SEMANTIC FRAMES
A Way for Automating the Management of Medicinal Documents
Juha Puustjärvi
Helsinki University of Technology, Innopoli 2, Tekniikantie 14, Espoo, Finland
Leena Puustjärvi
The Pharmacy of Kaivopuisto, Neitsytpolku 10, Helsinki, Finland
Keywords: Medical applications, Semantic web, Ontologies, Knowledge management, Semantic interoperability.
Abstract: Knowledge management concerns with acquiring, accessing and maintaining knowledge within an
organization. Knowledge management is also important because organizations view internal knowledge as
an intellectual asset from which they can draw greater productivity, create new value, and increase their
competitiveness. We have investigated the management of knowledge in pharmacies. It has turned out that
the volume of information coming in from a variety of information sources such as pharmaceutical
companies, medicinal wholesalers, social insurance institutions and other authorities is increasing all the
time. Further, the various formats such as paper, fax, email, and a wide variety of multiple electronic media
formats are still complicating the information management. In order to alleviate this problem, we have
introduced the notion of semantic frames, which are included in incoming XML-documents. The frames
specify how to integrate the incoming document into the medicinal ontologies and taxonomies in the
pharmacy system. Further, as our used ontology models the relationships of the incoming documents and
the daily duties, the integration of the documents and daily duties can be automatically done. The gain of
this approach is that the documents (medicinal instructions) are provided just-in-time, and tailored to their
specific needs. An essential prerequisite of our approach is that the healthcare organizations that send the
documents and the receiving pharmacies have to commit to the same medicinal ontologies, i.e., they have to
use the same vocabulary in specifying and interpreting the semantic frames.
1 INTRODUCTION
Healthcare is a field where the fast development of
drug treatment requires specialized skills and
knowledge. As a result the amount of new medicinal
information increases all the time.
The problems arising of increased medicinal
information are discussed in many practitioner
reports and public national plans, e.g., in (Hyppönen
et al., 2005; Dwivedi et al, 2007; Batenburg and
Broek, 2008); Ghani et al., 2008). These plans share
several similar motivations and reasons for the
implementation of medicinal information systems.
These include: reduction of medication errors, better
productivity, and financial savings.
At the same time the technology developed for
managing document has significantly developed.
However this new technology based on Semantic
Web (Daconta et al., Davies et al., 2002) is not yet
deployed in managing medicinal information
systems. This is regrettable since through this
technology many of the goals of the medicinal
information systems could be achieved in an elegant
way.
However, during the past few years several
organizations in the healthcare sector have produced
standards and representation forms using XML. For
example, patient records, blood analysis and
electronic prescriptions are typically represented as
XML-documents (Dolin et al., 2001; Jung, 2005;
Puustjärvi and Puustjärvi, 2006; Bobbie et al., 2005).
This generalization of XML-technologies sets a
promising starting point for the management of
medicinal documents. However, the introduction of
XML itself is not enough but also many other XML-
based technologies (Harold and Scott Means, 2002)
290
Puustjärvi J. and Puustjärvi L. (2009).
SEMANTIC FRAMES - A Way for Automating the Management of Medicinal Documents .
In Proceedings of the First International Conference on Computer Supported Education, pages 290-295
DOI: 10.5220/0001968802900295
Copyright
c
SciTePress
have to be introduced in order to alleviate the
complexity of managing medicinal information (Lin
and Hsieh, 2006; Raisinghani and Young, 2008).
We have investigated the management of
medicinal knowledge in pharmacies. It has turned
out that the volume of information coming in from a
variety of information sources such as
pharmaceutical companies, medicinal wholesalers,
social insurance institutions and other authorities is
increasing all the time. Further, the various formats
such as paper, fax, email, and a wide variety of
multiple electronic media formats are still
complicating the information management.
In order to alleviate this problem we have
investigated semantic exchange of pharmaceutical
information between medical information systems.
By semantic exchange we refer to the ability that the
communicating parties can unambiguously (based
on medicinal ontologies) interpret the exchanged
messages.
In order to alleviate this problem, we have
introduced the notion of semantic frames, which are
included in incoming XML-documents that are sent
by other medical organizations to pharmacies.
