=Paper=
{{Paper
|id=Vol-2114/paper11
|storemode=property
|title=A Systematic Approach to Define Requirements and Engineer the Ontology for Semantically Merging Data Sets for Personal-Centric Healthcare Systems
|pdfUrl=https://ceur-ws.org/Vol-2114/paper11.pdf
|volume=Vol-2114
|authors=Aleksandr Kormiltsyn
}}
==A Systematic Approach to Define Requirements and Engineer the Ontology for Semantically Merging Data Sets for Personal-Centric Healthcare Systems==
https://ceur-ws.org/Vol-2114/paper11.pdf
A Systematic Approach to Define Requirements
and Engineer the Ontology for Semantically
Merging Data Sets for Personal-Centric
Healthcare Systems
Aleksandr Kormiltsyn
Department of Software Systems, Tallinn University of Technology,
Akadeemia tee 15A, 12816, Tallinn, Estonia
alexandrkormiltsyn@gmail.com
Abstract. A fundamental paradigm shift to patient-centric health ser-
vices is needed to meet the challenges such as increasing healthcare costs,
ageing population, and unhealthy lifestyles. Designing patient-centric
services is challenging because of medical- and health data heterogeneity
and lack of standardization for Personal Health Records (PHR). Further,
designing patient-centric systems, especially in healthcare is a challenge
as they are influenced by emotions when users are influenced by emo-
tions triggered by privacy issues. In this study, we focus on the engineer-
ing method to design a patient-centric system that considers medical-
data heterogeneity and emotional security issues. We plan to develop a
methodology for designing a patient-centric system with a multi-agent
system (MAS) approach where agents operate with smart contracts. An
emotional security framework is integrated into the design of such sys-
tems. Multi-faceted validation and verification are used to evaluate the
engineered method. A framework for Evaluation in Design Science Re-
search (FEDS) is used to define the evaluation activities and tools.
Keywords: eHealth, Patient-centric systems, Smart contracts, Emo-
tional security, Multi-agent systems, Ontology
1 Introduction
Health information technology has the potential to improve medical quality,
patient safety, educational resources and patient-physician communication while
decreasing costs [7]. The development of central systems such as Smart Open
Services for European Patients (epSOS)1 and the Estonian Health Information
System (EHIS)2 increases interoperability on national- and international levels.
Still, the available data is often fragmented and the secondary use is limited
[11]. The term ’secondary use’ can be defined as the use of data collected for one
purpose to study a new problem. Secondary uses of data could include quality
1
http://www.epsos.eu
2
http://www.e-tervis.ee
measurement, public health surveillance, and patient access to data about their
illness [6]. The fragmentation stems from decentralized and disintermediated
data sets.
The main purpose of Personal Health Records (PHR) is that a patient is the
author and owner of his/her medical data that can be shared with other individ-
uals, including healthcare professionals, or automated clinical decision-support
services. The growing amount of personal data available increases the role of
patient-centric systems in healthcare. In contrast, Electronic Health Records
(EHR) are created and owned by healthcare providers.
The complexity of data analysis, processing and security increases together
with the number of different healthcare systems. Another problem is that the
data is created by healthcare professionals and based on disease episodes that
are only a part of the patients’ health-related data and history. For example,
several patients with chronic conditions measure their blood pressure, or blood
glucose regularly while these data sets are not stored in the context of EHR and
thus, limiting their future use.
Design and development of patient-centric systems are challenging because
users are influenced by emotions triggered by privacy issues. As users are not
forced to use such software, these issues lead to adoption failure, e.g., individuals
do not trust to share their sensitive medical data with an application that does
not provide transparency of personal data usage.
The importance for a healthcare domain is related to the possibility of sharing
PHR between different stakeholders. Individuals could manage their own data
and share it with healthcare providers and other stakeholders such as medical
researchers, health insurance providers, and automated Clinical Decision Sup-
port Systems (CDSS). Additional medical data available to healthcare provider
enables the provision of personalized guidelines to a patient. Personalized guide-
lines help patients to understand how to achieve their goals. On the other hand,
the availability of CDSS reduces the number of people requiring direct contact
with a doctor and increases the accessibility of medical services to people who
really need it. Further, researchers gain access to personal medical data.
