Content uploaded by Fazle Rabbi
Author content
All content in this area was uploaded by Fazle Rabbi on May 21, 2019
Content may be subject to copyright.
An architectural design for
self-reporting e-health systems
Suresh Kumar Mukhiya∗, Fazle Rabbi∗, Ka I Pun∗,†, Yngve Lamo∗
∗Western Norway University of Applied Sciences, Norway
{suresh.kumar.mukhiya, fazle.rabbi, ka.i.pun, yngve.lamo}@hvl.no
†University of Oslo, Norway
Abstract—Worldwide, public healthcare is challenged to de-
liver consistent and cost-efficient services. The cost of healthcare
is increasing primarily due to growing populations and more
expensive treatment. Health facilities in many countries are
reaching their full operational capacity, and their resources are
less than what is required to deliver the expected quality of
care. Under these circumstances, ICT has a major role to play
in mitigating the growing need for hospital care. In this paper,
we present a cloud-based interoperable architecture built on the
top of Service-Oriented Architecture (SOA) for Internet-based
treatment where patients can directly interact with underlying
healthcare systems such as Electronic Health Records (EHRs).
Based on this architecture, we also present a prototype for
screening and monitoring of mental and neurological health
morbidities. The proposed solution is based on the healthcare
interoperability standard, HL7 FHIR, that enables healthcare
providers to assess and monitor patients condition in a secure
environment.
Index Terms—service-oriented architecture, healthcare, self-
screening, HL7 FHIR, SMART on FHIR, Internet-based Treat-
ments
I. INTRODUCTION
E-health is an emerging field of medical informatics which
refers to the use of information and communications technolo-
gies in healthcare. According to World Health Organization
(WHO), e-health is the cost-effective and secure use of in-
formation and communications technologies to support health
and health-related fields, including healthcare services, health
surveillance, health literature, health education, knowledge and
research [44]. Today e-Health infrastructures and systems are
being considered as central to the provision of high qual-
ity, citizen-centered healthcare. A very common requirement
for e-health applications is to exchange information between
healthcare systems and to support patients while they are
staying at home. Although there exist many techniques such
as electronic surveys to get data from patients for screening
and monitoring, they are not integrated with existing electronic
health record systems (EHRs), the data are not standardized,
not interoperable and not easily accessible to researchers and
clinicians [7]. As a result, healthcare providers are in general
unable to use any standardized system for collecting and
analyzing information from patients while the patients are not
inside a hospital and cannot share the information with other
care providers if needed. To resolve these issues, we propose
This project is partially funded by INTROMAT(259293/o70).
an architectural design for a self-reporting e-health system that
uses healthcare standards for interoperability.
In this paper, we focus on the mental healthcare domain as
there is an increasing rate for this illness. WHO states that
one in four people in the world will be affected by mental or
neurological disorders at some point in their lives and around
450 million people currently suffer from such conditions [44].
People suffering from mental health problems may have diffi-
culty in comprehending, acting or processing the experiences
and information appropriately. This is due to the illness which
can affect the mental or neurological functioning including
concentration, memory, initiative, ability to plan, summing
up the information, experience of time, and generalization
[38]. Lack of such cognitive functioning can have a serious
impact on patients, families, friends, and society. Dealing
with such mental health patients can be economically [14],
[35], physically and emotionally challenging. Several types
of research have been carried out to facilitate automated and
personalized screening and treatment to such mental health
patients. Human Brain Project1, eMEN [16], and INTROMAT
[26] are among the European projects that address such mental
health problems. In particular, the goal of the INTROMAT
project is to offer personalized treatments for patients suffering
from mental and neurological disorders.
We envision developing an adaptive system which can
support citizen-centered mental healthcare. We focus on the as-
sessment and monitoring process in order to develop a patient-
centered healthcare solution. As part of the INTROMAT
project, we are developing core services that will be made
available to research projects through the Internet of Things
(IoT) and application platforms. Examples of such services
are data processing services, data storage, security and privacy
services, artificial intelligence (cognitive computing, machine
learning, deep learning), event processing, analytical services,
and others. We provide an overview of the proposed solution
which includes: (i) a cloud-based interoperable architecture
built on the top of SOA, named as INTROMAT Core, for
self-assessment and evaluation of mental or neurological mor-
bidity, (ii) HL7 FHIR standard questionnaire, (iii) a working
prototype of self-assessment and monitoring mobile app, and
(iv) a list of research challenges and opportunities in the field.
The paper is structured as follows: Section II outlines the
1https://www.humanbrainproject.eu/en/
main motivation for building a self-reporting e-health system,
Section III presents the INTROMAT core architecture, its
workflow and quality attributes. In Section IV we describe
a case study for depression with corresponding domain model
for self-reporting system. Section V outlines some of the re-
lated works. Section VI concludes the paper with a discussion
addressing some challenges and also mentions future work.
