Development of a Lightweight and Adaptable Clinical Terminology Server

Abstract

Health terminologies, vocabularies and ontologies are increasingly becoming indispensable computational artifacts since they allow the standardization and conceptualization of medical vocabulary. Its applications vary from the cleaning and improvement of data quality, learning, and serving as a basis for health information exchange and up to improve data analysis and decision making. In this context an important aspect is the software component responsible for storing and managing these linguistic artifacts often named as terminology server, ontology repository, metadata repository, among other names. In this paper we present the eHealth Interop Terminology Server, a clinical terminology server based on health information standards. The main characteristics is its adaptability and can be easily implemented as a component within a health information system as well as serve as the basis for a terminological portal. It uses service oriented architecture so all the communication is made using a web services API. We also perform a functional test with clinical terminologies in the context of a mental health care network.

Full text

ISBN 978-989-97433-8-0 E-book edition 2017 by SciKa
Book of abstracts of the
CENTERIS 2017 – Conference on Enterprise Information Systems /
ProjMAN 2017 – International Conference on Project MANagement /
HCist 2017 – International Conference on Health and Social Care Information Systems and Technologies
264
Development of a Lightweight and Adaptable Clinical
Terminology Server
Newton Shydeo Brandão Miyoshi, João Mazzoncini de Azevedo Marques,
Domingos Alves, Paulo Mazzoncini de Azevedo Marques
Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, São Paulo, Brazil
Abstract
Health terminologies, vocabularies and ontologies are increasingly becoming indispensable computational artifacts
since they allow the standardization and conceptualization of medical vocabulary. Its applications vary from the
cleaning and improvement of data quality, learning, and serving as a basis for health information exchange and up
to improve data analysis and decision making. In this context an important aspect is the software component
responsible for storing and managing these linguistic artifacts often named as terminology server, ontology
repository, metadata repository, among other names. In this paper we present the eHealth Interop Terminology
Server, a clinical terminology server based on health information standards. The main characteristics is its
adaptability and can be easily implemented as a component within a health information system as well as serve as
the basis for a terminological portal. It uses service oriented architecture so all the communication is made using a
web services API. We also perform a functional test with clinical terminologies in the context of a mental health
care network.
Keywords: clinical terminologies; ontologies; web services; health information exchange; concept matching.
1. Introduction
The application of information technologies in the health area is increasingly an essential tool for
both rapid resolution and coordination of various processes and for analysis and decision making in
individual and collective care. A central aspect in this context is the standardization of terms and
concepts and this is due to three main factors: (i) the highly heterogeneous and sometimes ambiguous
nature of medical language and its constant evolution; (ii) an enormous amount of data that is generated
constant by the automation of processes and from new technologies that arise and (iii) the need to
process, analyze and make decisions based on this information1.

ISBN 978-989-97433-8-0 E-book edition 2017 by SciKa
Book of abstracts of the
CENTERIS 2017 – Conference on Enterprise Information Systems /
ProjMAN 2017 – International Conference on Project MANagement /
HCist 2017 – International Conference on Health and Social Care Information Systems and Technologies
265
Over the years several artifacts have been created and formalized and standardized medical
knowledge such as: terminologies, codifications, controlled vocabularies, taxonomies, thesaurus and
ontologies2 as well as different formats for the formal representation of these artifacts. In this context,
the terminology and ontology server is the piece of software responsible for storing and managing these
computational linguistic artifacts. The terminology and ontology server acts as a terminology repository
and can add other functions and auxiliary tools to search terms and concepts.
The objective of this research is the design and development of a clinical terminology server based
on health information standards and has two main characteristics: (i) it is robust and functional as an
independent piece of software that can serve as the basis for the construction of A knowledge portal on
a specific topic; (ii) is flexible to act as a software component and can be easily integrated as part of an
existing health information system.
2. Methods
To allow the development of a clinical terminology server that can be deployed in several
environments we choose a service-based architecture. The service-oriented computing paradigm allows
applications to be quickly built and can be easily composed in a distributed and heterogeneous
environment3. In this context the terminology server named eHealth-Interop Terminology Server
follows a client-server architecture in which all functionality is made available in a set of restful web
services. As health information standards, the Common Terminologies Services version 24, the HL7
FHIR Terminology Service5 and the IHE Sharing Value Sets Profile6 were used.
2.1. Use Case
Although Brazil has a Unified Health System for all the population, the services organization and
care delivery is done in a regional context in health care networks. In the region of Ribeirão Preto, in
state of São Paulo, the mental health care network comprises about 26 municipalities and approximately
1,400,000 inhabitants. An ongoing project aims to establish a data interoperability platform to support
care coordination in this region. It was defined a set of information models to be exchanged in this
context. We use the solution presented here as the terminology server that will be the basis for the
standardization of mental health care information.
3. Results
A clinical terminology server called eHealth-Interop Terminology Server was designed and
developed. This server acts as the Value Set Repository actor defined in the IHE SVS profile as well as
the terminology module of the HL7 FHIR. In the next subsections we will detail the main components
that comprises the server architecture, the actors and operations that are made available by the API