In this research, we focus on the PHR usage in the healthcare processes
and involvement of patients as main responsible stakeholders by engineering a
method for designing decentralized patient-centric systems. The main proposed
method’s concepts are: the number of healthcare processes is not limited; trust,
transparency and influence of patient’s emotions during medical data sharing
are essential for the usage of such a system; medical data heterogeneity is un-
avoidable.
To enable these main concepts we use different techniques: Multi Agent Sys-
tems (MAS) to support an unlimited number of healthcare processes, smart
contracts are used by both human and non-human agents to enable trust and
transparency while sharing PHR, embedding emotional security framework in
the MAS design phase to avoid negative emotions while using such a system.
The ontology is used to describe medical data integration processes between
different agents.
92
The remainder of the paper is structured as follows. Section 2 presents related
work and Section 3 describes the problem statement as well as our contributions.
Next, Section 4 presents the design-science research method research actions in
details and Section 5 references the conference paper as initial results used as a
basis for the current research. Section 6 provides the evaluation plan for designed
methodology, Section 7 concludes the paper together with providing directions
for future research. In a final Section 8 acknowledgements to supervisors are
presented.
2 State of the Art
Literature review [3, 5, 22] shows that researchers provide solutions for integra-
tion challenges based on data-heterogeneity reduction by developing a new on-
tology or merging existing ones. This approach is effective for decreasing the
heterogeneity of different EHR standards and a limited number of processes.
However, the amount of PHR data is rapidly growing with the development of,
e.g., IoT devices. The number of processes using PHR is not limited, as pa-
tients visit different hospitals. Therefore, an efficient integration of PHR and
EHR requires a dynamic approach that focuses on the process rather than on
data standardization. Paying attention to data-flow processes is an important
part of EHR and PHR data integration as health data flows along processes
in distributed systems. Research [2] proposes smart contracts for logging the
patient-provider relationships that associate a medical record with viewing per-
missions and data retrieval instructions for the execution on external databases.
The role of emotions and associated quality goals that influence the adoption
and effective use of socio-technical systems is detailed in research [15]. In order to
design and develop a system, it is important to elicit requirements in the form of
functional, nonfunctional and emotional goals [15]. To do this, some research has
identified methodologies for segmenting the target users of a system as different
stakeholders have different goals and requirements [21]. Research [21] hypoth-
esizes that user segmentation is helpful in requirements engineering to direct
systems analysts and owners to segment potential users and select target users
as the first stage of requirements engineering. In [17], the authors introduce a
straightforward notation for modeling emotional goals in agent-oriented mod-
eling. Further, little is known about software engineering methodology defining
how best emotional requirements can be elicited from stakeholders. There is
no systematic methodology to help elicit such emotional requirements early on
before the systems are designed.
3 Problem Statement and Contributions
Healthcare services become more expensive and their accessibility to patients
decreases. Patient-centric healthcare systems enable PHR data collection and
processing by CDSS and often do not require physical contact between patient
and doctor. The PHR and EHR integration requires a solution for medical data
93
heterogeneity as PHR does not have common data standard. The processing of
integrated PHR and EHR needs to be transparent for patients so they adopt
the patient-centric system. In this paper, we fill two gaps: (1) the integration
process of PHR and EHR; (2) and embedding an emotional security framework
into the design of patient-centric systems. In the scope of integrated healthcare,
there are different main stakeholders: individuals, researchers that use medical
data, healthcare professionals, state organizational units as statistics and social
departments, and insurance companies.
We develop a methodology for patient-centric systems design. The methodol-
ogy includes an emotional attachment framework to engage users and alleviate
the users from feeling a negative emotion such as distrust, lack of ownership
private and confidential data. Lack of transparency in health data processing
leads to negative feelings for an individual. Target users for this methodology
are IT architects and designers because there is no current methodology in IT
for designing patient-centric systems.
Our planned contribution is a methodology for the design of dynamic PHR
and EHR integration processes. We plan to use intelligent agents in MAS and
smart contracts for sharing PHR. Intelligent agents are used for PHR collection
from personal devices and can provide feedback based on the collected data
analysis while blockchains provide transparency of PHR usage to the patient.
To eliminate emotional security risks, we expand traditional functional and non-
functional requirements of software system with emotional ones.
The main research question is how to dynamically merge and process inte-
grated PHR and EHR in the cross-organizational processes of the patient-centric,
decentralized healthcare system? To answer this question, a number of challenges
(sub-questions) need to be addressed.