Semantic frames are presented in RDF, and they
indicate how the document (medicinal instruction)
relates to the ontologies and taxonomies of the
pharmacy system. Hence, based on the semantic
frame, the integration of the document to the
knowledge base can be automatically done. This,
however, requires that the healthcare organizations
that send the medicinal documents and pharmacies
which receive the documents have to commit to the
same medicinal ontologies, i.e., use the same
vocabulary in specifying and interpreting the
semantic frames.
Further, our used ontology models the
relationships of the incoming documents and the
daily duties in pharmacies, and hence the integration
of the documents and daily duties can be also
automatically done. The gain of this approach is that
medicinal instructions are provided just-in-time, and
tailored to their specific needs.
The rest of the paper is organized as follows. In
Section 2, we give a motivating example of the
problems we have encountered in pharmacies in
managing the incoming information flows. Then, we
consider the way incoming information should be
managed in knowledge centric organizations. In
Section 3, we consider the role of medicinal
ontologies and taxonomies in managing documents.
We first consider the annotation of medicinal
instructions by medicinal taxonomies. Then we
present our used medicinal ontology, which
incorporates documents (medicinal instructions) and
the dispensation of medicinal products as well as
their relationships.
In Section 4, we consider semantic frames from
technical point of view. We first present the SOA-
architecture, where the exchange of semantic frames
takes place, and then we illustrate how semantic
frames are expressed in RDF, and how they are
incorporated in SOAP-messages. Finally, Section 5
concludes the paper by discussing the advantages
and critical issues of our presented approach.
2 MOTIVATING EXAMPLE
A significant problem in pharmacies is that there are
no commonly agreed practices for managing
incoming information. The information management
is more or less haphazard. On the other hand, it is
well known that this is also the case with many other
organizations. However, in pharmacies, the
incoming information is more critical in the sense
that it has direct dependencies on patients’
healthcare.
As an example, consider the activity where a
pharmacist dispenses drugs to the customer and
generates a dispensation note. This operation may
give rise for many kinds of additional
announcements for the customer. For example:
The pharmaceutical company may have
informed pharmacies that the shape of
the tabs has changed, the size of
package is changed, or the consistency
of the drug has changed.
Medicinal wholesalers may have
informed of the forthcoming changes on
certain medicinal products.
Social insurance institutions may have
informed the pharmacy of the changes
concerning the refunds of the dispensed
drug.
Healthcare authorities may have
informed the pharmacy about the
changes concerning the generic
substitution of the dispensed drugs. In
addition, healthcare authorities may
have set restrictions on physicians’
rights for prescribing certain drugs or
dispensation of certain drugs for certain
patients may be forbidden.
SEMANTIC FRAMES - A Way for Automating the Management of Medicinal Documents
291
The information of such changes is crucial for
the customer. However, due to the huge amount of
such incoming information there is no hope that the
pharmacist can ensure that she or he has transmitted
all the relevant information to the customer.
Even though all the incoming information were
stored in pharmacy’s information system the
problem still remains, as the relevant information
and its integration into daily duties are not specified
in a machine understandable form. On the other
hand, as the various systems are independently
developed, built based on proprietary solutions,
developed in piecemeal way, and tightly coupled
through ad hoc means, there is no easy way to
develop solutions that require the interoperability of
the various systems.
It is well known that these kinds of problems can
be avoided in knowledge centric organizations
(Davies et al., 2002; Antoniou and Harmelen, 2004).
By the term knowledge centric organization we refer
to the organization which incorporates Semantic
web technologies in information modelling,
presentation, storing and retrieval.
Storing a document in the knowledge base
includes the following steps:
1. Mark up the information with XML using a
relevant XML-Schema.
2. Annotate the information (annotations
specify how the connections to organization
taxonomies and ontologies can be done).
3. Integrate the information to organizations
ontologies.
4. Store the information in an application with
a Web service interface. If this is a new
Web service, it should be registered in the
organization’s registry, along with its
taxonomy classification.
This kind of management of incoming documents
and their associations with organizations taxonomies
and ontologies allows effective search and querying
functionalities.
A problem, however, is that annotating the
documents by appropriate taxonomies and
ontologies (i.e., step 2) has turned out to be hard as
there are no simple way for automating the
annotation. In our approach this problem is avoided
as we assume that the sender (creator) of the
documents has already annotated the document as it
has relevant information to make the annotation.
This however, requires that the communicating
parties commit to the same medicinal ontologies and
taxonomies with respect to the exchanged
documents.
We next consider the medicinal taxonomies and
ontologies on which our solutions are based on.