ISBN 978-989-97433-8-0 E-book edition 2017 by SciKa
Book of abstracts of the
CENTERIS 2017 – Conference on Enterprise Information Systems /
ProjMAN 2017 – International Conference on Project MANagement /
HCist 2017 – International Conference on Health and Social Care Information Systems and Technologies
266
Restful developed, and the use case of using clinical terminologies to support the mental health care
network.
3.1. Software Architecture
The eHealth-Interop Terminology Server architecture has the following main components:
Terminology Database, a document-based database to store all terminologies and concepts data, the
MongoDB7 was used as the database engine; Base Models, set of classes that represent the main data
models used to manage terminologies, ontologies, concepts, value sets and mappings; Mapping Module
is responsible for the high-level functionalities related to the management of mappings between
concepts of different terminologies; Knowledge Module is responsible for high level functionalities for
searching concepts, terminologies and value sets; the RESTful Web services is the set of functionalities
(API) that are exposed to client applications through the HTTP protocol, from these web services it is
possible to build APIs specific to a particular programming language such as Java, among others. The
eHealth Interop Terminology Server was built using the Loopback 3 framework8.
Following the recommendation of the IHE SVS Profile and HL7 FHIR, the authentication and
security features are optional but are represented in the following components: Auth DB, database for
user registration, roles and client applications, the MySQL relational database was used as database
engine; Auth Models, are the classes that allow the management of the data model stored in Auth DB
and finally; Auth Module, which has the business logic for managing the authentication and
authorization features.
3.2. RESTful Web Services Operations
The eHealth-Interop Terminology Server API is the communications layer with the client
applications, in which the main functionalities are available. The main actors involved are: Terminology
Server, Terminology Consumer and Terminology Administrator. The Terminology Server actor, which
is the solution described in this paper, acts as the Value Set Repository actor defined by the IHE SVS
Profile. Terminology Consumer is a representation of a client application and corresponds to the Value
Set Consumer defined by the IHE SVS Profile. Terminology Administrator is an important actor
responsible for some special restricted operations for managing all the data in the Terminology Server
and does not have a correspondent actor in IHE SVS profile.
3.3. Mental Health Care Network – Use Case
In order to support the information model defined for care coordination in the mental health care
network of Ribeirão Preto region, we used 3 international terminologies and 2 national classifications.
The international terminologies are: ICD-10 - International Statistical Classification of Diseases and

ISBN 978-989-97433-8-0 E-book edition 2017 by SciKa
Book of abstracts of the
CENTERIS 2017 – Conference on Enterprise Information Systems /
ProjMAN 2017 – International Conference on Project MANagement /
HCist 2017 – International Conference on Health and Social Care Information Systems and Technologies
267
Related Health Problems, ICF - International Classification of Functioning, Disability and Health,
ICPC-2 - International Classification of Primary Care. The national classifications are the Table of
Procedures, Medications and Orthotics, Prostheses and Materials of the Unified Health System in Brazil
and the Brazilian Classification of Occupations.
In addition, local terminologies were created to represent different concepts such as: sex, marital
status, sexual orientation, among others. These codifications, for the most part, are in agreement with
those used in the National Information System for primary care or based on information models from
national standard for discharge summary.
All these terminologies were instantiated into the eHealth-Interop Terminology Services and their
concepts and terms were imported using a client application acting as an “Terminology Admin” actor.
This client application parsed the different formats and send them using the web services API.
4. Conclusion and Future Work
In this paper we present the eHealth-Interop Terminology Server as a software component that can
be adapted to several health contexts and applications. The use of the service-oriented paradigm through
the Loopback framework has proved to be a quick and effective way to develop a solution adaptable to
different scenarios. The document-oriented database as the terminology database proved to be a good
solution because it allowed the storage of particular properties of each terminology without losing the
basic structure of the information models.
References
1. Konopka BM. Biomedical ontologies—A review. Biocybern Biomed Eng. 2015;35(2):75-86.
doi:10.1016/j.bbe.2014.06.002.
2. Bodenreider O. Biomedical ontologies in action: role in knowledge management, data integration and decision support.
Yearb Med Inform. 2008:67-79. http://www.ncbi.nlm.nih.gov/pubmed/18660879. Accessed May 12, 2017.
3. Papazoglou MP, Traverso P, Dustdar S, Leymann F. Service-Oriented Computing: A Research Roadmap. Int J Coop Inf
Syst. 2008;17(2):223-255. doi:10.1142/S0218843008001816.
4. Object Management Group. Common Terminology Services 2 (CTS2) Version 1.2.; 2015.
http://www.omg.org/spec/CTS2/1.2/. Accessed March 10, 2017.
5. HL7. Terminology Service - FHIR Relase 3 (STU). https://www.hl7.org/fhir/terminology-service.html. Published 2017.
Accessed May 2, 2017.
6. IHE. IHE IT Infrastructure ( ITI ) Technical Framework Volume 1 Integration Profiles. Vol 13.; 2016.
7. MongoDB. MongoDB for GIANT Ideas. http://www.webcitation.org/6qZlFWE1o. Published 2017. Accessed May 3,
2017.
8. Strongloop. LoopBack - Node.js framework. http://www.webcitation.org/6fwAow3zf. Published 2017. Accessed May 3,
2017.