The first sub-question is how to collect and process PHR data in the con-
text of integrated healthcare to improve the quality of diagnostics and cures?
To answer this question, we focus on the dynamic integration of EHR and PHR
considered from a process-based point of view. After defining the dynamic inte-
gration process, we turn to the shared cross-organizational processes.
The second sub-question is how to specify cross-organizational processes for
handling EHR and PHR in a dynamic and secure way in a distributed patient-
centric systems? The dynamic integration of EHR with PHR and the embed-
ded emotional security framework enables the effective adoption and usage of
integrated data in a different number of e-services. Smart contracts used by in-
telligent agents in multi-agent systems (MAS) eliminate trust issues between
stakeholders while providing a patient with PHR and EHR data security.
As the security of personal data is a very sensitive and important topic, the
third research question is how to perform a privacy-aware analysis of integrated
PHR and EHR in the context of the cross-organizational process in a distributed
patient-centric system? The privacy aware protocol [19] that is used in data-
driven EHR systems for private data analysis is extended to support integrated
PHR and emotional security aspects. The protocol is based on the emotional
94
requirements to PHR data sharing as well as the ontology for the collection of
emotional goals and secured views proposed in [19].
4 Research Methodology and Approach
For this research, the design-science research method is considered. This method
is solution-oriented and requires the creation of an innovative, purposeful artifact
for a special problem domain [24]. First, we consider the dynamic integration
process for integrating the PHR with EHR. The proposed process is used as
a foundation for designing cross-organizational processes where medical data
is used by different human- and non-human agents. Finally, we define security
protocol based on the emotional aspects of information security.
For designing dynamic integration processes, the requirements for PHR data
collection are defined. The ontology for the PHR and EHR data integration and
the integration process flow are presented. The requirements are defined with
the Agent-Oriented Modelling (AOM) approach as in [23, 14]. According to [20],
analyzing the problem of this socio-technical domain can be performed by us-
ing a goal model. The objective of goal models is to serve as communication
media between technical- and non-technical stakeholders for generating under-
standable domain knowledge. We use goal models in the context of requirements-
engineering processes.
The processing of integrated EHR and PHR requires an ontology definition
to reduce semantic- and syntactic heterogeneity. An ontology for the dynamic
integration process is designed using the Protégé tool3 together with the Web
Ontology Language (OWL)4 . OWL is used to describe taxonomies and classifi-
cation networks. An evaluation of the ontology is performed with the HermiT5
reasoner that is based on a novel hypertableau calculus [18] to provide much
more efficient reasoning than any previously known algorithm.
We describe the PHR and EHR integration process with the Business Process
Model and Notation (BPMN) [1], a graphical representation for specifying busi-
ness process model. BPMN is suitable for reasoning about cross-organizational
integration processes including PHR- and EHR integration process. The inte-
gration process is validated with the Signavio6 tool that checks BPMN diagrams
for existing conflicts.
For the evaluation of defined integration process of PHR and EHR, we build a
formal Colored Petri Nets (CPN) model in order to detect and eliminate eventual
design flaws, missing specifications, security and privacy issues [8, 10]. A CPN
is a graphical oriented language for the design, specification, simulation as well
as the verification of systems and describes the states of a modeled system and
the events (transitions) that cause the system to change states [13]. We refer the
reader to [9] for more information about CPN.
3
http://protege.stanford.edu
4
https://www.w3.org/2001/sw/wiki/OWL
5
http://www.hermit-reasoner.com/
6
https://www.signavio.com/
95
To define the cross-organizational processes where medical data is used by dif-
ferent human- and non-human agents, first intelligent BDI agents are described
as they offer a straightforward formalization of reasoning human agents with
intuitive concepts (beliefs, desires, and intentions) that closely match human
reasoning[4]. Next, best practices for designing decentralized healthcare systems
are proposed.
We define best practices for designing decentralized healthcare systems by
integrating smart contracts in the communication layer between different non-
human agents involved in the cross-organization process. To describe non-human
BDI agents, ontology built for the integration process in the previous step is
extended using OWL and validate it with HermiT reasoner.
Emotional requirements for PHR handling and sharing processes are included
in the early stages of the PHR and EHR data integration. The emotional frame-
work [16] is integrated into the AOM approach and extends the AOM goal model
with emotional goals in order to define the emotional requirements for PHR data
sharing. The ontology for the integration process is extended to cover the collec-
tion of emotional aspects. We use BPMN to describe the security protocol and
blockchain as a technology enabling transparency in the cross-organizational
process.
