patient undergoes treatment, data is continuously collected and analysed, allowing the digital twin to use several machine learning techniques to evaluate treatment progress or the extent of the disease. This framework demonstrates the potential of combining digital twinning and machine learning to revolutionize cancer diagnosis and treatment. For prediction,³⁶ presented disease prediction models utilizing a novel hybrid machine learning algorithm that processes comprehensive medical data. This model enables the anticipation of illnesses even before symptoms appear. Additionally,³⁷ proposed a machine learning-based electrocardiogram (ECG) classifier model for cardiac diagnostics and early detection of problems. Their cardiac models predict conditions with exceptional accuracy across various scenarios. These findings highlight the potential of Digital Twins in healthcare to create intelligent, comprehensive, and scalable health systems that improve patient-physician communication. Moreover, machine learning has also been applied to validate healthcare applications. For instance,³⁸ presented a method for generating realistic synthetic datasets that are statistically equivalent to real clinical data, demonstrating the effectiveness of a Generative Adversarial Network (GAN)-based approach. Furthermore, generative artificial intelligence (GAI), a rapidly emerging technology, has garnered research interest for its potential to evolve alongside digital twins in healthcare. GAI leverages advanced AI algorithms to automatically create, manipulate, and modify valuable and diverse data. It has shown remarkable potential in digital twin data analysis, including classification,^39,40 segmentation,^41,42 abnormality detection^43,44 and prediction^45,46

The functional layer encompasses the simulation and decision support services, which are central to the digital twin’s operation. Secure access to these services is facilitated using virtual private networks (VPNs) and multi-factor authentication (MFA), ensuring that only authorized personnel can interact with the simulation and decision support systems. Regular audits and continuous monitoring are implemented to detect any unauthorized activities or anomalies, allowing for swift responses to potential security threats.

(v) The business layer: the business logic in this layer aligns with the organization’s business strategy and guides the utilization of digital twins to meet specific goals. While the design of the architecture is not directly influenced by business logic, it plays a crucial role through specific inputs that drive the behaviour and decision-making of the digital twin. At the business layer, security policies are developed and enforced to align with the organization’s strategic goals and compliance requirements, such as the Health Insurance Portability and Accountability Act (HIPAA). These policies guide the use of digital twins, ensuring that their operation supports the healthcare organization’s objectives while maintaining data privacy and security. Anomaly detection systems are also employed to monitor for deviations from expected behavior, which could indicate potential security breaches.

Overall, In the proposed architecture, the interaction between layers is structured to facilitate the operation of a digital twin within a healthcare unit. At the physical layer, the digital twin embodies the physical components of the healthcare environment. This data is then transmitted through the data conversion layer, where real-time data from devices and static engineering data are integrated into a coherent format suitable for digital twin modeling. The information layer enriches this data with semantics and contextual information, stored in databases, ensuring confidentiality and accessibility through information query and insertion services. Integration with existing healthcare systems occurs here, following standards like HL7 and ISO. The functional layer embodies the virtual essence of the digital twin, providing simulation and decision support services crucial for monitoring, diagnosis, prediction, and management in healthcare operations. Finally, the business layer aligns digital twin usage with organizational goals, influencing behavior and decision-making. This layered approach ensures a holistic and effective digital twin implementation in healthcare settings.

However, a fundamental aspect of digital twins is their dynamic bidirectional mapping. They transcend being merely one-way maps, digital reflections, or simulation models of the physical world in the digital domain. In the healthcare sector, ensuring the continuous operation of ICT systems, including feedback mechanisms between the digital and physical layers, is paramount. One approach involves real-time monitoring and analysis of data generated by the digital twin to identify potential issues or inefficiencies in the physical environment, such as providing a platform for online and real-time monitoring of ongoing trauma management sessions.²² This is the process of automatically sending data when an event occurs. When an event updates the database, the DT system acknowledges it immediately. This feedback is then back-propagated to the physical layer through automated alerts or notifications to relevant personnel, prompting timely intervention and corrective actions. Additionally, predictive maintenance algorithms can anticipate equipment failures or maintenance needs based on data trends, enabling proactive measures to prevent disruptions. Furthermore, integration between ICT systems and physical devices, coupled with robust cybersecurity measures, ensures reliable communication and data exchange between layers, minimizing the risk of malfunctions or interruptions. This holistic approach to feedback propagation safeguards continuous operation in the healthcare sector, supporting the uninterrupted delivery of critical services.

