A Conceptual Data Model for Health Information Systems

Abstract

The development of Health Information Systems
based on dual models allows modifications to be conducted in
the layer of archetypes, reducing dependencies on software
developers. However, we identified a lack of conceptual
models to represent two-level database entities. This paper
proposes a novel conceptual data model, called ArcheER,
which is a dual modeling approach and aims to reduce
redundant entities and guarantee the creation of unique
electronic health records. ArcheER is an extension of the
Entity-Relationship model and is based on archetypes. A
CASE modeling tool based on ArcheER is outlined. Finally, to
illustrate the key features of the proposed model, an ArcheER
conceptual schema built for a legacy system is discussed, and
results collected from a test with 18 human subjects are
reported. Results indicated a reduction of 83,35% in the
representation of redundant entities and a gain of 78,9%
concerning the modeling of entities characterizing knowledge.

Full text

A Conceptual Data Model for Health Information
Systems
André Magno Costa de Araújo1, Valéria Cesário Times1, Sérgio Castelo Branco Soares1
1Center for Informatics, Federal University of Pernambuco, Recife, Pernambuco, Brazil
Abstract - The development of Health Information Systems
based on dual models allows modifications to be conducted in
the layer of archetypes, reducing dependencies on software
developers. However, we identified a lack of conceptual
models to represent two-level database entities. This paper
proposes a novel conceptual data model, called ArcheER,
which is a dual modeling approach and aims to reduce
redundant entities and guarantee the creation of unique
electronic health records. ArcheER is an extension of the
Entity-Relationship model and is based on archetypes. A
CASE modeling tool based on ArcheER is outlined. Finally, to
illustrate the key features of the proposed model, an ArcheER
conceptual schema built for a legacy system is discussed, and
results collected from a test with 18 human subjects are
reported. Results indicated a reduction of 83,35% in the
representation of redundant entities and a gain of 78,9%
concerning the modeling of entities characterizing knowledge.
Keywords: Novel Software Tools, Conceptual Data
modeling, Health Information Systems, Archetypes.
1 Introduction
Conceptual modeling is an important activity for
designing a database. The conceptual scheme is a concise
description of data requirements specified by the application
designer, including detailed descriptions about types of
entities, relationships and constraints [1]. Thus, the artifacts
generated from the conceptual data modeling are important
elements in building database systems. Currently, most Health
Information Systems (HIS) are built using traditional database
modeling technologies [2], in which both information and
knowledge concepts are represented in single level computer
systems using conventional data models. However, HIS must
handle a large number of concepts that often change or are
specialized after a short period of time and, consequently,
HISs based on such models are expensive to maintain.
Several research projects and many applications have
been developed from the specifications of the openEHR
system architecture and the concept of archetypes [3-8]. The
Open Electronic Health Record (openEHR) software
architecture for HIS is aimed at developing an open,
interoperable and computational platform for the Health
domain [9]. This architecture separates generic information
that represents the structures of the Electronic Health Records
(EHR) and demographic characteristics of the patients of a
reference model, from the constraints and standards
associated with the clinical data of a given specific domain,
which composes the knowledge model. An archetype consists
of a computational expression that is based on the reference
model and is represented by domain constraints and
terminologies [3] (e.g. data attributes of a blood test), while
templates are structures used to group archetypes for allowing
their use in a particular context of application, and are often
associated with a graphical user interface. On the other hand,
some authors have already proposed extensions of traditional
conceptual modeling techniques to represent HIS
applications. However, these extensions do not model EHR,
do not provide dual modeling constructors and are not based
on archetypes. In fact, little attention has been devoted to the
investigation of the following issue: which conceptual
constructors are needed to model the two-level database
entities of HIS applications?
This paper proposes a novel conceptual two-level data
model, named ArcheER, for helping database designers with
the modeling of HIS applications. ArcheER is an extension of
the Entity-Relationship (ER) model [1] and is based on the
openEHR definitions [3]. It also comprises a set of modeling
constructors with graphical representations for building health
information conceptual schemas and a set of knowledge-level
constructors that are based on archetypes. The ER model was
chosen because it is simple and widely used in both academia
an organizations for the development of DB applications, and
because it is capable of providing an abstraction of
implementation details as well as being easily mapped to
DBMS logical data models. Another contribution of this paper
is related to the development of a modeling tool for ArcheER.
