=Paper=
{{Paper
|id=Vol-3764/invited1
|storemode=property
|title=Digital Twins in Healthcare: A forefront for knowledge representation techniques
|pdfUrl=https://ceur-ws.org/Vol-3764/invited1.pdf
|volume=Vol-3764
|authors=Swarnendu Ghosh
|dblpUrl=https://dblp.org/rec/conf/nsg/Ghosh24
}}
==Digital Twins in Healthcare: A forefront for knowledge representation techniques==
https://ceur-ws.org/Vol-3764/invited1.pdf
Digital Twins in Healthcare: A forefront for
knowledge representation techniques
Swarnendu Ghosh1,2,†
1
IEM Centre of Excellence for Data Science, Kolkata, WB, India
2
Institute of Engineering & Management (UEM), Kolkata, WB, India
Abstract
Digital twins have recently gathered significant interest in the healthcare community. This concept
promises to unlock various previously unavailable services such as remote monitoring, advanced visu-
alization, simulation of medical procedures, predictive analytics, demographic studies, and so on. At
present research in this area is localized and conducted independently. Thus, effective deployment of
digital twins in healthcare is still a work in progress due to inconsistent data representation and isolated
innovation without effective integration at large scale. Knowledge representation plays a vital role in
structuring, integrating, and reasoning over heterogeneous healthcare data sources such as electronic
health records, genomics data, clinical guidelines, reports, medical literature, and more. The process of
digitization is relevant not only to patients but also to healthcare professionals, infrastructure facilities,
devices, insurance providers, and even historical records. This work proposes to thoroughly highlight this
research gap and the current initiatives addressing these issues. It aims to review and consolidate existing
efforts in standardizing data structures for healthcare digital twins, with a focus on interoperability,
representation and integration across diverse healthcare domains.
Keywords
Digital Twins, Healthcare Informatics, Knowledge Representation.
1. Introduction
The uncivilized human species have solely depended on their biological immunity and be-
havioural traits for treating physical ailments[1]. The earliest civilizations started adopting
unorthodox practices to aid in the general well-being and lifestyle improvement. There have
been numerous records of herbal remedies, spiritual and yogic practices, and even primitive
surgical methods in ancient civilizations like the Egyptian, Indian, Greeks and so on [2]. Around
the 15th century, the Renaissance period introduced several aspects of modern medicine such
as diagnosis, anatomical studies and various surgical tools. As time progressed each century
brought us new concepts such as the Germ theory[3] and introduction of anesthesia[4] in the
19th century and other advancements like vaccines[5], antibiotics[6], radiology[7], ECG[8] and
so on in the 20th century. This accelerated innovation has continued to increase the periphery
of modern medicine. With the dawn of the 21st century, the age of automation took over. With
2nd Symposium on NLP for Social Good (NSG), 25-26 April, 2024, University of Liverpool, United Kingdom
†
This work is done with infrastructure support from the Innovation & Entrepreneurship Development Cell, IEM
Kolkata, India
$ drghosh90@gmail.com (S. Ghosh)
0000-0002-2220-1677 (S. Ghosh)
© 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0, International (CC BY 4.0).
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
immense advancement in machine learning, internet of things, 5G, cloud computing platforms,
healthcare informatics have taken a new turn[9]. Present efforts are being focused on the
automation of the diagnosis, robot assisted surgeries, remote clinical processing and more. The
complete mapping of the human genome[10] has led to a new era of precision medicine that
combines genomic and proteomic studies for drug design. These advancements have created
the hotbed for the emergence of the concept of digital twin.
Figure 1: Hypothetical scenario in a healthcare ecosystem
Definition 1. A digital twin is an accurate electronic representation or model of something
that has physical existence created using real-world data and simulations to mimic its behavior,
characteristics, and functionalities
The concept of digital twins is being adopted by several industries[11] such as defence, trans-
portation, manufacturing, urban planning, automobiles, e-commerce, environmental monitoring,
and last but not the least healthcare.
Though several of previous studies have been conducted in this topic[12, 13, 14], we will
discuss the possibilities of integrating the digital twin ecosystem with the current healthcare
industry especially highlighting the role of knowledge representation. A hypothetical healthcare
ecosystem is described in fig. 1.