II. MOT IVATIO N
In this paper, we will be addressing the following three
challenges:
•heterogeneity of healthcare information across various
healthcare service providers
•lack of integrated Internet-based treatment that is acces-
sible for patients and healthcare providers
•lack of adaptiveness in patient-centered care.
Healthcare ICT technologies have substantial contributions
in uplifting the quality of life and reducing socioeconomic bur-
dens. Technology provides a variety of promising approaches
to mental health and neurological practices, training, and
supervision in several ways including therapeutic intervention
[20] and cognitive behavioral therapy (CBT) [42]. Despite this
high prevalence of ICT system in healthcare, the inadequacy
provision of interventions for the prevention and treatment
of mental or neurological problems has become a global
challenge. This is mainly due to lack of standardization
in data collection and data preparation methodologies, and
heterogeneity of the data [25]. This limits the sharing of data
among different healthcare providers. The limitation in sharing
health data is detrimental to patient care, provider satisfaction,
and healthcare cost [21]. Moreover, providing access to and
sharing of health data have been shown to benefit and empower
the patients [45]. The complexity of heterogeneity in data
and restriction in sharing patient data have hindered stan-
dardization and interoperability in the healthcare industry [1].
In this project, we propose to overcome these problems of
data collection and interoperability by following a common
health standard HL7 FHIR (Fast Healthcare Interoperability
Resource) [2], [11]. HL7 FHIR solutions are designed for
web standards like XML, JSON, HTTP, and OAuth, are con-
structed from a set of modular components called Resources
which can be grouped into working healthcare systems. HL7
FHIR specification component terminology is supported by
the coding system that defines the content being exchanged.
This coding system is established by external organizations
including SNOMED-CT, LOINC, RxNorm, ICD family and
others [2]. We selected required codes for mental health
care from these organizations. OpenEHR2is an alternative to
HL7 FHIR, which is maintained by the openEHR foundation.
Both HL7 FHIR and OpenEHR have open-standard and data
interoperability as the main vision. HL7 FHIR is created to
exchange data using REST API using the block of around 100
FHIR resources. In contrast to HL7 FHIR resources, openEHR
2https://www.openehr.org
uses over 300 more complex Archetypes which are designed
to provide a maximal set of data elements.
People suffering from mental health require continuous
monitoring and treatment. It is not only costly to provide the
continuous monitoring and treatment using hospital facilities
but it also has a social burden for patients and families.
Internet-based psychotherapy treatments play a major role to
resolve these issues and evidence illustrates that some forms
of Internet-based treatments have comparable outcomes as
conventional face-to-face psychotherapy [12]. Moreover, these
Internet-based interventions can be presented to a large popu-
lation flexibly and inexpensively. Internet-based psychotherapy
includes treatments programs that are delivered via the Internet
such as CBT, CRT, and Mindfulness [8]. Self-assessment is a
major component of these Internet-based treatments [31]. Self-
assessment provides a possible way such that people suffering
from such morbidity can self-evaluate and manage their ill-
ness. The concept of self-assessment is not new in the health-
care domain and dates early back to 1977 where Alert Bandura
published a theory of self-assessment process incorporating
self-observation, self-judgment, and self-evaluation [31]. This
is to say, self-assessment comprises behavior observation,
behavior evaluation and a response to the evaluation or
progress performance. Self-assessment can be a convenient
and economical tool for understanding areas for improvement,
accelerating self-esteem and enhancing self-awareness [13].
A self-reporting mobile application can provide a virtual
therapist in everyone’s pocket and help overcome the issue of
stigma. Such stigma towards neurological or mental disorders
is one of the main causes why some ethnic minority group does
not seek and adhere to clinical treatments [19]. Moreover, the
study done in [43] shows people suffering from mental health
issues undergoes a fear of discomfort in facing a practitioner or
choosing services, fear of community and psychosis. However,
there is a need for self-assessment tool that a therapist can
access in situations where the patients require the involvement
of therapist in their treatment program. We address this issue
by providing a prototype application for self-assessment and
monitoring which is accessible by therapists.
People with different mental or neurological health prob-
lems require various types of screening, monitoring, and
treatment. Based on the patients monitoring condition, the
intervention should be adapted accordingly. But the current
Internet-based treatment programs are not flexible enough to
support such customization of the treatment plan based on
patients need. In the INTROMAT project, we envision an
architecture where we continuously monitor the mental health
conditions and provide a personalized intervention. In this
paper, we mainly focus on self-reporting part and leave the
discussion of adaptive treatment for future work.