Finally, to emphasize the unique contributions of the proposed model, a comparison with existing architectures is presented. However, there is currently no established architecture for digital twins in healthcare monitoring service systems. Therefore, this study compares the proposed digital twin architecture with IoT architectures available in the literature, as shown in Table 1.

Table 1.

Comparison of the proposed digital twin architecture with existing IoT architectures.

Articles	Layers									Notable details
Articles	Physical	Pre-processing/ Data conversion	Access gateway	Network	Middleware	Remote service	Processing/ Knowledge	Application	Business	Notable details
8,15	✓			✓				✓		Basic three-layer architecture
8	✓		✓	✓	✓			✓		The five-layer architecture, additional 2 layers depending on specific objective
9	✓	✓		✓	✓			✓
15	✓			✓			✓	✓	✓
11	✓			✓		✓	✓	✓
12	✓			✓	✓			✓	✓
16	✓			✓	✓			✓	✓
The proposed architecture	✓	✓	✓	✓	✓	✓	✓	✓	✓	Comprehensive aspects of architecture design
The proposed architecture			Group in information layer				Group in functional layer			Comprehensive aspects of architecture design

According to Table 1, the three-layer architecture introduced as the fundamental IoT model, consisting of the physical, network, and application layers. However, the five-layer architecture has become more prevalent in recent research. Researchers have introduced two additional layers to the basic model to achieve specific goals. Accordingly, the proposed digital twin architecture is structured as a five-layer model, with each layer addressing the relevant aspects of the system. One of the contributions of the proposed architecture is its comprehensive approach. Another contribution lies in the information layer of the proposed architecture, which integrates with the hospital’s current ICT system. The database issue was raised for the first time compared to existing architectures. Moreover, the knowledge-sharing component in the information layer reflects a design not only for a single hospital but also for hospitals within a healthcare unit center network.

Case study analysis

Problem description

In this study, the outpatient care of the chemotherapy department of HATYAI – NAMOM Cancer Centre in Thailand was used as a case study. Figure 9 illustrates the service system description of outpatients who taking chemotherapy. Firstly, patients arrive at the patient’s queue for registration by a nurse. Subsequently, a diagnosis from a doctor is provided. Then, a patient takes a chemotherapy drug order from a nurse and picks the drug from a pharmacist. Finally, the patient leaves the system upon completion of the treatment process.

Figure 9.

Service system description.

Data type

According to details of the case study in previous sections, we could identify four main data types as follows.

(1) Physical data

This data type included all physical entities in the operating environment, for example, patients, nurses, doctors, equipment, etc. In addition, the building was another physical data required for creating the floor plan in a digital twin model.

(2) Time data

In the context of a digital twin, time data plays a crucial role in synchronizing the virtual representation (the digital twin) with its real-world counterpart. The digital twin is a virtual model that mirrors the behaviour and characteristics of a physical object, system, or process. To generate time data for a digital twin, you can consider the two main aspects as real-time data and historical time data.

The digital twin should be synchronized with the real-world system it represents. This involves capturing real-time data from sensors, devices, or other sources in the physical system and updating the digital twin accordingly. The time stamps of the data collected from the real-world system can serve as the time data for the digital twin.

Besides, the value of historical data for training and validating the behaviour of the digital twin depends on its specific purpose. Historical time data can be collected from the real-world system or relevant sources and used to create and validate the digital twin’s model.

(3) Treatment Protocol

By integrating the treatment protocol into a digital twin, healthcare providers can benefit from enhanced decision support, personalized treatment plans, and continuous monitoring. This can lead to improved patient outcomes, reduced healthcare costs, and more efficient healthcare delivery.