The main goal of such tool is to provide application designers
with computer support to assist in the database modeling
activities of healthcare applications.
The dual modeling approach has been used by several
researchers and is not unique to archetypes [10]. However, in
this paper the focus lies on the use of dual modeling based on
the concept of archetypes, since [8],[11],[12] reported the use
of this approach as essential to achieve interoperability and
standardization of EHR.
2 Related work and motivation
Späth and Grimson [8] used the openEHR specification
to map the structure of an EHR into a proprietary database
system. They examined the reuse of archetypes available in
the repository of openEHR by specializing some of them and
then proposing a new set of archetypes to support biomedical
knowledge discovery. To achieve this, they studied the
database schema to reorganize it according to the concept of
archetypes, by mapping each field of the database to an
archetyped element. Some difficulties were reported while
doing so, including a lack of consolidated modeling tools and
lack of mechanisms to determine overlapping archetypes as
236 Int'l Conf. Software En
g
. Research and Practice
|
SERP'16
|
ISBN: 1-60132-446-4, CSREA Press ©

well as solve the semantic conflicts that may appear when
archetypes are mapped to the chosen DBMS.
Bernstein et. al [6] conducted a study in Denmark about
the patterns of the development of healthcare computer
systems. This research indicated that the Danish healthcare
systems were based on several information models and
heterogeneous technology platforms, developed by different
software vendors. Besides, it showed the need for replacing
traditional standards of software development in the Health
domain, and reported the importance of the openEHR
architecture as a new pattern for the development of computer
systems for healthcare.
Despite the development of HIS based on the openEHR
specifications being a multidisciplinary research area, (there
are already varied studies published by the scientific
community [13-15]) there is a consensus that the openEHR
architecture definitions have to evolve and address some open
problems [8]. This paper points out that the difficulty in
applying the openEHR concepts to a given problem domain
for enabling the two-level data modeling is due to the lack of
a methodology to express which are the data requirements
requested by users and how these might be modeled.
This paper goes one step beyond previous works by
specifying an ER-based conceptual data model enabling the
definition of which archetypes, patient demographic
properties, hospital administrative information and clinical
data are important and should be taken into account during the
conceptual modeling of a healthcare database application. The
main concepts of the ArcheER modeling proposal are detailed
in the following section.
3 The ArcheER conceptual model
ArcheER is a conceptual data model that aims to allow
the specification of a health application domain using the
concepts of dual modeling. The proposed set of ArcheER
modeling constructors extends the basic ER modeling
elements with a set of archetyped components listed in Figure
1. ArcheER represents entity types alongside their
relationships and properties, and a conceptual schema is
composed of hospital administrative data and archetyped
information. An archetyped entity denotes a set of entities that
must have a set of generic data structures. Each of those
structures is defined as an attribute of those entities, and
organizes data through data structuring elements that are
neither dependent of the DBMS storage format nor of the
application development technology.
In order to model relationships between archetyped
entities, and a relationship between a conventional entity and
an archetyped entity, ArcheER proposes a new relationship
type, called Party Relationship. It is worth noting that an
administrative entity may not be modeled as an archetyped
entity, i.e. it may not be represented as an entity type with a
set of generic data structures, therefore not being able to have
party relationship associations with other entity types.
However, this must not be mandatory and an administrative
entity may benefit from the use of generic data structures as
well.
Figure 1. The Main Modeling Components of ArcheER
One of the advantages of ArcheER is the elimination of
data redundancy by defining a uniqueness constraint based on
the concept of roles played by the actors being modeled.
According to this constraint, every instance of a relationship
involving the demographic information of an actor and an
entity of the type clinical care or administrative must be
modeled as a relationship between a role played by the actor
and the entity type clinical care or administrative.Observe
that the use of openEHR definitions requires an understanding
of which actors should be considered while modeling an
application domain, how they relate to each other, how they
play their roles and which capabilities they have. This
understanding is important and must not be neglected.
However, this cannot be enforced automatically by the
DBMS nor can it be seen as a data model constraint to
guarantee a unique EHR.