3 MEDICINAL TAXONOMY AND
ONTOLOGY
3.1 Medicinal Taxonomy
In order to standardize semantic metadata specific
taxonomies are introduced in many disciplines.
In general, taxonomy is a way to classify or
categorize a set of things into a hierarchy (Daconta
et al., 2003). It is a tree like structure consisting of a
root and branches where each branching point (i.e., a
node) and leaf is an information entity. In the
context of information technology taxonomy is
generally understood as the classification of
information entities in the form of a hierarchy, ac-
cording to the presumed relationship of real-world
entities that they represent.
The logic behind taxonomy is that when one
goes up the taxonomy toward the root, the
information entities become more general, and
respectively when one goes down towards the leaves
the information entities become more specialized.
To illustrate this, a simply taxonomy of medicinal
groups is presented in Figure 1. The idea behind this
classification is that the medicinal instructions can
be annotated by the metadata items (the branching
points and the leaves) represented in the tree.
Medical product group
Cardiac drug
Pain drugs
Stomach drugs
Prescription
based pain
drugs
Over the counter
pain drugs
Pain drugs
for topical
use
Aspirin Panadol Ibuprofen
Figure 1: A simple Medical product group taxonomy.
For example, a reason for missing many relevant
documents in keyword based searching is that the
keywords used with queries and documents
metadata descriptions are not standardized by
appropriate taxonomies, e.g., a document may be
CSEDU 2009 - International Conference on Computer Supported Education
292
annotated by a keyword “over the counter drug”
while a user uses keyword “self-care drug” in
searching the information.
3.2 Medicinal Ontology
The term ontology originates from philosophy where
it is used as the name of the study of the nature of
existence (Gruber, 1993). In the context of computer
science, the commonly used definition is “An
ontology is an explicit and formal specification of a
conceptualization” (Antoniou and Harmelen, 2004).
So it is a general vocabulary of a certain domain.
Essentially the used ontology must be shared and
consensual terminology as it is used for information
sharing and exchange. On the other hand, ontology
tries to capture the meaning of a particular subject
domain that corresponds to what a human being
knows about that domain. It also tries to characterize
that meaning in terms of concepts and their
relationships.
Ontology is typically represented as classes,
properties attributes and values. So they also provide
a systematic way to standardize the used metadata
items. Metadata items describe certain important
characteristics of their target in a compact form
(Baeza-Yates and Ribeiro-Neto, 1999). The
metadata describing the content of a document (e.g.,
an electronic prescription) is commonly called
semantic metadata. For example, the keywords
attached to many scientific articles represent
semantic metadata
Each ontology describes a domain of discourse.
It consists of a finite set of concepts and the
relationship between the concepts. For example,
within electronic prescription systems patient,
medicinal product, and e-prescription are typical
concepts. These concepts and their relationships are
graphically presented in Figure 2, where ellipses
represent classes and boxes represent properties.
The ontology of Figure 2 includes for example
the following information:
Taxonomy item is a class, and a parent item
may be associated to a taxonomy item.
Hence, we can model a taxonomy (a
hierarchy of concepts) by the medicinal
ontology.
Dispensation is a class meaning that each
instance of the class is stored in the
knowledge base. (Semantically a
dispensation is a purchase of a prescription
based medicinal product, i.e., a business
transaction in a pharmacy). Dispensation is
related to an e-prescription, which is related
to a medicinal product. Further an e-
prescription relates to a patient and
physician.
Instruction is a class, and an instruction
may replace another instruction. Further, an
instruction may be related to medicinal
product, patient or physician.
e-prescription
physician
isPrescribedBy
patient
isTargetedAt
name
id
id
name
id
Instruction
medicinal_product
includes
physician_association
patient_association product_association
dispensation
relatesTo
id
name
Tramadol
Norvasc
typeOf typeOf
replaces
taxonomy_item
parent_item
IsAnnotatedBy
Figure 2: A medical ontology.
A significant point in our used medicinal
ontology is that it indicates the possible relationships
of the class dispensation and class instruction.
Hence, in dispensation activity we can query
whether a given dispensation relates to one or more
instructions. And most importantly, such a query can
be automatically generated within each dispensation
activity, and in the case of non empty result, the
relevant instructions are shown to the pharmacist. It
is also possible to visualize the instructions by using
Semantic web technologies as proposed in
(Puustjärvi and Puustjärvi, 2008).