III. ARCHITECTURE
In this section, we outline the cloud-based INTROMAT core
architecture which is based on SOA. The proposed architecture
is built around the healthcare process model as shown in Fig. 1.
The process model is based on evidence-based treatment
Assessment
Problem
Identification/
Treatment Plan
Intervention
Evaluation
Monitoring
followup
12
3
4
5
Fig. 1. Work-flow Model for mental or neurological clinical interventions
where the healthcare provider optimizes patient outcomes
through a series of interactions during which they (1) obtain
and examine patients’ perspectives and clinical information
(assessment); (2) identify the problem and create a treatment
plant (problem identification/treatment plan); (3) administer
clinical and behavioural interventions; (4) monitor patient
progress; and (5) evaluate the progress and ensure follow-
up if required. Clinical and behavioral intervention includes
evidence-based treatments including Cognitive Based Therapy
(CBT), Cognitive Remedial Therapy (CRT) and mindfulness.
Our SOA based solution is centered on the exchange of
information using HL7 FHIR resources. The architecture uses
services that are given by application components through
some communication protocol over a network [6]. The refer-
ence model for SOA [32] is an abstract framework that illus-
trates how significant entities communicate within a service-
oriented environment. SOA is a standard for creating and using
distributed components as services that may be maintained by
different ownership domain.
In the rest of the section we outline the main components
of the INTROMAT core architecture (Section III-A), discuss
communication between different components (Section III-B),
and discuss its quality attributes (Section III-C).
A. Components
Fig. 2 is a reference model [32] illustrating significant
components and relationship between them with their relevant
environments. As illustrated in the figure, the service provider
component will reside in the cloud to perceive the benefits
of distributed architecture. In this section, we outline the
most significant entities in the architecture and discuss the
communication between them.
1) Mobile Client: Mobile client facilitates data acquisition,
data transmission from wearable sensors. Moreover, mobile
apps can be used to provide evidence based treatments.
2) Web Client: A web client provides interfaces for login,
authentication, and authorization for practitioners, abilities
to create evidence-based treatments for patients and provide
a comprehensive dashboard to visualize lists of patients,
associated interventions, progress, and their activities. Both
mobile and web clients form the service requester/consumer
components of the architecture.
3) Authorization Server: The authorization server is an
OpenID connect [40] compliant web server with an ability to
authenticate users and grant authorization access tokens. More-
over, authorization server manages scopes and permission of
the clients, introspects token and requests for the resource
server.
4) SMART on FHIR: SMART on FHIR [36] performs the
role of service broker [6] and provides identity and access
management, access to data and launch sequence management.
SMART on FHIR incorporates OAuth2 [22] for authorization,
OpenID connect [40] for authentication, data models from
FHIR, and supports EHR UI integration through SMART
launch specification like EHR context and UI embedding for
web apps [34].
5) Resource Server: Resource server is an FHIR [2] com-
pliant web server with an ability to respond to FHIR REST
requests consuming access tokens granted by Authorization
Server. This component performs the role of service provider
[6] in the SOA architecture.
6) Health Provider Clients: Health Provider clients can be
external tools and applications, national resources, regional
clinical and core systems, as well as legacy systems. While
external EHRs following FHIR standards can communicate
with Resource Server inside the INTROMAT Core using
web services, the legacy systems require a middleware, that
transcribes legacy standard into FHIR standard.
7) Data Analysis: Data Analysis is one of the important
aspects of the INTROMAT Core Infrastructure. Data collected
and stored in the core will be available to researchers for anal-
ysis and to practitioners for medical interventions. The core
will provide services through IoT and application platforms
enforcing research services including data processing, artificial
intelligence, event processing and analytical services. Explain-
ing such research services and how it will be implemented is
not within the scope of this paper.
B. Communication Flow
SMART on FHIR specifies three different workflows includ-
ing contextless flow, EHR launch flow, SMART application
launch flow [36]. Our architecture supports all three types
of workflows, but for building self-reporting technology, we
adopt contextless workflow and is explained in Fig. 3. Devel-
opment and support for other workflows are entitled to the
extension of this work and further development.
The contextless communication between different compo-
nents is illustrated by the sequence diagram in Fig. 3 and the
steps involved as follows.
1) SMART applications (mobile/web clients) makes con-
formance request to the Resource Server.
2) The Resource Server responds back with valid confor-
mance endpoints (/token, /authorize, /manage).