The ArcheER constructors inherited from ER are mostly
used for modeling the operational aspects of a hospital
organization (i.e. entity type Administrative), while the
archetyped entity types are mainly concerned with the
representation of (i) metadata and the context of the
application being modeled (i.e. entity type Structuring); (ii)
patient’s demographic information (i.e. entity type
Demographic.); (iii) clinical data (i.e. entity type Clinical
Care) and (iv) constraints, terminologies of health area,
internal coding of vocabulary and textual information given
by a domain specialist (i.e. entity type Knowledge). While the
first three entity types represent the information level of the
dual modeling approach, the last type of entity and its
specializations compose the second level and are useful for
generating knowledge at runtime. The definition of each type
of constructor is given below.
3.1 Structuring constructors
The ArcheER data model provides the following
modeling constructors for structuring health care information:
Composition, whose attributes represent the metadata of an
ArcheER conceptual schema; and Section, which organizes
the remaining modeling constructors of ArcheER into themes
Int'l Conf. Software En
g
. Research and Practice
|
SERP'16
|
237
ISBN: 1-60132-446-4, CSREA Press ©

or subjects that represent the context of the application being
modeled.
3.2 Demographic constructors
The modeling of demographic information requires the
identification of actors who compose the hospital application
domain, and the definition of their roles and capabilities in the
health area. For modeling actors, ArcheER specifies the
following set of constructors of demographic entities that
represent the specialization of an actor in a Health domain.
The definition of each type of demographic entity is given
below:
Agent: Expresses a software agent or any device that
communicates with the healthcare application.
Person: Corresponds to an entity type that represents
a generic description of people who are part of the
context of the application being modeled.
Group: Models parts of the real world that interact
with each other and are grouped to represent the
purpose of being together.
Organization: Denotes an abstraction of all
companies involved in a health application domain.
Role: Represents a generic description of a role
played by a given actor.
Capability: Models the qualification of an actor to
play a certain role in a healthcare domain.
Party Identity: Indicates how an actor is identified in
a healthcare application, and allows an actor to be
identified in several ways.
Contact: Expresses the possible ways of contacting
an actor.
Address:Indicates how the contact information of
actors is formatted.
3.3 Health care constructors
The ArcheER modeling constructors that represent
health care information are in charge of defining all the
semantics of EHR - hence, the information they model
represent the main target to be archetyped. For the modeling
of clinical care information, ArcheER proposes the following
entity types:
Admin Entry: Expresses all the administrative
information of patients in the modeling of EHR.
Note that this entity type concerns the modeling
requirements of the patient´s administrative
information that compose the EHR of the patient and
does not refer to administrative aspects of a service
provider organization in health. In this work, for the
modeling of these administrative issues of a health
service organization, we assume that the use of
traditional ER constructors will suffice, thus, in fact,
only clinical care and demographic information are
modeled using archetypes.
Observation: Represents any event or clinical status
associated with the patient.
Instruction: Expresses all future actions to be
administered to the patient.
Activity: Specifies the activities of an instruction.
Action: Specifies the actions of an instruction.
Evaluation:Represents general information about
the clinical care of patients, based on diagnosis,
assumptions, risk assessments and observations.
3.4 Knowledge constructors
The entity type Knowledge expresses the terminology
and constraints related to attributes, also called generic data
structures. This entity allows the second level of dual
modeling to be displayed in an ArcheER schema.
Furthermore, ArcheER adds a new constructor of
relationships, called Knowledge Relationship, to express
associations between generic data structures and instances of
the entity type Knowledge.
ArcheER extends the definition of ER relationship types
to enable the creation of direct relationships between the
generic structure attributes of archetyped entities and the
entity type Knowledge. The following relationship
cardinalities are considered by our ArcheER proposal: 1:1, 1:
N and M: N.
The entity type Knowledge is specialized in the
following entity types: Free Text, Internal Code and
Terminology, which are directly related to the generic data
structures through the Knowledge-type relationship. These
specializations model knowledge of a given Health domain –
in other words, they represent the second level of ArcheER
dual modeling. The entity type Free Text represents free text
information given by the domain specialist, while the entity
type Internal Code denotes codes of a health vocabulary (e.g.
procedures, billing tables, international classification of
diseases) used for the exchange of information between EHR
applications. Lastly, the entity type Terminology represents
terms and concepts designed to standardize, promote and
disseminate health knowledge.