(4) Patients Physiological Status Data

The digital twin can integrate various sources of patient data, including electronic health records (EHRs), medical devices, wearables, and real-time sensor data. This data can provide a comprehensive view of the patient’s condition, medical history, and treatment response. Consequently, the digital twin can continuously monitor the patient’s vital signs, biomarkers, and other relevant data in real time. It can compare the real-time data with the simulated treatment plan, identifying any deviations or potential issues. This enables early intervention and adjustment of the treatment protocol if necessary.

Digital twin architecture implementation

In this section, the proposed digital twin architecture was applied to the case study. Figure 10 illustrates the digital twin architecture in the chemotherapy department of HATYAI – NAMOM Cancer Centre.

Figure 10.

Digital twin architecture in chemotherapy department of HATYAI – NAMOM Cancer Centre.

Since the critical design factors for the digital twin in a healthcare unit are the physical entities and information within an operational environment, the initial step in the physical layer involves identifying entities such as patients, nurses, doctors, technicians, pharmacies, scales, blood pressure monitors, kiosks, desks, computers, chairs, and beds. The relationships between these physical entities and the information within the environmental conditions and constraints are established. Data is generated as a patient arrives at each station and receives the required services. Furthermore, data flows between workstations. For instance, patients are assigned to a queue at the first station and measured for weight, height, heart rate, and blood pressure. This data flows from station to station until the patient completes treatment.

Subsequently, all physical entities and necessary data in the physical layer can be transformed into digital replicas in the functional layer through the data conversion layer and information layer by employing IoT and standalone databases, respectively. In the data conversion and information layer, a standalone database is applied, as it can effectively provide the necessary data with minimal risks regarding data security. Subsequently, the necessary data is input into the database and connected via a local Wi-Fi network, adhering to data security and privacy standards.

Then, digital replication can be created using simulation software in the functional layer. In this layer, feedback mechanisms in digital twins are essential for continuous improvement in healthcare processes. Real-time data collection from various sources ensures the digital twin has up-to-date information for accurate analysis and decision-making. Patient-reported outcomes and clinical data integration provide comprehensive insights into patient health and treatment efficacy. Consequently, the functional layer of healthcare digital twins utilizes various feedback mechanisms to meet specific goals such as queue management, doctor assignment, bed assignment, and symptom monitoring. Continuous monitoring and data analysis enable these mechanisms to be utilized for ongoing improvements, enhancing the efficiency, effectiveness, and quality of healthcare delivery. By employing real-time data, patient feedback, and advanced analytics, digital twins represent a transformative approach to modern healthcare.

Queue management in healthcare facilities involves overseeing patient flow from check-in to discharge. Digital twins utilize real-time data from patient check-ins, sensor data, and digital kiosks to monitor patient movement and wait times continuously. This data is used to predict bottlenecks and optimize patient routing. For example, if the digital twin detects a surge in patient volume at the diagnosis station, it can recommend opening additional consultation rooms or reallocating staff to high-demand areas. Historical data analysis helps identify inefficiencies and inform long-term process improvements and resource allocation strategies.

Additionally, the assignment of doctors to patients is another critical area where digital twins provide significant benefits. The digital twin continuously tracks doctors’ schedules, specialties, patient loads, and performance metrics. This information is used to assign doctors to patients based on availability, expertise, and patient needs, ensuring an optimal match and reducing wait times. Feedback on assignment efficiency, patient outcomes, and doctor performance is analysed to refine algorithms, improve matching accuracy, and enhance patient satisfaction. Over time, this leads to better utilization of medical staff and improved patient care.

Efficient bed assignment is vital for hospital operations, particularly in managing capacity and patient care. The digital twin integrates real-time data on bed occupancy, patient acuity levels, and discharge schedules. This information is used to allocate beds based on current and predicted availability, patient conditions, and care requirements. By ensuring that patients are assigned to appropriate beds, the digital twin helps optimize the use of hospital resources. Continuous monitoring of bed utilization patterns informs capacity planning and helps improve turnaround times for bed availability, leading to more efficient hospital operations.

Moreover, symptom monitoring is a key aspect of personalized healthcare enabled by digital twins. Wearable devices, health apps, and patient-reported data provide continuous updates on symptoms and health metrics. The digital twin analyses this data to detect early signs of health deterioration, triggering alerts for medical intervention and personalizing treatment plans. Aggregated symptom data helps identify common trends and effective interventions, guiding the development of best practices and improving predictive models. This proactive approach enhances patient outcomes and reduces the risk of complications.