Web Client
Authorization
Server
Resource
Server
DB
Mobile Client
Devices and
Sensors
External tools and
applications
Other
National
Resources
Regional clinical
and core systems
EHR Applications
Legacy System
Middleware
Health Provider Clients
INTROMAT CORE INFRASTRUCTURE
SMART on FHIR
apps
Fig. 2. Service-oriented architecture for adaptive healthcare system
3) The end users generate an OAuth 2.0 authorization grant
request (with valid scopes and headers) and redirect the
end users to the authorization server.
4) The authorization server validates the identity and cre-
ates access tokens for the valid clients.
5) The end user creates a request to the Resource Server
with Authorization headers (access token)
6) The Resource Server performs token introspection in
order to know more about the user and scopes.
7) The authorization server responds with valid scopes,
permissions and other token parameters including the
token type and token expiration.
8) The Resource Server makes a DB query with valid
requests.
9) The DB responds back with requested resources.
10) The Resource Server then forwards the requested re-
sources to the SMART applications.
C. Quality Attributes
In this section, we present the quality attributes [27] of
the proposed INTROMAT architecture to characterize the run-
time behavior of the system, its design and user experience.
To describe the quality attributes of the architecture, we have
adopted the notational convention keywords ‘MUST’, ‘MUST
NOT’, ‘REQUIRED’, ‘SHALL’, ‘SHALL NOT’, ‘SHOULD’,
‘SHOULD NOT’, ‘RECOMMENDED’, ‘MAY’, and ‘OP-
TIONAL’ in this section and are to be interpreted as described
in RFC 2119 [23].
1) Interoperability: Mandl and Kohane recognized impli-
cations of the inflexibility of the contemporary EHR sys-
tem and proposed a need for health platform with inherent
characteristics like liquidity of data, suitability of applications
(modularization and interoperability), based on open-standard
and supports diverse applications [34]. The same study em-
phasis interoperability is a key requirement for the success
of Healthcare Information Systems. One of the approaches
to make system interoperable is to follow open-standard for
defining syntactic and semantic meaning of information. We
adopt one of such open-standards HL7 FHIR [2], [3], which
is based on web standards including XML, JSON, HTTP,
OAuth, ontology-based, and OpenID connect. To enforce
interoperability, health records SHOULD be stored based on
FHIR standards. However, legacy EHR systems MAY choose
to utilize middleware that converts legacy data structure into
FHIR standard before a successful communication can be
established.
2) Security: The authorization server and the resource
server MUST be TLS-secured [24] and should be improved
using the contemporary practices mentioned in [24]. The
authorization server SHOULD issue short-lived tokens and
have a mechanism open to administrators and end users to
eliminate tokens in the case of a security conflict. Moreover,
the authorization servers MUST NOT use the value of the
launch code as a mechanism for transferring the authenticated
state. Using such a mechanism can lead to a session injection
attack and a session fixation attack [29].
SMART APPS
(Web/Mobile)
Authorization
Server
Resource
Server
1. Conformance Statement Request
2. Conformance endpoints
3. Redirect to Auth URL
to verify identity
4. Authorized - access token
5. Resource request with Auth Header (Access Token)
6. Token Introspection
7. Token Validated -
Scopes & Permissions
DB
8. DB Query
9. DB Response
10. Resource Response
Fig. 3. Communication between different components
3) Modifiability: Modifiability incorporates evolvability,
customizability, configurability, and extensibility [17]. SOA-
based architecture facilitates modifiability [32] by allowing
manageable growth of large-scale enterprise systems. These
enterprise systems or components are independent of vendors,
products, and technologies. This makes it easy for individual
components to be managed and modified. For example, the
Resource server in the architecture (described in Section III-A)
MAY update the HL7 FHIR version or create an additional
service that consumes the data and performs business intel-
ligence, without affecting other components. Similarly, the
authorization server MAY create a customized interface for
managing authorized clients, their scopes and permissions
without broadcasting its development complexity, structure
and patterns, and technological compliances to other compo-
nents. However, the constituting components MUST follow a
common standard for data storage and transmission.
4) Scalability: Scalability can be motivated by simplifica-
tion of the architectural components, distribution of services
across many components [32] and control of configurations
and interactions between constituting components [17]. SOA,
as mentioned in Section III-A, enforces scalability by orga-
nizing services into several components communicating over
a network. Each component of the architecture can be updated
and evolved in terms of hardware and software independent of
other components. For example, the data storage capacity of
resource server can be increased or decreased without affecting
other components.
5) Testability: As described in Sections III-C3 and III-C4,
each component of the architecture is independent and MUST
be tested independently to ensure their integral functional
components. In addition to unit testing, functional testing
and domain testing [18], detailed integration testing MUST
be performed in order to ensure different components can
communicate with one another.