3.5 Data entry constructors
The ArcheER modeling constructors used to define
attributes are called data entry constructors, since such
attributes comprehend entries with any kind of data that are
represented by generic data structures. Hence, generic data
structures are defined as attributes of archetyped entities of
ArcheER. For each element of these data structures, a data
type must be specified. An entry may have a single clinical
statement (e.g. a short description about the history of the
current illness), or otherwise contain a large amount of data
(e.g. the list of values of a laboratory test, tabular data
reporting a hospital infection, a hierarchical structure
containing all procedures, materials and medications of a
patient's hospital bill, an entire microbiology result or a
psychiatric examination note). An entry defines the semantics
of multiple formats of data which are properties of the
archetyped entities of ArcheER.
238 Int'l Conf. Software En
g
. Research and Practice
|
SERP'16
|
ISBN: 1-60132-446-4, CSREA Press ©

For modeling generic data structures, the ArcheER
model provides the following types of attributes:
ITEM_SINGLE: represents a data structure with a single
element; ITEM_LIST: represents a list of data items or values,
where each element of this list may assume a value or not,
may be referenced by a name and may have an index to
indicate its position within the list; ITEM_TREE: models a
data structure that is logically represented as a tree; and
ITEM_TABLE: defines a data structure with lines and
columns, where the line represents the specification of an
element, and the column the information value.
3.6 ArcheER constraints
A set of constraints aiming at ensuring the uniqueness
of the EHR is specified in Object Constraint Language (OCL)
[16] notation. Thus, the relationship between demographic
and clinical care entities and the relationship between
demographic and administrative entities of the patient are
restricted by two constraints, respectively: Context Clinical
Care inv: Health->forAll (oclType=Role) and Context
AdminEntry inv: Health->forAll (oclType=Role or
OclType=Administrative). Consequently, each instance of
entities CareEntry and AdminEntry is related with
demographic information of patients only by means of the
roles played by the actors. The benefit of defining constraints
over these relationships using the concept of roles is that,
while actors of a Health domain are modeled as generic
entities, their specific characteristics are represented as roles.
This ensures the conceptual modeling of the uniqueness of
demographic information, since new instances of a given actor
are created only through the roles played by him.
In order to model actors, roles and capabilities, we
propose four constraints, which are aimed at enforcing the
uniqueness of EHR and explained as follows. Constraint
Context Actor inv: Actor.allInstances->forAll (ar | self.Actor
< > ar.Actor implies self < > ar) specifies the actors'
uniqueness constraint and enforces that each instance of an
actor entity of ArcheER is unique. The constraint Context
Actor inv: self.Actor_Role->notEmpty() implies
self.Actor_Role->forAll (r1 | self.Role < > r1.Role implies
self < > r1) indicates that each actor is not allowed to have
two instances with the same role, while constraint Context
Role inv: self.Actor_Role-> includes (self.Actor) guarantees
that in order to create a new instance of a given role, a
corresponding instance of an actor must exist. In addition, the
constraint Context Capability inv: self.Role_Cap-> includes
(self.Role) defines that, for each instance of the entity
Capability, a corresponding instance of the entity Role must
exist. Entity Address models the details of each instance of the
entity Contact; thus, an instance of entity Address can exist
only if there is a corresponding instance of entity Contact.
This is enforced by the constraint Context Address inv:
self.address-> includes (self.Contact).
The entity Administrative may be related to demographic
information (i.e. to instances of the entity Demographic) and
to clinical concepts as well (i.e. to instances of the entity
ClinicalCare). For relationships with demographic
information, a constraint is specified to ensure that this
relationship is always established through instances of the
entity Role, while for the relationship with a clinical care
entity, there must be an entity AdminEntry. The constraint
Context Administrative inv: demographic-> forAll
(oclType=Role or oclType=AdminEntry) guarantees this.
3.7 The ArcheER case tool
ArcheERCASE is a computational modeling tool that
builds conceptual data schemas based on ArcheER. It is a
graphic design software, not a technology-oriented tool, since
both the data schema elaborated using this tool and the
configuration metadata of this tool are stored in XML format.
Figure 2. The ArcheERCASE Tool
The main goal of ArcheERCASE is to provide
application designers with computer support to assist in the
database modeling activities of healthcare applications.
Details about ArcheERCASE, including the system prototype
architecture, the ArcheERCASE Data Dictionary and the
Graphic Module of ArcheERCASE can be found at
www.r2asistemas.com.br/ArcheER. Figure 2 depicts the
graphic environment of this tool alongside graphic notations.