2. Digital Twins in Healthcare: Scope, Target & Purpose
The healthcare industry has already started adopting the concept of digital twins in various
different ways. Till now the innovation has mostly been localized. The objective of this paper
is to propose a more global approach to make digital healthcare a reality. The applications of
digital twins in healthcare can be classified into several categories based on the application
scope, target areas and purpose that the digital twins would serve.
2.1. Scope
Digital twins is may be implemented at the level of a single patient[15], an entire organization[16]
or even a geographical area[17]. Each level of implementation serves different use cases and
application scenarios. A patient level twin may be developed by measuring physiological
parameters, gene sequencing, radiology scans, medical histories, psychometric profiles and so on.
At an organizational level, models of various organs, and biological processes can aid in running
simulations or providing training. Molecular modelling is often necessary for drug designs and
vaccine development. Furthermore various clinical infrastructures, services, equipment, etc.
can also have virtual counterparts to enable process simulations and optimization of various
clinical activities. Virtualization is not only limited to specific persons or organizations. They
may also span over geographical regions to aid in various types of demographic surveys, and
community based healthcare modelling.
2.2. Target
The next obvious aspect of digital twins is to figure out what components of the healthcare
industry can be represented using a digital twin.
• Multiomics: Multiomics modelling[18] covers various aspects regarding the molecular
dynamics of the human body. The genetic sequence and its electronic representation can
be considered as a genetic twin of a person. This genetic profile may be used to prepare
personalized treatment plans. Other than that models of various bio-molecules allow us
to simulate molecular interactions. Prediction and 3D visualizations of protein structures
is a big application area in this regard.
• Drugs: Molecular interactions can also be used for designing candidate drugs[19]. Digital
twins of such molecules may be used to measure docking feasibility. This can be further
extended to create precision medicine that takes genetic variations into considerations.
• Diseases: Modelling genetic variants of disease[20] causing microorganisms is a promis-
ing area to explore possible mutations and drug interactions or vaccine efficacy.
• Epidemics: Modelling mobility patterns of diseases[21] and various other environmental
factors that may trigger healthcare concerns leads to a better understanding of epidemics
and plan accordingly. Geographic twins of epidemic events can provide the platform for
planning containment zones and vaccination drives.
• Biological Systems: Various organs, and respective biological processes like digestion,
circulation, neural impulses can also have digital twins[22]. Generic models of these
organs or processes can be used for training purpose. Even patient specific models can
allow us to develop surgical or treatment plans by running simulations. Digital twins also
help in designing personalized prosthetics.
• Infrastructure: Even in administration we can have twins of healthcare facilities[23],
and personnel to optimize administration and services. Various clinical equipment and
devices can also have digital twins. These can be helpful in training and simulations.
Electronic Health Records can provide decision support for defining health policies and
insurance parameters.
2.3. Purpose
The final piece of the digital twin ecosystem is to define the purpose that is served by the twin.
We have already discussed about the various scopes and targets for digitization. These twins
would be useful in several scenarios.Digital twins may be used for training [24] of healthcare
personnel on new equipment or clinical procedures may carried out. Surgical simulations [25]
may be performed on patient specific twins of target organs. Demographic surveys allow us
to plan region specific healthcare services[26] and also create decision support systems for
epidemics and community healthcare. Past records may be used to forecast future outcomes[27]
of treatment protocols, health camps, and also anticipate maintenance needs and provide
supportive evidences for taking decisions. As discussed before, various vaccine[28] and drug
design rely on digital twins for simulating molecular interactions. Personalized medicine[18] is
also an application that consider genetic profile, medical history and lifestyle factors. Finally
healthcare institutions can use digital twins to optimize their services, and reduce costs.
3. Holistic Healthcare Framework
The proposed holistic healthcare framework is a hypothetical model that ideologically estab-
lishes the potential of the digital twin industry. The objective of this study is to recognise the
challenges of integrating digital twins into the existing healthcare industry and to propose
essential steps for initiating a globally connected healthcare industry that can exploit modern
technologies such as deep learning, internet of things(IoT), 5G/6G communications and cloud
computing. But before we proceed we must understand that a digital twin is not a nuclear
computational module working in isolation. It is an ecosystem consisting of various stake-
holders such as patients, healthcare facilities, healthcare personnel, technology developers,
device manufacturers, regulatory bodies, ethics committees, cybersecurity service providers,
educational & research institutions, financial support providers, government, advocacy groups
and more. It is built upon a versatile and robust technological stack [29]that interacts with
various external data repositories and under the supervision of ethics committees and regulatory
bodies.