IV. CAS E STU DY
To assess this architecture, we have implemented a proto-
type based on the proposed architecture for self-assessment
and monitoring of mental health problems. We use the proto-
type for our case study which include depression (MADRAS-S
[33], PHQ-9 [30], MDI [10]), anxiety disorder (GAD-7 [41]),
ADHD (ASRSV1.1 [15]) and bipolar disorder (ASRM [5]).
The self-assessment tool, developed using React Native3, is in
the form of mobile application for both IOS and Android.
In this section, we present the domain model of the self-
assessment mobile application and discuss how information
from the mobile application is exchanged from the device
to Resource Server in the INTROMAT core based on FHIR
standard.
3https://facebook.github.io/react-native/
A. Domain model
We use Diagram Predicate Framework (DPF) [39] for
domain modeling. DPF formalizes software development ac-
tivities such as metamodelling [4] and model transformations
based on category theory and graph transformations [9]. By
applying DPF we can formalize clinical guidelines and clinical
domain models at the different abstraction levels in form of
diagrammatic specifications. The diagrammatic nature of DPF
also facilitates visual representations of guidelines that can be
presented at different level of abstraction. A model in DPF is
represented by a diagrammatic specification S= (S, CS: Σ)
which consists of a graph Sand a set of constraints CS
specified by a predicate signature Σ.
The predicate signature is composed of a collection of
predicates, each having a name and an arity (shape graph). A
constraint consists of a predicate from the signature together
with a binding to the subgraph of the models underlying graph
which is affected by the constraint. Table I shows a sample
DPF predicate signature with three predicates: multiplicity,
injective and commutative. The table shows the arity, visu-
alization and the semantic interpretation of the predicates. In
Fig. 4 we present a portion of the domain model we used
for this case study. It illustrates a model M1which is typed
by metamodel M0. The model M1is constrained with some
predicates which specify the following constraints:
1) For each response instance, there exists a source refer-
ence and a questionnaire reference.
2) A questionnaire instance may have one or more ques-
tion items. An example of question item is shown in
Listing 1.
3) A question item may have zero or more answer options.
FHIR provides two ways to specifies options: answer-
ValueSet and answerOption. Listing 1 shows an example
of answerValueSet. In addition, each answer option has a
score which is used to get a total score from the response
of a user. The score of the options is created by the FHIR
concept of extension and StructureDefinition as shown
in Listing 2.
4) A question item must have at least one code; An answer
option must have at least one code. Listings 1 and 3
shows example of code.
5) A response instance must have the same number of
response items as the number of question items the
questionnaire instance has.
6) For every response item of a response instance, there
exists an answer.
A questionnaire instance has a way to interpret the total
score. This can be realized by the concept of extension.
Listing 2 shows an example of how a total score should be
calculated and interpreted. Moreover, it gives an evaluation
expression that can be used to calculate total score by means
of a path based navigation and extraction language, fhirpath4.
The constraints mentioned above are imposed on M1by
means of three DPF predicates: multiplicity, injective and
4http://hl7.org/fhirpath/
TABLE I
ASA MPL E DPF PR EDI CATE S IGN ATURE
Predicate p,
Symbol
Arity,
α(p)
Visualization
Semantic Interpretation
Multiplicity, [n..m]
Injective, [inj]
Commutative,
[=]
1 2
fX Y
f[n..m]
1 2
f3
gX Y
fZ
g
[inj]
1
4
f3
g
h
X
Z
fY
g
h[=]
2kWk
Ɐy ∈ f(x) where x ∈ X
there exists g(y’) where y’ ∈ y.
Ɐx ∈ X : m ≤ | f(x) |≤ n,
with 0 ≤ m ≤ n
Ɐx ∈ X : f;g = h;k.
Questionnaire Item AnswerOption
Code
Response
questions answerOptions
responseItems
[0..*]
[1..*]
Score
[=] answer
[inj]
[1..*]
Class DataType
attribute
reference
hasCode
hasScore
Model, M1
Model, M0
Patient
questionnaire
[1..1]
[1..1]
[1..*]
ResponseItem
source
Fig. 4. Example: Diagrammatic model with constraints in DPF
commutative (see Table I). We transform this DPF model
with constraints into an HL7 FHIR profile where Response,
Questionnaire, Patient classes are mapped to Questionnair-
eResponse, Questionnaire, and Patient FHIR resource type
respectively. The purpose of these constraints is to link and
validate the responses against the questions being asked by the
self-assessment tool. Therefore, the benefits of using an FHIR
profile on top of existing HL7 FHIR QuestionnaireResponse
is to maintain the validity of the self-assessment tools.