4Results
4.1 Experimental design
To validate ArcheER, we conducted two data modeling
experiments with two distinct set of human subjects. In both
experiments, nine Brazilian professionals with at least two-
year experience in conceptual modeling and database design
were asked to build two conceptual schemas to model a
problem domain. The experiment is based on a hospital
scenario located in Northern Brazil, for urgency care. The full
description of the problem domain is available at
www.r2asistemas.com.br/archeER.
The goal of this research is not to determine the best
conceptual data model for HIS applications, but to better
understand some important differences between ArcheER and
ER conceptual models.
Int'l Conf. Software En
g
. Research and Practice
|
SERP'16
|
239
ISBN: 1-60132-446-4, CSREA Press ©

To accomplish that, we computed the time each
participant took to complete a given modeling task. Also, for
each conceptual data schema generated, we observed whether
the uniqueness of EHR was represented, and whether
terminologies used in health standards were identified and
modeled. Observing the software artifact produced by the
ArcheER approach (i.e. each ArcheER conceptual scheme),
our experiments measure the difference with respect to the ER
model in the following aspects: (i) elapsed time for building a
conceptual data scheme; (ii) number of redundant entities
produced by each conceptual data schema; and (iii) number of
entities that represent terminologies and standards in health of
each conceptual data schema produced.
To perform our experiments, each selected participant
received the following support instruments: a) instruction
about the ER model, b) instruction about the ArcheER
approach, c) record sheet, and d) description of the problem
domain. In our experiments, the following hypotheses were
considered: Hypothesis 1 (H1): The use of the ArcheER
approach reduced the time needed to build a conceptual data
scheme of a problem domain. Hypothesis 2 (H2): The use of
ArcheER warrants the uniqueness of EHR. Hypothesis 3 (H3):
ArcheER allows the identification of terminologies and health
standards used in a given problem domain.
The variables considered in our experiments are: F1 –
Conceptual Modeling Technique for building data schemes;
Level of factor T1: Conceptual scheme designed with the ER
model (F1→T1); Level of factor T2: Conceptual scheme
designed with the ArcheER model (F1→T2). The metrics
collected in our experiments are TSB – Time spent for
building the conceptual data scheme, QRE – Quantity of
redundant entities and QEK – Quantity of entities
characterizing knowledge (terminologies and standards). The
subjects selected to take part in this study were divided into
two working groups (i.e. G1 and G2), chosen by lottery. To
eliminate the influence of previous experience of the selected
subjects, we used the design of Latin Square experiment 2x2.
Considering that Exp1 and Exp2 correspond to the
experimental objects that were randomly attributed by lottery
to the variables, the experiment design is described in Figure
3a.
Figure 3. Design and statistical results
For interpreting the raised hypothesis, we have used the t
distribution test. This test is often chosen when the average
population is less than 30 and there is a normal (or
approximately normal) distribution. For this work, the
distribution of t sampling with n-1 degrees of freedom was
adopted. Figure 3b has the values of each metric computed
after the application of statistical tests.
Figure 4. Results of experimental design
Results indicated that the time for building a conceptual
data scheme is similar for both modeling approaches.
However, for the other two metrics, it is possible to say that,
as shown in Figure 4a, the quantity of redundant entities in
conceptual schemes designed by the ER Model is greater than
the respective number of redundant entities in schemes
designed using ArcheER. Moreover, the quantity of redundant
entities was reduced in 77.5 % for group 1 and in 89.2 % for
group 2 with the adoption of ArcheER approach. Actually, in
conventional modeling, for each new role an actor plays in a
health domain, new instances are created to represent it, which
possibly generates data redundancy in the DBMS – i.e., if a
doctor needs to be represented as a patient, a new instance of
patient is created by storing information about this person
redundantly in the EHR.
Regarding the quantity of entities that denote knowledge,
the ArcheER approach identified more entities than the ER
approach, as shown in Figure 4b. This increase represents a
gain of 75.7% for group1 and 82.1% for group 2. As the use
of health terminologies and standards is common in the Health
domain, the previous identification of terminologies and
standards during the conceptual modeling phase can provide a
better understanding of which archetypes are needed for an
application to generate knowledge during runtime.