3.1. Technological Stack
The necessary technological stack (as illustrated in fig. 2) for an ideal digital twin ecosystems
would have multiple functional modules. Obviously the ecosystem would be built on top of
the real world. This physical layer consists of the patients, healthcare professionals, medical
facilities, equipment, and other relevant institutions. The virtual world would start with the data
layer that would consist of the electronic data acquired from various sources such diagnostic
equipment, patient records, reports, genomic profiles and so on. Once we have acquired the
data the obvious next concerns are addressed by the communication layer and storage layer.
Figure 2: A schematic diagram for the technological stack for a digital twin ecosystem
Communication protocols may be defined as per implementation of the IoT infrastructure
built upon 5G or 6G network backbone. The storage layer would deal with the necessary
storage infrastructure. Blockchain techniques may be used for decentralization of information.
The representation layer deals with the computational representation of the acquired data.
The data would generally be acquired from different sources with different representation
format. A common modelling language is needed for a consistent embedding. For a globally
holistic healthcare framework common schema would be needed to assemble and integrate
multi-modal information in a consistent data structure. The computation layer consists of
the high performance computing infrastructure along with the advanced machine learning
algorithms, image processing and language processing toolkit and models along with rendering
engines. Finally the application layer would provide the necessary interface for visualization,
simulation, and analysis. This can be carried out through standard modes of human computer
interaction or even through augmented and virtual reality platform. Other than this there is a
obvious needs of communication with external repositories such as electronic health records,
government databases, insurance records, demographic information, pharmaceutical and disease
repository, multiomics data banks and more. Additionally there is a need for cyber-security
protocols across the stack and supervision from ethics and regulatory bodies.
3.2. One Species One Schema
For a digital twin ecosystem we need to operate with data acquired from various sources
such as patients, equipment, healthcare facilities, and external data banks. All these data have
vastly different formats and structure. The biggest challenge of this digital twin ecosystem
is to progress towards the one species one schema concept. This refers to a single model that can represent various elements of the healthcare industries
along with their relevant parameters and contexts. This would allow us to create the knowledge
graphs for healthcare digitalization[30]. The need of a common data structure is graphically
summarized in fig. 3.
Figure 3: The variety among data sources and the need of a common data structure.
This schema for a global knowledge base must be built with some specific properties that
ensures its sustainability. These properties, here abbreviated as M.U.S.C.L.E., refers to the
aspects of
• Multi-modality: Data may come from various sources but must be mapped to the same
schema
• Uniformity: Standardization of data representation is mandatory
• Scalability: The schema should be able to grow with the growing complexities of the
data
• Consistency: The updates in the schema should be non-disruptive in nature
• Longevity: The schema must be able to adapt and grow with time to remain relevant
with the fast moving pace of medical research as it can be a costly affair to revamp the
entire schema
• Ethics: Each transactions in the knowledge based must be traceable to the responsible
healthcare personnel or facility and must be explainable to avoid ethical or legal issues.
3.2.1. Knowledge Base Transactions
The knowledge base that is built upon the schema should be accessed and altered through
serialized transactions. Transactions to update the knowledge base may be at three different
levels:
1. Content : refers to nuclear updates that deals with specific parameters. These updates
do not disrupt other nodes or edges. e.g. Updating the height and weight of a patient
2. Context : refers to updates that have effect in its immediate neighbourhood. These
updates tends to display contextual significance. Updating diagnosis based on a test
report result. Test report parameters will be updated which will trigger update in the
diagnosis parameters
3. Concept: refers to the updates in the knowledge base that leads to addition, alteration or
deletion of entire conceptual branches. A proposed treatment by a physician opens up a
new branch for the healthcare facility that requires addition of several nodes and edges
such as hospital beds, OT reservations, insurance parameters and so on.