{
"linkId": "LittleInterest",
"code": [
{
"system": "http://loinc.org",
"code": "44250-9"
}
],
"text": "Little interest or pleasure in
doing things",
"type": "choice",
"required": true,
"answerValueSet": "http://loinc.org/vs/LL358
-3"
}
Listing 1. An example question item extracted from PHQ-9
"extension": [
{
"url" : "http://hl7.org/fhir/
StructureDefinition/questionnaire-
calculated-value",
"valueExpression" : {
"description" : "Minimal or none (0-4), Mild
(5-9), Moderate (10-14), Moderately
severe (15-19), Severe(20-27)",
"language" : "text/fhirpath",
"name" : "score",
"expression" : "QuestionnaireResponse.item.
repeat(answer.valueCoding.extension.
valueDecimal)"
}
}
],
Listing 2. Specifying scoring criteria using FHIR for PHQ-9
{
"code": "LA6569-3",
"display": "Several days",
"extension": [
{
"url": "http://hl7.org/fhir/
StructureDefinition/valueset-
ordinalValue",
"valueDecimal": 1
}
],
},
Listing 3. Specifying option score for PHQ-9
B. Depression
In this section, we present a self-assessment tool for screen-
ing patients with depression and a CBT for managing depres-
sion in adults. We implemented MADRS-S [33] using SMART
on FHIR. The MADRS-S score is used to evaluate the level
of depression of a patient. Based on the level of depression,
different CBT modules are created and assigned to patients.
One of such CBT training is offered by Helse Bergen, Norway
for depression management called eMeistring5. Anyone who
wants treatment using the eMeisting program must be referred
from a General Practitioner, specialist health service, other
doctor or psychologist. The patients are offered a CBT to man-
age mild to moderate depressions. The patients who receive
treatment through the eMeistring program are requested to give
their consent to use their information for doing research.
A CBT program consists of several modules and each
module includes reading, writing, listening or watching tasks.
The patients are encouraged to perform these tasks regularly.
Moreover, to better understand how patient activities affect the
mood, each patient is provided with a mobile client to create
a weekly plan consisting of a list of activities. The patients
can also annotate the activities with a positive or negative flag
to indicate their mood. These activities are shared with their
therapist who helps them personalize the CBT based on their
activities and mood.
5https://helse-bergen.no/emeistring
V. RE LATE D WOR K
Balsari and et al. [7] proposed a concept of federated,
patient-centric healthcare system after the Government of India
announced to provide automated healthcare to 500 million
citizens in India. The proposed healthcare system comprises
various features including API enabled, authentication and
authorization, open-standards (e.g., FHIR), block-chain based
and the ability to share the personal health record with
researchers and medical practitioner. The concept of trans-
forming collected data into EHR from wearable sensors and
using it for medical and research intervention is similar to
the INTROMAT project. However, in the INTROMAT project
we have broader scope focusing on the entire mental health
domain. We propose a cloud-based interoperable architecture
based on SOA that uses the personal health data in machine
learning to get deeper insights and use these insights for
creating a personalized Internet-based treatment.
The ALZCARE [37] project developed a mobile applica-
tion for screening dementia in elderly people. The proposed
prototype contains questionnaires tests, whose response could
be exported as XML in FHIR format. The proposed system
incorporated the web-based clinical Information System for
clinical settings and functionality for patient tracking. The
system is based on RESTful client-server architecture, while
our work is based on SOA. Moreover, we use the latest HL7
FHIR standard, support multiple mental health cases, and
provide service for a higher level of cognitive computation.
HABIT [28], funded by the New Zealand Ministry for
Business, outlines an IT infrastructure to provide appropri-
ate data management and scalable system for youth mental
health intervention. The paper reports the initial design of the
platform and intended HABITs platform requirements. While
the vision of having identity management, assessment storage,
reasoning, usage logging, and consent is similar to the INTRO-
MAT project, the approach of implementation, technological
eco-system preferences, assessment by a data-driven approach
using wearable sensors and cognitive computation, usage of
a recent version of the HL7 standard and SMART on FHIR
technology are distinct.
VI. CONCLUSION
This paper presents some results from the INTROMAT
project. We give an overview of the proposed solution in-
cluding a cloud-based SOA for self-assessment and evaluation
of mental or neurological morbidity, HL7 FHIR standard
questionnaire, a working mobile application prototype for
self-assessment and monitoring, a list of research challenges
and opportunities in the field. The proposed solution solves
the problem of interoperability and stigma on the one hand
and provides a self-assessment tool that is accessible and
available to patients and healthcare providers on the other.
Future work includes the use of mobile applications to collect
biological markers using wearable sensors. We will be using
these sensors data to perform a cognitive computation to
derive useful insights and utilize them to create personalized
evidence-based treatments. These treatments will be provided
through various types of applications such as mobile, VR
application, and voice assistant. In the INTROMAT project,
a team of researchers and IT professionals are developing
prototypes based on the proposed architecture and evaluation
of the architecture with respect to quality standards ISO 25010
are within the scope of the future work.