To meet our goal of designing the architecture of a patient-centric system,
three milestones are defined. First, the design of collecting and processing the
PHR data, then the specification of a cross-organizational process followed by
design of a secure PHR sharing protocol based on the emotional requirements. To
define requirements for the PHR data collection, we use AOM. The ontology for
the PHR and EHR integration process is defined using OWL and the integration
process is described with BPMN. The specification of a cross-organizational pro-
cess includes a definition of non-human agents described with AOM and OWL.
The usage of smart contracts is considered in the context of the dynamic inte-
gration process. Secure PHR sharing protocol for the PHR includes a definition
of emotional requirements for PHR data sharing described with AOM and the
emotional framework, followed by the ontology for PHR emotional aspects col-
lection described with OWL. Finally, we design a security protocol for PHR
sharing using BPMN and blockchain technology.
5 Preliminary or Intermediate Results
In the workshop paper [12], we consider the dynamic integration of EHR and
PHR from the process based point of view. The process workflow emphasizes the
EHR and PHR context difference while the system requirements are defined with
an AOM goal model describing functional and quality goals. The ontology for
the integration process defined in the workshop paper is a foundation describing
main concepts for the integration process. Finally, the BPMN representation of
the integration process shows different stages of PHR- and EHR data preparation
and -processing. All these findings set the scope of our research and we use them
as a basis for future work.
96
6 Evaluation Plan
We use a combination of different approaches for the evaluation. First, CPN
is used to evaluate processes in the design phase to eliminate, or minimize the
security risks. Also, it is possible to consider concurrency conflicts, dependabil-
ity issues and detect and eliminate eventual design flaws as well as security
and privacy issues with CPN. For architecture design, we use a combination of
Requirements Bazaar for requirements gathering and also for the evaluation of
the architecture. The Architecture Tradeoff Analysis Method (ATAM) is used
for evaluating and analyzing in a qualitative and empirical way the architec-
ture of a patient-centric system with the related functional, non-functional and
emotional requirements. ATAM is mostly applied for exploring the use of best-
practice architectural styles, for exploring quality attributes of architectures and
for evaluating existing systems. ATAM also helps to modify an architecture, or
integration work with new systems. Also, we use the HermiT reasoner for vali-
dating the ontology for the integration process and BPMN for reasoning about
cross-organizational integration processes. We also conduct user-centered eval-
uations to capture user experiences with the design of the system to measure
their level of engagement and emotional aspects. This evaluation we conduct us-
ing the emotional attachment framework as a practice lens to direct interviews
and usability testing activities.
7 Conclusions
In this paper, we describe current issues and challenges in electronic healthcare,
focusing on the need to create a systematic approach to designing a decentral-
ized patient-centric system that meets these challenges. We then define three
milestones to meet our goal of designing the architecture of such a system. To
solve the PHR heterogeneity issue in healthcare processes, first, the design of
collecting and processing PHR is proposed. However, the patient involvement in
healthcare processes creates additional challenges for such processes and there-
fore, we present the specification for cross-organizational processes working with
PHR. Shifting the paradigm to patient-centric increases the role of patient emo-
tions and privacy issues. Thus, we design the secure PHR sharing protocol that
enables transparency of the PHR usage to the patient and based on the patient’s
emotional requirements.
8 Acknowledgments
The author would like to thank both supervisors Alex Norta from Tallinn Tech-
nical University of Technology and Antonette Mendoza from the University of
Melbourne.
97
References
1. Allweyer, T.: BPMN 2.0: introduction to the standard for business process mod-
eling. BoD–Books on Demand (2016)
2. Azaria, A., Ekblaw, A., Vieira, T., Lippman, A.: Medrec: Using blockchain for
medical data access and permission management. In: Open and Big Data (OBD),
International Conference on. pp. 25–30. IEEE (2016)
3. del Carmen Legaz-Garcı́a, M., Martı́nez-Costa, C., Menárguez-Tortosa, M.,
Fernández-Breis, J.T.: A semantic web based framework for the interoperability
and exploitation of clinical models and ehr data. Knowledge-Based Systems 105,
175–189 (2016)
4. Casali, A., Godo, L., Sierra, C.: A graded bdi agent model to represent and reason
about preferences. Artificial Intelligence 175(7-8), 1468–1478 (2011)
5. Dogac, A., Laleci, G.B., Kirbas, S., Kabak, Y., Sinir, S.S., Yildiz, A., Gurcan, Y.:
Artemis: Deploying semantically enriched web services in the healthcare domain.