This case study focuses on integrating the digital twin of the cancer center with the existing ICT systems of its host hospital which is valuable to the organizational strategy in the business layer. The following main areas and concrete examples can be illustrated as follows.

(1) Improvement in diagnosis and treatment service

Comprehensive healthcare services and patient monitoring systems can be developed by integrating the digital twin with EHR (Electronic Health Record) or CDR (Clinical Data Repository). Healthcare system monitoring such as queue management, and doctor and bed assignments are employed. For instance, when the digital twin detects a surge in patient volume at any station, it can recommend reallocating staff to high-demand areas of the healthcare unit.

(2) Precision medicine

Symptom monitoring functions are important for patients. This enables physicians at the host hospital to remotely observe patients at the cancer care center, track vital signs in real-time, and provide precision treatment advice. Additionally, this function assists nurses in monitoring patients’ vital signs during chemotherapy. Currently, nurses check vital signs every 15 minutes during chemotherapy; real-time tracking can reduce their workload and enable continuous detection of abnormalities, thereby facilitating timely medical intervention.

(3) Predictive healthcare service

Additionally, using the EHR or CDR helps in promptly assigning patients for chemotherapy at the healthcare facility. Treatment protocols stored in the hospital’s database guide decisions regarding patient transfers. For example, if the cancer care center closes at 4 pm and patient transfer takes 1 hour, then patients requiring less than 4 hours of treatment can be transferred if beds are available by 11 am. Conversely, patients needing more than 4 hours of treatment cannot be transferred.

(4) Telehealth care

Hospital congestion is a significant challenge that can lead to longer wait times, increased patient stress, and strained healthcare resources. The integration of digital twins in telehealth care provides a solution to this problem. A digital twin of a healthcare facility, such as a cancer center, continuously monitors the availability of chemotherapy beds. If the digital twin identifies many empty chemotherapy beds, it can suggest transferring patients from the overcrowded main hospital to the cancer branch center. This alleviates hospital congestion while ensuring patients receive the necessary care promptly. Therefore, telehealth care, supported by digital twins, enhances the overall patient experience by reducing wait times and ensuring timely access to care. Patients transferred to the cancer center for chemotherapy can receive focused, specialized treatment in a less congested environment, improving their satisfaction and health outcomes. In addition, physicians can closely monitor patients’ symptoms and advise care from the main distance hospital.

Discussion

In this study, we developed generic digital twin architecture in a healthcare unit inspired by the works of literature on IoT architectures and generic digital twin architecture for industrial energy systems. However, the general approach in the literature was a 3-layer architecture approach. The three-layer architecture may not be comprehensive enough for system design and modelling, as it may not adequately capture all the essential components of the system. It might overlook certain aspects or dependencies that were crucial for a complete understanding of the system’s architecture. In addition, the three-layer architecture did not give a reliable solution⁴⁷ and hided details related to functionality and information flow.^48,49

The proposed generic digital twin architecture in a healthcare unit was based on a five-layer concept, offering a higher level of abstraction compared to a three-layer architecture. This increased abstraction provided a more conceptual view of the system, allowing for a broader understanding of its components and interactions. Nevertheless, the proposed architecture retained its usefulness in identifying crucial components and functionality for a digital twin in a healthcare unit. It successfully outlined the basic structure and components necessary for such a digital twin without imposing any specific technological constraints or dependencies. This approach allowed for flexibility and adaptability to various technologies, ensuring that the architecture could be implemented using different technological solutions based on specific requirements and preferences.

Specifically, the information layer considered two types of databases: standalone units and organized databases with information management. The inclusion of these database types aims to ensure the security and privacy of patient information. By utilizing a standalone unit or implementing efficient information management practices for an organized database, the architecture strives to establish a robust framework that safeguarded and ensured the confidentiality of patient data within the system.