4.2 Modeling an outpatient emergency with
ArcheER
In this section, we describe the main difficulties
encountered in modeling HIS using traditional approaches,
and later we comment on the advantages of modeling HIS
using ArcheER. For the sake of didactics, we present in
Figure 5 a data schema extracted from a HIS produced by
manufacturers of a Health Software in Brazil. This HIS
concerns an ambulatory emergency that is performed daily at
a Hospital located in Northern Brazil.
240 Int'l Conf. Software En
g
. Research and Practice
|
SERP'16
|
ISBN: 1-60132-446-4, CSREA Press ©

Observing the data schema, it is possible to see that the
initial difficulty is due to the variety of roles played by the
actors in a Health domain, such as workers of a hospital,
physicians responsible for patient care, nurses, and other
health professionals that sometimes act as health care
providers, but occasionally might be seen as a patient who
receives care themselves. Besides, the current approaches for
database modeling do not provide any constraints to limit this
redundancy. Actually, in conventional modeling, for each role
played by an actor in a Health domain, new instances are
created to represent it, and thus data redundancy may be
added to the DBMS.
Figure 5. Legacy Data Schema
It is possible to see, in Figure 5, that entities
representing demographic information (i.e. Doctor, Hospital,
Hospital_Staff, Patient and Nursing_Staff) reflect this
modeling practice. In other words, if an actor plays a role,
new instances are created for each entity, making their
information redundant in the EHR.
In the ArcheER model proposal, actors are modeled in
their more generic way, with new instances being created
from the roles played. Therefore, an actor may have several
roles in an organization and still keep its record unique. As
shown in Figure 6, the entity Person_EHR represents the most
generic characteristics of the actor, while entities
Hospital_Staff, Patient, Nursing_Staff and Doctor represent
the roles played by this actor in EHR. To play a role, the actor
must have training that qualifies them to perform the referred
role –in this case, the Council entity illustrated in Figure 6
represents the professional record that the actor needs to have
in order to play the role of a physician.
Besides the roles played in a Health domain, an actor
may take the form of an organization that provides health
services, or that is directly involved in the application context.
In this sense, the entity Hospital of Figure 6 represents the
organization responsible for providing services to the patient.
This case shows that the advantages of the ArcheER model go
beyond the input of demographic information into the EHR
modeling: due to the specified constraints, a demographic
entity may only be related to other concepts of EHR (i.e.
clinical care, administrative) by means of a role played. In this
case, if necessary, only new instances of the roles played by
an actor are created, keeping its most generic characteristics
preserved, thus ensuring the uniqueness of EHR. As Figure 6
shows, all relationships having an entity that represents patient
care (i.e. OutPatient) are established by means of the roles
identified in the described application.
Figure 6. Demographic Conceptual Schema
Figure 7 portrays entities that model clinical care,
administrative and knowledge information. Entities Snomed,
List_Presc and ICD show the knowledge modeled in the
ArcheER conceptual schema. The first entity expresses the
terminology and constraints of health care regarding the
construction of laboratory examinations, while the entity
Item_Presc models an internal coding that standardizes the
prescription items of a hospital. Finally, the entity ICD
represents the terminology used to define the patient diagnosis
internationally.
Fig. 7.Clinical Conceptual Schema
For modeling clinical care information, ArcherERCASE
provides the following entities types: Admin_Entry,
Observation, Evaluation, Instruction, Action and Activity. All
those types represent abstractions of clinical concepts found in
Int'l Conf. Software En
g
. Research and Practice
|
SERP'16
|
241
ISBN: 1-60132-446-4, CSREA Press ©

a Health domain. One can see in Figure 7 that the entities
denoting the concepts of patient clinical care are Exams,
Prescription, History, Evolution and Clinical_Information.
The importance of having modeling constructors that
represent such concepts is justified by the following aspects:
firstly, it helps in the understanding of how to identify and
classify EHR clinical information, and secondly, each instance
of a clinical care entity represents a potential archetype that
may be reused.
5 Conclusions
This paper proposed a novel conceptual data model,
named ArcheER, based on archetypes and dual modeling. The
benefits of using dual modeling constructors in the conceptual
modeling of health information systems have not been studied
so far. The modeling constructors that compose ArcheER and
the modeling technique selected for diagrammatic
representation were chosen from openEHR specifications.