VII. ACKN OWLEDGEMENT
This publication is a part of the INTROducing Mental health
through Adaptive Technology (INTROMAT) project, funded
by Norwegian Research Council (259293/o70). INTROMAT
is a research and development project in Norway that employs
adaptive technology for confronting these issue. The opinions,
findings, discussions, recommendations, and conclusions illus-
trated in this chapter are those of the authors and do not reflect
the views of the funding agencies.
REFERENCES
[1] The Healthcare Industry and Data Quality (Cambridge, MA 02139-4307
U.S.A., 1996), Massachusetts Institute of Technology.
[2] Hl7 FHIR SMART app launch, Retrieved December 28, http://hl7.org/
fhir/smart-app- launch/.
[3] HL7 FHIR services, Retrieved on January 2, 2019, https://www.hl7.org/
fhir/services.html.
[4] AGG. Attributed graph grammar system, Retrieved on 7 January 2019.
[5] ALTMAN, E. G. , HE DEK ER , D., P ET ERS ON, J . L. , AND DAVI S, J. M.
The altman self-rating Mania scale. Biological Psychiatry (1997).
[6] AR SAN JAN I, A. CH5.3 - Service-oriented modeling and architecture.
IBM developer works (2004).
[7] BAL SAR I, S ., FO RTE NKO , A., B LAYA, J. A ., GROP PER , A., JAYAR AM,
M., MATTH AN , R., S AHASRANAM, R., S HANKAR, M., SAR BAD -
HIKARI, S. N., BIERER, B . E., MAN DL, K . D. , MEH EN DAL E, S. ,
AN D KHAN NA, T. Reimagining health data exchange: An application
programming interface-enabled roadmap for India. Journal of Medical
Internet Research (2018).
[8] BAR AK, A ., H EN, L ., B ONIEL-NISSIM, M ., A ND SHAPIRA, N. A
comprehensive review and a meta-analysis of the effectiveness of
internet-based psychotherapeutic interventions. Journal of Technology
in Human Services (2008).
[9] BAR R, M. , AN D WELL S, C. Category Theory for Computing Science.
1998.
[10] BE CH, P., RASMUSSEN, N. A. , OL SEN , L. R., NOERHOLM, V., A ND
ABILDGAARD, W. The sensitivity and specificity of the Major De-
pression Inventory, using the Present State Examination as the index of
diagnostic validity. Journal of Affective Disorders (2001).
[11] BE NDE R, D., AN D SART IPI , K. HL7 FHIR: An agile and RESTful
approach to healthcare information exchange. In Proceedings of CBMS
2013 - 26th IEEE International Symposium on Computer-Based Medical
Systems (2013).
[12] BE RGE R, T., B OE TTC HER , J. , AND CA SPAR , F. Internet-based guided
self-help for several anxiety disorders: A randomized controlled trial
comparing a tailored with a standardized disorder-specific approach.
Psychotherapy (2014).
[13] BE ST M, C ARSWELL RJ, A. S. Self-evaluation for nursing students.
Nurse Outlook. PubMed PMID: 2362838 (Jul-Aug 1990).
[14] BL OOM , D., AND CA FIER O,E.A. The global economic burden of
noncommunicable diseases. Pgda working papers, Program on the
Global Demography of Aging, 2012.
[15] DAIGRE, C. , RAMOS-Q UIROGA, J. A. , VALERO , S. , BOS CH , R., RON-
CE RO, C. , GO NZA LVO, B., NOGUEIRA, M ., AND CA SAS , M. Adult
ADHD self-report scale (ASRS-V1.1) symptom checklist in patients
with substance use disorders. Actas Espanolas de Psiquiatria (2009).
[16] EMEN. emen digital mental health european project, Retrieved on 17
December 2018.
[17] FIELDING, R. Architectural Styles and the Design of Network-based
Software Architectures. PhD thesis, 2000.
[18] FR EED MA N, R. S. Testability of software components. IEEE Transac-
tions on Software Engineering (1991).
[19] GA RY, F. A. Stigma: Barrier to mental health care among ethnic
minorities, 2005.
[20] GO SS, S TE PHE N,E.A.Technology in Mental Health. 2016.
[21] HALAMKA, J. D., A ND TRIPATHI, M. The HITECH Era in Retrospect.
New England Journal of Medicine (2017).
[22] HA RDT, D. [OAuth 2.0] The OAuth 2.0 Authorization Framework [RFC
6749]. RFC 6749 (2012).