Information Systems 31(4), 321–339 (2006)
6. Hersh, W.R.: Adding value to the electronic health record through secondary use
of data for quality assurance, research, and surveillance. Clin Pharmacol Ther 81,
126–128 (2007)
7. Hoyt, R.E., Yoshihashi, A.K.: Health Informatics: Practical guide for healthcare
and information technology professionals. Lulu. com (2014)
8. Jensen, K.: Coloured petri nets. In: Discrete Event Systems: A New Challenge for
Intelligent Control Systems, IEE Colloquium on. pp. 5–1. IET (1993)
9. Jensen, K., Kristensen, L.M.: Coloured Petri nets: modelling and validation of
concurrent systems. Springer Science & Business Media (2009)
10. Jensen, K., Kristensen, L.M., Wells, L.: Coloured petri nets and cpn tools for
modelling and validation of concurrent systems. International Journal on Software
Tools for Technology Transfer 9(3-4), 213 (2007)
11. Köpcke, F., Trinczek, B., Majeed, R.W., Schreiweis, B., Wenk, J., Leusch, T.,
Ganslandt, T., Ohmann, C., Bergh, B., Röhrig, R., et al.: Evaluation of data com-
pleteness in the electronic health record for the purpose of patient recruitment into
clinical trials: a retrospective analysis of element presence. BMC medical informat-
ics and decision making 13(1), 37 (2013)
12. Kormiltsyn, A., Norta, A.: Dynamically integrating electronic-with personal health
records for ad-hoc healthcare quality improvements. In: International Conference
on Digital Transformation and Global Society. pp. 385–399. Springer (2017)
13. Leiding, B., Norta, A.: Mapping requirements specifications into a formalized
blockchain-enabled authentication protocol for secured personal identity assurance
14. Lister, K., Sterling, L., Taveter, K.: Reconciling ontological differences by assistant
agents. In: Proceedings of the fifth international joint conference on Autonomous
agents and multiagent systems. pp. 943–945. ACM (2006)
15. Mendoza, A., Miller, T., Pedell, S., Sterling, L., et al.: The role of users emotions
and associated quality goals on appropriation of systems: two case studies
16. Miller, T., Pedell, S., Lopez-Lorca, A.A., Mendoza, A., Sterling, L., Keirnan,
A.: Emotion-led modelling for people-oriented requirements engineering: The case
study of emergency systems. Journal of Systems and Software 105, 54–71 (2015)
17. Miller, T., Pedell, S., Mendoza, A., Keirnan, A., Sterling, L., Lopez-Lorca, A.A.:
Emotionally-driven models for people-oriented requirements engineering: the case
study of emergency systems. IEEE Transactions on Software Engineering (2014)
98
18. Motik, B., Shearer, R., Horrocks, I.: Hypertableau Reasoning for Description Log-
ics. Journal of Artificial Intelligence Research 36, 165–228 (2009)
19. Nguyen, T.A., Le-Khac, N.A., Kechadi, M.T.: Privacy-aware data analysis mid-
dleware for data-driven ehr systems. In: International Conference on Future Data
and Security Engineering. pp. 335–350. Springer (2017)
20. Norta, A., Mahunnah, M., Tenso, T., Taveter, K., Narendra, N.C.: An agent-
oriented method for designing large socio-technical service-ecosystems. In: 2014
IEEE World Congress on Services. pp. 242–249. IEEE (2014)
21. Sherkat, M., Miller, T., Mendoza, A.: Does it fit me better? user segmentation
in requirements engineering. In: Software Engineering Conference (APSEC), 2016
23rd Asia-Pacific. pp. 65–72. IEEE (2016)
22. Sonsilphong, S., Arch-int, N., Arch-int, S., Pattarapongsin, C.: A semantic inter-
operability approach to health-care data: Resolving data-level conflicts. Expert
Systems 33(6), 531–547 (2016)
23. Sterling, L., Taveter, K.: The art of agent-oriented modeling. MIT Press (2009)
24. Von Alan, R.H., March, S.T., Park, J., Ram, S.: Design science in information
systems research. MIS quarterly 28(1), 75–105 (2004)
99