4 REPRESENTING SEMANTIC
FRAMES
Our used architecture is based on Service Oriented
Architecture (SOA) (Singh and Huhns, 2005). It
SEMANTIC FRAMES - A Way for Automating the Management of Medicinal Documents
293
provides flexible methods for connecting pharmacy
systems to the other relevant systems (Figure 3).
SOAP protocol
Social insurance
institutions
system
Pharmaceutical
company’s
system
Medicinal
wholesalers
system
Web service
interface
Web service
interface
Web service
interface
Pharmacy’s
Information
system
Web service
interface
Figure 3: The interaction of medicinal information
systems.
We now illustrate how we can send the
medicinal intructions as well as their semantic
frames by the SOAP protocol.
SOAP was orginally intended to provide
networked computers with remote-procedure call
services written in XML. It has since become a
simple protocol for exchanging XML-messages over
the Web.
A SOAP-message is comprised of a SOAP
header, SOAP envelope and SOAP body which
contains the application-specific message that the
backend application will understand. As illustrated
in Figure 4, we incorporate our used semantic frame
and the instruction in the SOAP body.
HTTP Header
SOAP Envelope
SOAP Header
Headers
SOAP Body
Semantic Frame Instruction
Figure 4: Incorporating a semantic frame in a SOAP
message.
An example of our used SOAP-message is
presented in Figure 5, where the semantic frame is
an RDF-description. The namespaces “mo” and “to”
specify the used ontology and taxonomy,
respectively. The semantic frame indicates that the
instruction “The shape of the tablet has changed” is
an instance (type) of class “instruction”. Further, the
description indicates that the product_association of
the instruction is Tramadol. In addition, the
Instruction is annotated by the taxonomy “medicinal
product group” by the keyword “Prescription based
pain drugs”.
<SOAP-ENV: Envelope
xmlns:SOAP-ENV=“http://schemas.xmlsoap.org/soap/envelope/
SOAP-ENV:encodingStyle=”http://schemas.xmlsoap.org/soap/encodig/”>
<SOAP-ENV:Body>
<sematic-frame>
<rdf:RDF
xmlns : rdf=”http://www.w3.org/1999/02/22-rdf-syntax-ns#”
xmlns : xsd=”http://www.w3.org/2001/XMLSchema#”
xmlns : mo=“http://www.lut.fi/ontologies/medical-ontology#”
xmlns : mt=“http://www.lut.fi/ontologies/medical-taxonomy#”>
<rdf:Description rdf:about=”instruction_123”>
<rdf:type rdf:resource=“&mo;instruction”/>
<mo : product _association>Tramadol</mo : recipient>
<to : medicinal_product_group>
Prescription_based_pain_drugs
</to : medicinal_product_group>
</rdf : Description>
</rdf:RDF>
</semantic-frame>
<instruction>
<comment> The shape of the tablet has changed </comment>
</instruction>
</SOAP-ENV: Body>
</SOAP-ENV: Envelope>
Figure 5: A SOAP-message including a semantic frame
and an instruction.
5 CONCLUSIONS
Medicinal knowledge is expanding every day. As a
result neither the pharmacists nor other workers in
the health care sector can keep up without the help
of modern information and communication
technology. We have considered this problem in the
case where a pharmacist dispenses drugs to the
customer and generates a dispensation note.
The dispensation may require a variety of
relevant announcements for the customer. However,
due to the huge amount of such incoming
information there is no hope that the pharmacist can
ensure that she or he has transmitted all the relevant
information to the customer.
In order to alleviate this problem we have
introduced the notion of semantic frames, which are
included in incoming XML-documents that are sent
by other medical organizations to pharmacies.
Semantic frames specify how the incoming
document (medicinal instruction) relates to
medicinal ontologies the integration of the document
to the knowledge base can be automatically done.
CSEDU 2009 - International Conference on Computer Supported Education
294
Further, as the used medicinal ontology models the
relationships of the incoming documents and the
daily duties, the integration of the documents and
daily duties can be automatically done. The gain of
this integration is that medicinal instructions can be
provided just-in-time, and tailored to their specific
needs.
An essential prerequisite of our introduced
approach is that the communicating commit to the
same medicinal ontologies, i.e., use the same
vocabulary in specifying and interpreting semantic
frames. This is a critical issue with our approach.