In following a software engineering approach, the architecture adopts a modular design, enabling independent development and scalability of each layer. This design philosophy promotes easier maintenance and facilitates future enhancements. Moreover, the architecture prioritizes security and privacy, incorporating robust measures such as encryption, access controls, and secure data transmission protocols. Compliance with healthcare regulations, including HIPAA (Health Insurance Portability and Accountability Act), ensures the confidentiality of patient data. Additionally, the architecture adheres to industry standards like HL7 (Health Level Seven) and ISO (Organization for Standardization), ensuring seamless integration with existing healthcare systems and fostering interoperability across diverse environments. It employs CI/CD (Continuous Integration and Continuous Deployment) pipelines for automated testing, deployment, and monitoring, thereby enhancing the reliability and stability of the digital twin architecture. Furthermore, the architecture integrates feedback mechanisms between the digital and physical layers, enabling proactive maintenance and efficient issue resolution. Advanced predictive analytics algorithms anticipate equipment failures, enabling preemptive actions to prevent disruptions in healthcare operations.

This study provides a concrete example of how digital twin architecture can be applied through a case study. The findings from the digital twin case study in the HATYAI–NAMOM Cancer Centre’s chemotherapy department demonstrate how this technology can be generalized to other healthcare settings by leveraging real-time data collection, predictive analytics, and process optimization. Digital twins can optimize operations in emergency departments, outpatient clinics, and other medical environments by monitoring patient inflow, resource allocation, and patient progress in real-time. The system also enables personalized and precision medicine through continuous symptom monitoring and tailored treatment plans, making it ideal for managing chronic diseases or rehabilitation. Predictive capabilities improve resource management in hospital networks, surgical scheduling, and bed capacity planning. Telehealth and remote patient monitoring, particularly in rural areas and home care, can be enhanced by digital twins, offering timely medical interventions while reducing hospital visits. The technology also improves queue management and workflows in primary care and diagnostic centers by analyzing patient loads and dynamically adjusting staff or resources. Feedback loops for continuous improvement benefit areas such as surgical outcomes, pharmacy management, and overall healthcare processes. Additionally, enhanced data security and privacy, as shown in the case study, ensure that digital twins can handle sensitive patient data across various healthcare settings. Overall, the adaptable framework of digital twins offers a scalable solution for optimizing patient care, resource allocation, and operational efficiency across multiple healthcare environments.

Research limitations

Computational resources

Effective handling of real-time data is essential for digital twins to provide accurate simulations and decision support. This requirement can strain environments with limited computational resources or high data volumes, potentially impacting performance. Leveraging cloud-based platforms or edge computing can enhance processing capabilities, distributing the computational load and reducing latency for timely data analysis.

Scalability

Consequently, scalability is another concern, as the architecture must handle increasing data volumes without degrading system performance. Designing with modular components allows for scalable expansion, while cloud infrastructure enables dynamic resource allocation based on demand, ensuring continuous and efficient operation.

Challenges and future research directions

Implementing the proposed five layers of digital twin architecture in healthcare, particularly in a chemotherapy department, presents unique challenges that require strategic solutions as follows.

Integrating data

A primary challenge is integrating data from diverse sources like IoT devices, Electronic Health Records (EHR), and Clinical Data Repositories (CDR). The varying data formats and standards make seamless data flow and interoperability critical yet challenging. Adopting standardized data formats and communication protocols such as HL7 (Health Level Seven) and FHIR (Fast Healthcare Interoperability Resources) can mitigate these challenges.

Data privacy and security

Middleware solutions further facilitate data harmonization across multiple sources, ensuring effective system communication. Additionally, ensuring data security and privacy is paramount due to the prevalence of cyber threats; the architecture must comply with strict regulations, encrypting data to prevent unauthorized access.

Continuity of healthcare service

In addition to privacy and security, a comprehensive backup system is critical to maintain continuity of service in case of system failures. Healthcare operations cannot afford downtime, as patient care must continue uninterrupted. Therefore, incorporating redundant systems and automated failover mechanisms can ensure that digital twins remain operational even during unexpected disruptions.

User adoption

User adoption and training present challenges; healthcare professionals may encounter difficulties in adopting new digital twin technologies, with resistance to change or unfamiliarity potentially hindering implementation. Comprehensive training programs and user-friendly interfaces can facilitate user adoption. Ongoing support and feedback mechanisms help address concerns, improve user experience, and promote engagement with the technology.

