2. Massively Multiplayer Online Role Playing Games (MMORPG)
2.5 La imagen-juego en el videojuego
The MUSTANG terminology server (Medical UMLS-based Terminology Server for Authoring, Navigating and Guiding the Retrieval to heterogeneous Knowledge Sources) is being developed to support indexing and data-recall facilities. MUSTANG, a CORBAmed-based central multilingual terminology server, provides the semantic foundation for a repository of XML-document forms. MUSTANG delivers terminological services to applications by standardized interfaces and provides access to terminology resource related to the Unified Medical Language System (UMLS). In addition, MUSTANG provides functions to retrieve medical information from Internet knowledge sources (e.g. PubMed) via MeSH-based queries. MUSTANG is implemented on a Windows platform using the ORACLE database management and development software [11], [12], [22].
The terminology, together with forms and IP protocols, represents the pivotal element for communication of patient-related data between sites and services. The terminology service is a basic feature of future infrastructures. One of the major drawbacks in communications between different medical systems in the past (at the information exchange level) was the use of different terminologies. Descriptive definitions and terminologies vary between countries and even
between physicians in the same country. To achieve data integration on a semantic level, TOSCA has worked to build a standardized terminology.
Numerous coding systems (e.g. SNOMED nomenclature, MeSH thesaurus, ICD-10 and ICPM classification) can be used as a possible basis for standardization. However, these key systems are standalone vocabularies and are inadequately maintained as far as software engineering is concerned. With the aim of creating an “ophthalmology platform” health service, a suitable architecture of terminological services with a clearly defined server has been established. This server communicates with the information broker and subsystems as well. All TOSCA applications rely on a telemedical infrastructure for communication, based on established and new technical standards, such as XML for data transfer via networks, DICOM for image transfer and HTTPS for the interaction of different application systems within the telemedical communications infrastructure. All sensitive patient-related data is transferred according to European safety standards. The communication platform includes brokering services, telescreening, image processing, a reference image database and a monitoring system.
Figure 6-4 The architecture of the MUSTANG Terminology Server.
To enable integration of different systems and services, developed within the TOSCA project, as well as of patient data management systems, a standardized ophthalmological terminology for Glaucoma and Diabetic Retinopathy has been developed. Selected medical terms have been adopted from previous and current European projects (e.g. OPHTEL, MUSTANG) and adjusted for the TOSCA project. New terms relevant for TOSCA have been added to the terminological component.
The terminological component is the basis for structured documentation and reports. Descriptive definitions, hierarchical relations among concepts and language-specific terms of a common terminology for Diabetic Retinopathy and Glaucoma are provided. The terms and definitions for the medical concepts are
in accordance with international standards (ICD, SNOMED, MeSH, etc.) The terminology also provides the functionality necessary for maintaining the consistency of medical contents and other components.
A central data dictionary is a prerequisite for integration of systems and subsystems. It defines the syntax and semantics of data. The data dictionary contains definitions of parameters, including data types, value sets, units, etc., that are used for documenting data. The basic resources of an evaluated TOSCA terminology are various coding systems and vocabularies that are relevant to the project. These extensive resources have to be managed by a special information system (for usability reasons). In TOSCA two such systems are used within this task: the MUSTANG Terminology Server and the KAMATO® terminology management module.
2.2.1 TOSCA Terminology in KAMATO
The terminological concepts for diabetic retinopathy and glaucoma are also stored in a KAMATO terminology module. KAMATO stands for Knowledge Acquisition and Management Tool, a system developed by AdaKoS GmbH. The Terminology Component of KAMATO ensures terminological control of knowledge bases and enables access to Internet knowledge sources (e.g. PubMed) via MeSH-based queries.
With the aid of a special search function, terminological concepts available in the UMLS Metathesaurus are automatically imported into KAMATO from the MUSTANG server. The preferred term, synonyms, source codes and (if available) semantic types and descriptions are added as well. KAMATO supports the following coding systems: UMLS-CUI, ICD-9, ICD-10, MeSH, Read Code, SNOMED 2, SNOMED Int 3.5, HL7 Terminology, WHOART, ICPM, ULMER and BLS. Further coding systems can be added on request. The concepts are arranged in a hierarchical order, in parent–child concepts. Figure 6-5 presents the parent concepts of the TOSCA terminology project. Figure 6-6 shows the hierarchical order of the “Eye disease” concept.