ArcheER is an extension of the ER data model because
this model has been recognized in literature as a simple and
efficient approach for the elicitation of data requirements,
providing the abstraction required for representing the
concepts of archetypes through its graphical notation. As
main contributions, we highlight a reduction in the
representation of redundant entities and a gain concerning the
modeling of entities characterizing knowledge. Also, a CASE
modeling tool based on ArcheER was presented and a set of
OCL constraints was specified, illustrating how the ArcheER
model provides uniqueness to the EHR. The specification of
semi-automatic generation of archetypes in ADL from the
data requirements modeled by an ArcheER conceptual
schema is a possibility for future research.
Acknowledgment
This work was partially supported by Fundação de
Amparo à Ciência e Tecnologia do Estado de Pernambuco
(FACEPE), under the grants APQ-0173-1.03/15 and IBPG-
0809-1.03/13.
6 References
[1] Elmasri R, Navathe S. B. “Fundamentals of Database
Systems”. Addison-Wesley, 6th ed, 2011.
[2] Marco E, Thomas A, Jorg. R, Asuman D, Gokce L. “A
Survey and Analysis of Electronic Healthcare Record
Standards”; ACM Computing Surveys, pp. 277–315, 2005.
[3] Mu-Hsing K, Tony S, Andre W.K, Elizabeth M.B,
Daniel K.G. “Health big data analytics: current perspectives,
challenges and potential solutions”; Int. J. Big Data
Intelligence, pp.114-126, 2014.
[4] Späth M. B, Grimson J. “Applying the archetype
approach to the database of a biobank information
management system”; International Journal of Medical
Informatics, pp. 1-22, 2010.
[5] Chen R, Klein G. O, Sundvall E, Karlsson D, Åhlfeldt H.
“Archetype-based conversion of EHR content models: pilot
experience with a regional EHR system”; BMC Medical
Informatics and Decision Making, pp. 9-33, 2009.
[6] Garde S, Hovenga E, Buck J, Knaup P. “Expressing
clinical data sets with openEHR archetypes: A solid basis for
ubiquitous computing”; International Journal of Medical
Informatics, pp.334–341, 2007.
[7] Lezcano L, Miguel A. S, Rodríguez S. C. “Integrating
reasoning and clinical archetypes using OWL ontologies and
SWRL rules”; Journal of Biomedical Informatics, pp. 1-11,
2010.
[8] Oriol X, Teniente E, Tort A. “Fixing Up Non-executable
Operations in UML/OCL Conceptual Schemas”; In: LNCS.
Vol.8824, pp. 253-496, 2014.
[9] Dinu V, Nadkarni P. “Guidelines for the Effective Use
of Entity-Attribute-Value Modeling for Biomedical
Databases”; International Journal of Medical Informatics, pp.
769-779, 2007.
[10] Bernstein K, Bruun R. M, Vingtoft S, Andersen S. K,
and Nøhr C. “Modelling and implementing electronic health
records in Denmark”; International Journal of Medical
Informatic, pp. 213-220, 2005.
[11] Jeffrey A. L, Jeffrey L. S, Blackford M. “Method of
Electronic Health Record Documentation and quality of
primary care”; J Am Med Inform Assoc, pp.1019-1024, 2012.
[12] Georg D, Judith C, and Christoph R. “Towards plug-
and-play integration of archetypes into legacy electronic
health record systems: the ArchiMed experience”; BMC
Medical Informatics and Decision Making, pp.1-12, 2013.
[13] Bernd B, “Advances and Secure Architectural EHR
Approaches”; International Journal of Medical informatics,
pp.185-190, 2006.
[14] Martínez C. C, Menárguez T. M, Fernández B. J. T,
Maldonado J. A. “A model-driven approach for representing
clinical archetypes for Semantic Web environments”; Journal
of Biomedical Informatics, pp.150–164, 2009.
[15] Buck J, Garde S, Kohl C. D, Knaup G. P. “Towards a
comprehensive electronic patient record to support an
innovative individual care concept for premature infants using
the openEHR approach”; International Journal of Medical
Informatics, pp.521-531, 2009.
[16] Arlow J, Neustadt I. “UML 2 and the Unified Process:
Practical Object-Oriented Analysis and Design”. Addison-
Wesley, 2nd ed, 2005.
242 Int'l Conf. Software En
g
. Research and Practice
|
SERP'16
|
ISBN: 1-60132-446-4, CSREA Press ©