3.3. Unique Biological Identifier (UBID)
For the realization of a global knowledge repository, the first and most important step is to
develop a unique biological identifier. A unique globally standardized biological identifier must
be created. This may be similar to other identification documents like passports, social security
numbers, taxpayer identification numbers, Aadhar Card, VoterCard etc. However, a biological
identifier must be associated to a biological signature of a being such as the fingerprints, retina
scans or even genome sequence[31]. This must be standardized by the respective government
as per uniform global standards. Most importantly it should be mandatory to associate with
all healthcare related procedures similar to a taxpayer identification being connected with all
financial transactions.
3.4. Globalization
Besides the UBID being necessary to unify healthcare transactions, a digital twin system would
also require innovation tracking [32] and development of global policies, and standardization.
We must understand that isolated innovation is the source of inconsistencies. Existing predictive
models must be standardized and existing literature must digitized to create a viable ontology.
Additionally an effort must be made to ensure every clinic, hospital, pharmacy, doctor, patients,
equipment is connected to the common Healthcare Framework.
4. Conclusion
It is evident that revolution in global healthcare requires a migration to the digital twin ecosystem.
However, several significant actions must be taken. From development of unique biological
identifiers to moving towards a unified knowledge repository built on robust backbone schema.
The proposed work tries to establish the need of "One Species One Schema" and the role that
knowledge representation plays in the transformation of modern healthcare. The schema would
need to be carefully designed to ensure long-term sustainability. The ultimate goal would be
an unified knowledge graph that connects patients, healthcare facilities, professionals and all
other stakeholder to ensure quality of medical services.
References
[1] L. Abi-Rached, M. J. Jobin, S. Kulkarni, A. McWhinnie, K. Dalva, L. Gragert, F. Babrzadeh,
B. Gharizadeh, M. Luo, F. A. Plummer, et al., The shaping of modern human immune
systems by multiregional admixture with archaic humans, Science 334 (2011) 89–94.
[2] B. V. Subbarayappa, The roots of ancient medicine: an historical outline, JOURNAL OF
BIOSCIENCES-BANGALORE- 26 (2001) 135–143.
[3] L. Pasteur, On the germ theory, Science (1881) 420–422.
[4] D. H. Robinson, A. H. Toledo, Historical development of modern anesthesia, Journal of
Investigative Surgery 25 (2012) 141–149.
[5] C. Mark, J. G. Rigau-Pérez, The world’s first immunization campaign: the spanish smallpox
vaccine expedition, 1803–1813, Bulletin of the History of Medicine (2009) 63–94.
[6] A. Fleming, Penicillin, British medical journal 2 (1941) 386.
[7] A. M. Thomas, A. K. Banerjee, The history of radiology, OUP Oxford, 2013.
[8] W. B. Fye, A history of the origin, evolution, and impact of electrocardiography, The
American journal of cardiology 73 (1994) 937–949.
[9] S. D. Pruitt, J. E. Epping-Jordan, Preparing the 21st century global healthcare workforce,
Bmj 330 (2005) 637–639.
[10] J. C. Venter, M. D. Adams, E. W. Myers, P. W. Li, R. J. Mural, G. G. Sutton, H. O. Smith,
M. Yandell, C. A. Evans, R. A. Holt, et al., The sequence of the human genome, science 291
(2001) 1304–1351.
[11] D. Jones, C. Snider, A. Nassehi, J. Yon, B. Hicks, Characterising the digital twin: A
systematic literature review, CIRP journal of manufacturing science and technology 29
(2020) 36–52.
[12] T. Erol, A. F. Mendi, D. Doğan, The digital twin revolution in healthcare, in: 2020
4th international symposium on multidisciplinary studies and innovative technologies
(ISMSIT), IEEE, 2020, pp. 1–7.
[13] T. Sun, X. He, Z. Li, Digital twin in healthcare: Recent updates and challenges, Digital
Health 9 (2023) 20552076221149651.
[14] H. Elayan, M. Aloqaily, M. Guizani, Digital twin for intelligent context-aware iot healthcare
systems, IEEE Internet of Things Journal 8 (2021) 16749–16757.