Figure 6-5 Terminology module of KAMATO.
Figure 6-6 Terminology module of KAMATO: Hierarchical order of the “Eye disease” concept.
With the terminology module of KAMATO, the ophthalmological concepts for diabetic retinopathy and glaucoma can be presented in structured hierarchical order. The MUSTANG terminology server allows new concepts, based on UMLS, to be imported automatically. The structured order and the clear user
interface allow easy and fast maintenance of the terminological concepts as well as multilingual usage.
Within the TOSCA project, the terminological concepts in KAMATO are used for the “Patient-centred Knowledge-based Information System for Glaucoma”. The concepts provide the basis for indexing documents and contents of the knowledge base. Indexing enables terminological control of the knowledge base and accurate retrieval of documents and knowledge contents (Figure 6-7).
Figure 6-7 Patient-centred Knowledge-based Information System for Glaucoma: preferred term and code set for “Visual field defect”.
The terminology module of KAMATO also enables access to Internet knowledge sources (e.g. PubMed) via MeSH-based queries (Figure 6-8).
Figure 6-8 Access to PubMed via KAMATO: Publications on the selected term “Glaucoma, open-angle”.
2.2.2 Data Sets
Terminologies and data sets are closely related. Within the GALEN project the DIABCARD data set has been used as a basis for a terminology server which was then applied for developing a medical information system. It turned out that additional effort is needed to harmonize data sets and terminology with the goal of implementing interoperable EHRs.
The EFMI Special Topic Conference 2003 was held in Rome, under the heading “The content of the Electronic Health Record: Clinical datasets for continuity of care and pathology networks”. Its working hypothesis was as follows: Clinical
Dataset describes the set of predefined data Entries stored, shared or presented as a unit within clinical applications, messages and EHR systems. A Clinical Dataset is able to describe a
particular aspect of the patient's status, of a procedure, of a clinical document in relation to a given health issue or task, across heterogeneous contexts.
A set of recommendations was developed on the basis of the current situation in healthcare telematics:
x standardization efforts on the HER,
x need for sharing and aggregating clinical information,
x need for clinical datasets to improve the coherence of HER contents, x a gradual process of convergence on clinical datasets.
The participating experts agreed on 20 recommendations, which were grouped under the following headers:
1. create a network of organizations interested in clinical datasets, within the context of the ongoing actions on HER;
2. stimulate the production and systematization of clinical entries (a kind of archetype) and clinical datasets (a kind of template),
3. define the role of the responsible organizations and produce the guidelines for accurate documentation on packages of clinical datasets, 4. set up a repository of templates, based on a standard processable
formalism,
5. set up a gradual process of convergence, assisted by the repository of templates.
Group 4 is of particular interest in the context of terminology:
x Production, retrieval, distribution, implementation, comparison and systematization of clinical datasets should be facilitated by setting up a suitable repository of templates – with an appropriate version control – and by stimulating the production of authoring tools;
x The upload and download of the descriptions of clinical datasets to and from the repository of templates, as well as their usage, should be free of charge;
x Several organizations have already developed their own clinical datasets, according to a specific local formalism. It is expected that suitable software tools will be able to assist in the semiautomatic translation and refinement of the existing clinical datasets into a unique standard formalism, possibly passing through a set of intermediate formalisms convenient for developers and users of each particular type of clinical data set.
The collection of data sets has just started. Initial results will be discussed at a continuation workshop in October 2004 and will be available on the EFMI web pages (www.EFMI.org).
3 Summary
Specialist terminology has always been used in the medical field for communication between professionals. In early days it was Greek- and Latin- based and not to be understood by others, especially the patient. Coding was used for the analysis and description of diseases and the healthcare system. Structured documentation for better communication and easy data entry has been a step forward in exchanging clinical data between professionals, based on classifications.
The Internet is changing existing habits. More than 60% of adults in some countries have access and 75% of them use the Internet to look for healthcare information[14]. This fact is influencing the decisions of patients regarding treatment etc. Terminology has to be adapted to the consumers’ clinical and mental models of diseases and their treatment. This will be a challenge for the coming years and it has to take into account the fact that health professionals also have to be knowledgeable about health consumers and their communication profiles. Health telematics stands a chance of benefiting all participants in the healthcare setting; first, however, it needs to organize the communication, storage and retrieval of data and knowledge as well as the structuring of information in a global healthcare system.