Evolving with technological advancements

In the functional layer of digital twin architecture within a healthcare unit, decision support serves as the core service. These decision support systems can be categorized based on their functionalities, including monitoring, diagnosis, prediction, management, and reconfiguration services.

However, successfully establishing these decision support services requires high-fidelity virtual modeling and robust information exchange, which can be challenging when faced with scarce, biased, or noisy data. Consequently, integrating advancements in technologies such as AI and machine learning into the functional layer presents a significant challenge. This synergy has the potential to bridge the gap between digital twin technology and advanced healthcare solutions, pushing the boundaries of what is possible in the digital age.

Computer system validation

Finally, the biggest challenge is Computer System Validation (CSV) if the digital twin is intended for use in diagnostics or treatment recipes. Risk assessment, validation plans, and regulatory compliance will be critical paths to success.

By focusing on these areas, healthcare institutions can effectively integrate digital twins, optimizing resources and enhancing patient care in chemotherapy departments and beyond.

Conclusion

Nowadays, healthcare systems around the world including Thailand have recognized the importance of digital transformation to create invention or changes in the form of care and services in the digital age to increase the efficiency of the healthcare system. This study focuses on applying digital twins in monitoring service systems in a healthcare unit. To prevent the lack of architecture that could be a barrier to successful digital twin implementation, the architecture designing of the digital twin in a healthcare unit is considered.

We employed a methodical research approach to deduce the architecture design patterns. In the first step, the systematic literature review (PRISMA) was adopted to investigate and answer the specific research questions to be a guideline for designing the architecture. Subsequently, the case study of outpatient care at the chemotherapy centre in HATYAI – NAMOM Cancer Centre in Thailand in which digital twin was applied. Finally, a novel architecture of digital twin for a healthcare unit was proposed by building on existing design patterns.

With the aim of ensuring the security and privacy of patient information, a five-layer architecture was proposed. The 1^st layer, known as the physical layer, represented the physical entities. The 2^nd layer, the data conversion layer, was established to transmit data from devices to the hospital’s database. The 3^rd layer, the information layer, gathered and enriched the data with semantics and related contextual information. The 4^th layer, the function layer, identified service categories such as simulation, real-time monitoring, and decision support. Lastly, the 5^th layer, the business layer, aligned with the organization’s business strategy and guided the utilization of digital twins to achieve specific goals. In addition, specific security measures, encryption protocols, and access control policies, which is essential for safeguarding sensitive healthcare data were implemented across all layers of the digital twin architecture. These measures not only protect the integrity and confidentiality of patient information but also ensure the secure and effective operation of digital twins within healthcare environments.

In conclusion, the architecture comprises five layers, each serving distinct functions while ensuring seamless integration and data flow. The proposed digital twin architecture for healthcare units offers a comprehensive solution that prioritizes patient rights, security, and privacy. By adopting a software engineering approach focused on modularity, scalability, security, and interoperability, the architecture provides a robust framework for effectively and efficiently managing healthcare environments. Continuous feedback mechanisms and predictive maintenance ensure the uninterrupted operation of ICT systems, supporting the delivery of critical services in the healthcare sector.

Footnotes

Author contribution

The first author contributed to manuscript preparation,conceptualization,data collection,and analysis. The corresponding author handled manuscript review,revision,and communication with the journal. Co-authors contributed to manuscript review and revision.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This work (Grant No. RGNS 65-228) was supported by Office of the Permanent Secretary,Ministry of Higher Education,Science,Research and Innovation (OPS MHESI),Thailand Science Research and Innovation (TSRI) and Prince of Songkla University.

Ethical statement

ORCID iD

Dollaya Buakum

References

Verdouw

Tekinerdogan

Beulens

, et al. Digital twins in smart farming. Agric Syst 2021; 189: 103046.

Tekinerdogan

Verdouw

. Systems architecture design pattern catalog for developing digital twins. Sensors 2020; 20: 5103.

Moher

Liberati

Tetzlaff

, et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 2009; 151: 264–269.

