VOL. 21 NO. 4, DECEMBER, 2016
Contents | Author index | Subject index | Search
| Home
Supporting better treatments for meeting health
consumers' needs: extracting semantics in social data for
representing a consumer health ontology
Yunseon Choi
Abstract
Introduction. The purpose of this paper is to provide a
framework for building a consumer health ontology using
social tags. This would assist health users when they are
accessing health information and increase the number of
documents relevant to their needs.
Methods. In order to extract concepts from social tags, this
study conducted an empirical study on terms collected from a
social networking site. The semantics of tags were analyzed
and a concept list was developed by using the middle-out
strategy.
Analysis. This study analysed the semantic values of tags by
employing Latent Semantic Analysis (LSA). This is a method
for extracting and representing the contextual-usage
meaning of words by analyzing relationships between
documents and the terms they contain and word semantics.
Results. The process of building an ontology using social
tags shows how using this consumer health ontology could
improve user access and retrieval. It demonstrates how terms
extracted from tags are related to each other with similarity
and relationships within hierarches in the ontology.
Conclusion. The study has implications for better design of
ontology applications that support the search for health-
related resources. This will enhance the communication
between health consumers and professionals.
change font
Introduction
As a large number of online health resources have become
available, there has been a great increase of the number of
health consumers replying on online health resources
available on the World Wide Web (Andreassen,
Bujnowska-Fedak, Chronaki, Dumitru, and Pudule, 2007;
Fox, 2011; Rice, 2006; MacLean and Heer, 2013). It has
been reported that health consumers should be able to
have effective access and utilise relevant health
information to meet their needs (Nutbeam, 2008; World
Health Organisation, 2011). A Pew Research Center survey
indicates that 72% of U.S. adult Internet users have looked
for health information online (Fox and Duggan, 2013).
Studies also show that most consumers lack the skills to
access and use effectively online health resources (Friel,
Bond, and Lahoz, 2015; Gray, 2005; Jain and Bickham,
2014; Ratzan and Parker, 2000; Rowlands et al., 2013).
There have been efforts to provide access to reliable health
information on the World Wide Web, and MedlinePlus and
InformedHealthOnline are such examples. MedlinePlus is
maintained by the National Library of Medicine and it is a
Web-based consumer health information service (Miller,
Lacroix, and Joyce, 2000). InformedHealthOnline is
published by the German Institut für Qualität und
Wirtschaftlichkeit im Gesundheitswesen (or IQWiG) and is
the English-language version of the German website which
provides health information to the public and patients.
Information in health or medical domains is critical and
should be provided to health consumers without difficulty.
However, the growing amount of health information on
the web has increased concern about effective access to
quality health information because terminology, currently
used for organising health or medical information, is
generated by professionals and may not be familiar to
users. The terminology gap between users' and
professionals' vocabulary in describing medical-related
web documents was also uncovered by a study on indexing
consistency of social tagging in comparison with
professional indexing (Choi, 2014). Health consumers and
healthcare professionals tend to use different terms to
describe health-related concepts, for example, dry mouth
vs. xerostomia and flu vs. influenza (Vydiswaran, Vinod,
Hanauer, and Zheng, 2014). This terminology gap in the
health domain prevents health consumers from accessing
health information relevant to their information needs.
For example, when a health consumer tries to find
information related to nosebleed symptoms, she/he may
not find the resources including only the term epistaxis in
the meta tags, title and text (Zielstorff, 2003). In large
medical health consumer websites, it has been reported
that when a consumer's terms are different from
physician-defined terms, the search returned no results,
for example, heart attack vs. myocardial infarction (Zeng,
Kogan, Ash, and Greenes, 2001) and shakes vs. tremor
(Zielstorff, 2003).
On the other hand, as networked information resources on
the web continue to grow rapidly, digital information
environments have led librarians and information
professionals to manage digital resources on the web.