[23] (IETF) , I. E. T. F. Key words for use in RFCs to Indicate Requirement
Levels. RFC 2119 (1997).
[24] (IETF) , I. E. T. F. Recommendations for Secure Use of Transport Layer
Security (TLS) and Datagram Transport Layer Security (DTLS). RFC
7525 (2015).
[25] INCORPORATED, D. Unstructured data in electronic health record (ehr)
systems: Challenges and solutions (report). Retrieved 17 december 2018,
17 December 2018.
[26] INTROMAT. Intromat, Retrieved on 1 January 2019, http://intromat.
no/.
[27] ISO/IE C 25000. ISO/IEC 25000:2005 Software Engineering - Software
product Quality Requirements and Evaluation (SQuaRE) - Guide to
SQuaRE, 2005.
[28] JI M WARREN, EWAN TEMP ERO , I. W. A. S. S . H. M . S., A ND ME RRY,
S. Experience Building IT Infrastructure for Research with Online Youth
Mental Health Tools. ASWEC (2018).
[29] KOL ˇ
SE K, M. Session fixation vulnerability in web-based applications.
Acros Security (2002).
[30] KRO ENK E, K ., SP IT ZER , R. L., AN D WILLIAMS, J. B. W. The PHQ-9.
Journal of General Internal Medicine (2001).
[31] LEVI NE, E. L. Introductory Remarks for the Symposium ”Organiza-
tional Application of self-appraisal and self-assessment: Another look”.
Personnel Psychology 33, 2 (jun 1980), 259–262.
[32] MAC KENZIE, C. M. , LASKEY, K., MCCABE, F., BROW N, P. F., AN D
MET Z, R. Reference Model for Service Oriented Architecture 1.0.
OASIS Standard, 2006.
[33] MADR S. Madrs, Retrieved on 7 January 2019.
[34] MANDEL, J. C., K RE DA, D. A ., M AND L, K. D., KOHANE, I. S. , AN D
RAMONI, R. B. SMART on FHIR: A standards-based, interoperable
apps platform for electronic health records. Journal of the American
Medical Informatics Association (2016).
[35] (NHE A), N. H. E . A. National health expenditure data, 2017.
[36] ON FHIR , S. Smart on FHIR, Retrieved December 28, http://www.
techterms.com/definition/social media.
[37] PROT OPAP PAS, V., TSIOURIS, K ., CHONDROGIORGI, M., TSIRONIS,
C., KONITSIOTIS, S., A ND FOTIADIS, D. I . ALZCARE: An information
system for screening, management and tracking of demented patients.
In 2016 38th Annual International Conference of the IEEE Engineering
in Medicine and Biology Society (EMBC) (2016).
[38] ROTO NDI , A. J ., SINKULE, J ., HAAS , G. L., SPRING, M . B.,
LITSCHGE, C. M., NE WH ILL , C. E., GA NGU LI , R., AND ANDERSON,
C. M. Designing Websites for Persons With Cognitive Deficits: Design
and Usability of a Psychoeducational Intervention for Persons With
Severe Mental Illness. Psychological Services (2007).
[39] RUT LE, A . Diagram Predicate Framework: A Formal Approach to
MDE. PhD thesis, 2010.
[40] SA KIM UR A, N. , BR ADL EY, J., B . JO NES , M., MEDEIROS, B., A ND
MORTIMORE, C. Final: OpenID Connect Core 1.0 incorporating, 2014.
[41] SP ITZ ER , R. L. , KRO ENK E, K., WILLIAMS, J . B., AND L¨
OWE, B. A
brief measure for assessing generalized anxiety disorder: The GAD-7.
Archives of Internal Medicine (2006).
[42] SUCALA, M., SCHNUR, J . B., CONSTANTINO, M. J., MIL LER , S. J .,
BRACKMAN, E. H. , AND MONT GO MERY, G. H . The therapeutic
relationship in E-therapy for mental health: A systematic review. Journal
of Medical Internet Research (2012).
[43] SW EEN EY, A., G ILLARD, S., WYK ES , T., AN D ROSE , D. The role of
fear in mental health service users’ experiences: a qualitative exploration.
Social Psychiatry and Psychiatric Epidemiology (2015).
[44] WHO. World health report, 2018, http://www.who.int/whr/2001/media
centre/press release/en/.
[45] WOO DS , S. S. , SC HWARTZ , E. , TUE PK ER, A ., PRES S, N. A., NAZI ,
K. M., TU RVEY, C. L ., AND NICHOL, W. P. Patient experiences with
full electronic access to health records and clinical notes through the
my healthevet personal health record pilot: Qualitative study. Journal
of Medical Internet Research (2013).