The problem here, as well as with any model that is
based on Semantic web technologies, is that the
creators of the documents (instructions) are
burdened with generating semantic frames.
However, the generation of semantic frames can be
automated by using the tools developed for the
Semantic web.
On the other hand, the introduction of a new
technology is also an investment. The investment on
new ICT-technology includes a variety of costs
including software, hardware and training costs.
Training the staff on semantic web technology is a
big investment, and hence many organizations like
to cut on this cost as much as possible. However, the
incorrect usage and implementation of a new
technology, due to lack of proper training, might
turn out to be more expensive in the long run.
Anyhow, the only way to evaluate definitely our
proposed approach of using semantic frames is to
implement the system in the case of real
applications, i.e., in the environment comprising of
several pharmacies and medicinal organizations.
REFERENCES
Antoniou, G., Harmelen., F., 2004. A semantic web
primer. The MIT Press.
Baeza-Yates, R., Ribeiro-Neto, B., 1999. Modern
Information Retrieval. Addison Wesley.
Batenburg, R., Van den Broek, E. (2008) Pharmacy
information systems: the experience and user
satisfaction within a chain of Dutch pharmacies,
International Journal of Electronic Healthcare. Vol. 4,
No.2 pp.119-131.
Bobbie, P., Ramisetty, S., Yussiff, A-L., Pujari, S. (2005)
Desgning an Embedded Electronic_prescription
Application for Home_Based Telemedicine using
OSGi Framework. Available at:
http://cse.spsu.edu/pbobbie/SharedFile/NewPdfs/eRx-
Final2Col.pdf
Daconta, M., Obrst, L., Smith, K. (2003) The semantic
web. Indianapolis: John Wiley & Sons.
Davies, J., Fensel, D., Harmelen, F., 2002. Towards the
semantic web: ontology driven knowledge
management. John Wiley & Sons.
Dolin, R.H, Alschuler, L., Beerb, C., Biron, P.V., Boyer,
S.L., Essin, D., Kimber, E., Lincoln, T., Mattison, J.E.
(2001) The HL7 Clinical Document Architecture. J.
Am Med Inform Assoc, Vol. 8 No.6, pp. 552-569.
Dwivedi, A., Bali, R.K., Wickramasinghe, N., Naguib,
R. (2007) How workflow management systems enable
the achievement of value driven healthcare delivery.
International Journal of Electronic Healthcare, Vol. 3,
No.3 pp.382-393.
Ghani, M., Bali, R.K., Naguib, R., Marshall, I.M.,
Wickramasinghe, N.S. (2008) Electronic health
records approaches and challenges: a comparison
between Malaysia and four East Asian countries,
International Journal of Electronic Healthcare,Vol. 4,
No.1 pp.78-104-77.
Gruber, T. R., 1993. Toward principles for the design of
ontologies used for knowledge sharing. Padua
workshop on Formal Ontology.
Harold, E., Scott Means W. (2002) XML in a Nutshell.
O’Reilly & Associates.
Hyppönen, H., Salmivalli, L., Suomi, R. (2005)
Organizing for a National Infrastructure: The Case of
the Finnish Electronic Prescription. In Proc. of the
38th Hawaii International Conference on System
Sciences.
Jung, F. (2005) XML-based prescription drug database
helps pharmacists advise their customers. Available
at:
http://www.softwareag.com(xml/aplications/sanacorp.
htm.
Lin, B., Hsieh, C. (2006) Critical factors for assessing
service quality of online pharmacies: a research
framework. International. Journal of Electronic
Healthcare, Vol. 2, No.4 pp.398-414.
Puustjärvi, J., Puustjärvi, L. (2006). The challenges of
electronic prescription systems based on semantic web
technologies. In Proc. of the 1st European Conference
on eHealth, pp. 251-261.
Puustjärvi, J., Puustjärvi, L. (2008) Using semantic web
technologies in visualizing medicinal vocabularies. In
the proc of the IEEE 8th International Conference on
Computer and Information Technology.
Raisinghani. M.S. , Young, E. (2008) Personal health
records: key adoption issues and implications for
management, International Journal of Electronic
Healthcare. Vol. 4, No.1 pp.67-77.
Singh, M., Huhns, M. (2005) Service Oriented
Computing: Semantics Proceses, Agents. John Wiley
& Sons.
SEMANTIC FRAMES - A Way for Automating the Management of Medicinal Documents
295