Thus, this trend has required new tools for organizing and
providing more effective access to the web. Subject
directories or Web directories are such tools for internet
resource discovery since subject directories organise Web
documents by subject areas. Yet, studies have shown that
subject directories based on traditional organisation
schemes are not sufficient for the web (Golub, 2006;
Nowick and Mering, 2003; Macgregor and McCulloch,
2006). This is because they were developed using
traditional library schemes which have been developed
with a focus on physical library collection. Web
documents, however, were originally organized and
indexed by professionally-generated keywords. This means
they do not reflect intuitively and instantaneously
expressed users' current needs (Macgregor and McCulloch,
2006).
Although there have been efforts to involve users in
developing information organization systems, they are not
necessarily based on users' real languages. Accordingly,
social tagging has received significant attention as a
promising way to solve this challenge since users' tags
reflect their interests and their languages. Social tags are
good sources for identifying users' terms. Several
researchers have discussed the impact of tagging on
retrieval performance on the web (Bao, 2007; Choi, 2009;
Choy and Lui, 2006; Golder and Huberman, 2006;
Heymann, Koutrika, and Garcia-Molina, 2008; Sen et al.,
2013; Yanbe, Jatowt, Nakamura, and Tanaka, 2006).
Although social tags have been discussed regarding its
usefulness as additional access points for classification and
retrieval (Trant, 2009; Choi, 2014), there has been little
research conducted on the use of social tags to improve
practices in information organization. Since social tags
provide additional access points as user-generated terms,
using them would improve information access and
promote effective reasoning for retrieval.
In terms of information organization, ontologies have been
used for information organization and information
integration. Ontology is a shared understanding of a
domain that can be communicated between people and
computers (Ding, 2001). Especially, in the medical and
health services, information systems should be able to
communicate difficult and complex concepts. However,
analysing the structure and concepts of medical
terminologies cannot be easily achieved.
There have been very few studies conducted on building
health or medical ontologies which features concepts and
vocabularies familiar to health consumers. Mayo
consumer vocabulary, a taxonomy of consumer health
terms and concepts, was developed and maintained by
Mayo Clinic (Seedorff et al., 2013). The Consumer Health
Vocabulary Initiative resulted in the creation of the Open
access collaborative consumer health vocabulary, which
was designed to complement the existing framework of the
Unified medical language system and to aid the needs of
consumer health applications (US. National Library of
Medicine, 2012). However, this vocabulary is not
implemented using a knowledge representation language
such as Web Ontology Language which supports semantic
search and knowledge reasoning.
The aforementioned important components of effective
health information organization are applied in this study:
Due to the unfamiliarity of health consumers to
current terminology used for organizing health or
medical information, medical information systems
need to include user-friendly vocabulary.
Considering the characteristics and quality of social
tags in representing users' views, social tags should
be utilised to improve practices in information
organization.
To establish a closer link between health consumers'
information needs and professionals' responses, a
powerful semantic-based ontology needs to be built.
This paper is part of a larger research project which aims
to answer questions about how we can assist users when
they are accessing health information in order to increase
the number of documents they find relevant to their needs.
The ultimate goal of the project is to build a consumer
health ontology by utilising social tags assigned to health-
related documents. The main objective of this paper is,
therefore, to provide the framework for a consumer health
ontology by discussing the process of building an ontology
featuring social tags. This paper intends to show how
social tags can be utilised for developing class hierarchies
in the ontology in order to identify unambiguously implicit
relations among social tags.
Ontologies for information organization and
information integration
Definitions of ontologies
The term ontology has been used in several disciplines,
from philosophy to computer science. As a branch of
philosophy, ontology studies the structures of the objects,
properties and relations of reality (Smith, 1997). In
computer science, into which the term came from artificial
intelligence, the ontology is a model of the representation
of objects in the world with properties and relationships
(Garshol, 2004). An ontology is defined as a formal,
explicit specification of a conceptualisation (Gruber, 1993;
Studer, Benjamins, & Fensel, 1998):
Conceptualisation refers to 'an
abstract, simplified view of the world
that we wish to represent for some
purpose' (Gruber, 1993
, p. 1).
Explicit refers to the 'type of concepts
used, and the constraints on their use
are explicitly defined' (Studer,
et al.,
1998, p. 25).