Scheibmeir

Malaiya

. An API development model for digital twins. In 2019 IEEE 19th International Conference on Software Quality, Reliability and Security Companion (QRS-C). IEEE, 2019, pp. 518–519.

Voigt

Inojosa

Dillenseger

, et al. Digital twins for multiple sclerosis. Front Immunol 2021; 12: 669811.

Harper

Ganz

Malakuti

. Digital twin architecture and standards. IIC Journal of Innovation 2019; 12: 72–83.

Al-Qaseemi

Almulhim

, et al. IoT architecture challenges and issues: lack of standardization. In 2016 Future technologies conference (FTC). IEEE, 2016, pp. 731–738.

Sethi

Sarangi

. Internet of things: architectures, protocols, and applications. Journal of Electrical and Computer Engineering 2017; 2017.

Banu

Sujatha

. IoT architecture a comparative study. Int J Pure Appl Math 2017; 117: 45–49.

10.

Uslu

BÇ

Okay

Dursun

. Analysis of factors affecting IoT-based smart hospital design. J Cloud Comput 2020; 9: 1–23.

11.

Mashal

Alsaryrah

Chung

T-Y

, et al. Choices for interaction with things on Internet and underlying issues. Ad Hoc Netw 2015; 28: 68–90.

12.

Said

Masud

. Towards internet of things: survey and future vision. Int J Comput Network 2013; 5: 1–17.

13.

Wang

Daneshmand

Dohler

, et al. Guest Editorial-Special issue on internet of things (IoT): architecture, protocols and services. IEEE Sensor J 2013; 13: 3505–3510.

14.

T-J

Ling

F-Y

, et al. Research on the architecture of internet of things. In 2010 3rd international conference on advanced computer theory and engineering (ICACTE). IEEE, 2010, pp. 485–487.

15.

Khan

Zaheer

, et al. Future internet: the internet of things architecture, possible applications and key challenges. In 2012 10th international conference on frontiers of information technology. IEEE, 2012, pp. 257–260.

16.

Grieves

. Digital twin: manufacturing excellence through virtual factory replication. A whitepaper by dr. Michael grieves. White Paper, 2014, pp. 1–7.

17.

Lee

Bagheri

Kao

H-A

. A cyber-physical systems architecture for industry 4.0-based manufacturing systems. Manufacturing letters 2015; 3: 18–23.

18.

Steindl

Stagl

Kasper

, et al. Generic digital twin architecture for industrial energy systems. Appl Sci 2020; 10: 8903.

19.

Heidel

. Industrie 4.0: the reference architecture model RAMI 4.0 and the Industrie 4.0 component. Beuth Verlag GmbH, 2019.

20.

Omotosho

Ayegba

Emuoyibofarhe

, et al. Current state of ICT in healthcare delivery in developing countries. Nternational Journal of Online Engineering 2019; 15: 91–107.

21.

Adebayo

Bankole

Adebayo

. The impact and challenges of information and communication technology on health care delivery in Nigeria. Madonna University, Nigeria Faculty of Law Law Journal, 2021, vol 6.

22.

Ricci

Croatti

Montagna

. Pervasive and connected digital twins—a vision for digital health. IEEE Internet Computing 2021; 26: 26–32.

23.

Tao

, et al. Enabling technologies and tools for digital twin. J Manuf Syst 2021; 58: 3–21.

24.

Tao

Wang

, et al. Digital twins and cyber–physical systems toward smart manufacturing and industry 4.0: correlation and comparison. Engineering 2019; 5: 653–661.

25.

Zhao

Liao

, et al. Modeling of cyber-physical systems and digital twin based on edge computing, fog computing and cloud computing towards smart manufacturing. In International Manufacturing Science and Engineering Conference. American Society of Mechanical Engineers, 2018, p. V001T005A018.

26.

Kamel Boulos

Zhang

. Digital twins: from personalised medicine to precision public health. J Personalized Med 2021; 11: 745.

27.

Karakra

Fontanili

Lamine

, et al. Pervasive computing integrated discrete event simulation for a hospital digital twin. In 2018 IEEE/ACS 15th international conference on computer systems and Applications (AICCSA). IEEE, 2018, pp. 1–6.