[15] P. Barbiero, R. Vinas Torne, P. Lió, Graph representation forecasting of patient’s medical
conditions: Toward a digital twin, Frontiers in genetics 12 (2021) 652907.
[16] A. Karakra, F. Fontanili, E. Lamine, J. Lamothe, Hospit’win: a predictive simulation-based
digital twin for patients pathways in hospital, in: 2019 IEEE EMBS international conference
on biomedical & health informatics (BHI), IEEE, 2019, pp. 1–4.
[17] A. Perzynski, K. Berg, J. Dalton, Digital twin neighborhoods for precision population
health: Initial results of community conversations, Innovation in Aging 7 (2023) 53–54.
[18] M. Cellina, M. Cè, M. Alì, G. Irmici, S. Ibba, E. Caloro, D. Fazzini, G. Oliva, S. Papa, Digital
twins: The new frontier for personalized medicine?, Applied Sciences 13 (2023) 7940.
[19] K. Subramanian, Digital twin for drug discovery and development—the virtual liver,
Journal of the Indian Institute of Science 100 (2020) 653–662.
[20] R. Laubenbacher, J. P. Sluka, J. A. Glazier, Using digital twins in viral infection, Science
371 (2021) 1105–1106.
[21] S. Alrashed, N. Min-Allah, I. Ali, R. Mehmood, Covid-19 outbreak and the role of digital
twin, Multimedia Tools and Applications 81 (2022) 26857–26871.
[22] X. Zhong, F. Babaie Sarijaloo, A. Prakash, J. Park, C. Huang, A. Barwise, V. Herasevich,
O. Gajic, B. Pickering, Y. Dong, A multidisciplinary approach to the development of
digital twin models of critical care delivery in intensive care units, International Journal
of Production Research 60 (2022) 4197–4213.
[23] Y. Peng, M. Zhang, F. Yu, J. Xu, S. Gao, Digital twin hospital buildings: an exemplary case
study through continuous lifecycle integration, Advances in Civil Engineering 2020 (2020)
1–13.
[24] M. W. Zackoff, D. Davis, M. Rios, R. D. Sahay, B. Zhang, I. Anderson, M. NeCamp, I. Rogue,
S. Boyd, A. Gardner, et al., Tolerability and acceptability of autonomous immersive virtual
reality incorporating digital twin technology for mass training in healthcare, Technical
Report, LWW, 2023.
[25] K. Hagmann, A. Hellings-Kuß, J. Klodmann, R. Richter, F. Stulp, D. Leidner, A digital
twin approach for contextual assistance for surgeons during surgical robotics training,
Frontiers in Robotics and AI 8 (2021) 735566.
[26] F. Pilati, R. Tronconi, G. Nollo, S. S. Heragu, F. Zerzer, Digital twin of covid-19 mass
vaccination centers, Sustainability 13 (2021) 7396.
[27] A. Allen, A. Siefkas, E. Pellegrini, H. Burdick, G. Barnes, J. Calvert, Q. Mao, R. Das, A
digital twins machine learning model for forecasting disease progression in stroke patients,
Applied Sciences 11 (2021) 5576.
[28] T. Zohdi, Machine-learning and digital-twins for rapid evaluation and design of injected
vaccine immune-system responses, Computer Methods in Applied Mechanics and Engi-
neering 401 (2022) 115315.
[29] S. N. G. Gourisetti, S. Bhadra, D. J. Sebastian-Cardenas, M. Touhiduzzaman, O. Ahmed, A
theoretical open architecture framework and technology stack for digital twins in energy
sector applications, Energies 16 (2023) 4853.
[30] A. Carbonaro, A. Marfoglia, F. Nardini, S. Mellone, Connected: leveraging digital twins
and personal knowledge graphs in healthcare digitalization, Frontiers in Digital Health 5
(2023) 1322428.
[31] C. Lippert, R. Sabatini, M. C. Maher, E. Y. Kang, S. Lee, O. Arikan, A. Harley, A. Bernal,
P. Garst, V. Lavrenko, et al., Identification of individuals by trait prediction using whole-
genome sequencing data, Proceedings of the National Academy of Sciences 114 (2017)
10166–10171.
[32] J. Barlow, Managing innovation in healthcare, World scientific publishing company, 2016.