28.

Karakra

Lamine

Fontanili

, et al. HospiT’Win: a digital twin framework for patients’ pathways real-time monitoring and hospital organizational resilience capacity enhancement. In International Workshop on Innovative Simulation for Healthcare. CAL-TEK Srl, 2020, pp. 62–71.

29.

Karakra

Fontanili

Lamine

, et al. A discrete event simulation-based methodology for building a digital twin of patient pathways in the hospital for near real-time monitoring and predictive simulation. Digital Twin 2022; 2: 1.

30.

Karakra

Fontanili

Lamine

, et al. HospiT'Win: a predictive simulation-based digital twin for patients pathways in hospital. In: 2019 IEEE EMBS international conference on biomedical & health informatics (BHI) 2019, pp.1-4. IEEE.

31.

Björnsson

Borrebaeck

Elander

, et al. Digital twins to personalize medicine. Genome Med 2020; 12: 1–4.

32.

Croatti

Gabellini

Montagna

, et al. On the integration of agents and digital twins in healthcare. J Med Syst 2020; 44: 161.

33.

Burattini

Montagna

Croatti

, et al. An ecosystem of digital twins for operating room management. In 2023 IEEE 36th International Symposium on Computer-Based Medical Systems (CBMS). IEEE, 2023, pp. 770–775.

34.

Guo

. Deep learning-empowered clinical big data analytics in healthcare digital twins. IEEE ACM Trans Comput Biol Bioinf 2023.

35.

Moztarzadeh

Jamshidi

Sargolzaei

, et al. Metaverse and healthcare: machine learning-enabled digital twins of cancer. Bioengineering 2023; 10: 455.

36.

Kulkarni

Quraishi

Raparthi

, et al. Hybrid disease prediction approach leveraging digital twin and metaverse technologies for health consumer. BMC Med Inf Decis Making 2024; 24: 92.

37.

Sharma

Kumar

. AI and digital twins transforming healthcare IoT. In 2024 14th International Conference on Cloud Computing, Data Science & Engineering (Confluence). IEEE, 2024, pp. 6–11.

38.

Arvanitis

White

Harrison

, et al. A method for machine learning generation of realistic synthetic datasets for validating healthcare applications. Health Inf J 2022; 28: 14604582221077000.

39.

Dhinagar

Thomopoulos

Laltoo

, et al. Efficiently training vision transformers on structural MRI scans for alzheimer’s disease detection. In 2023 45th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC). IEEE, 2023, pp. 1–6.

40.

Dong

Zhang

Qiu

, et al. An arrhythmia classification model based on vision transformer with deformable attention. Micromachines 2023; 14: 1155.

41.

Rahman

Valanarasu

JMJ

Hacihaliloglu

, et al. Ambiguous medical image segmentation using diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 11536–11546.

42.

Kim

. Diffusion adversarial representation learning for self-supervised vessel segmentation. arXiv preprint arXiv:220914566: 2022.

43.

Sergin

Huang

Chang

T-S

, et al. Image-based novel fault detection with deep learning classifiers using hierarchical labels. IISE Transactions 2024: 1–19.

44.

Wolleb

Bieder

Sandkühler

, et al. Diffusion models for medical anomaly detection. In International Conference on Medical image computing and computer-assisted intervention. Springer, 2022, pp. 35–45.

45.

Wang

. Predicting glaucoma pr requiring surgery using clinical free-text notes and transfer learning with transformers. Translational vision science & technology 2022; 11: 37.

46.

Allen

Siefkas

Pellegrini

, et al. A digital twins machine learning model for forecasting disease progression in stroke patients. Appl Sci 2021; 11: 5576.

47.

Kumar

Mallick

. The Internet of Things: insights into the building blocks, component interactions, and architecture layers. Procedia computer science 2018; 132: 109–117.

48.

Alhazmi

. A survivable internet of things scheme. J Adv Res Comput Appl 2018; 13: 19–26.

49.

Zyrianoff

Heideker

Silva

, et al. Architecting and deploying IoT smart applications: a performance–oriented approach. Sensors 2019; 20: 84.