The creation of an ontology of clinical document names.

The efficient use of documents from heterogeneous computer systems is hampered by differences in document-naming practices across organizations. Using an open-consensus method, the Document Ontology Task Force, with support from the Veterans Health Administration, addressed this pervasive problem by developing a clinical document ontology. Based on the analysis of over 2000 clinical document names, the ontology was used to formulate a terminology model which is currently being used to guide the creation of fully-specified document names in LOINC (Logical Observations, Identifiers, Names and Codes). Incorporation into LOINC will enable homogeneous management of documents in a widely distributed environment and will also give rise to a rich polyhierarchy of document names.

Relevance of mandates, notifications and threads in the management of continuity of care.

Availability of electronic healthcare records (EHCR) and geographical networks allows nowadays to realise a set of functionalities to support continuity of care. Actual exchange of complete clinical information, common centralised records, common decisions within an agreed protocols are not mandatory. Sub-optimal alternatives, involving easier management, are possible. In fact, the crucial challenge for continuity of care is the mutual awareness of the multiple perspectives by the actors contributing to patient's care. The various actors should timely know changes in the status of: mandates, i.e. who is involved in the care provision and thus is responsible for a "local" record. knowledge about the patient, i.e. active and inactive problems, impressions, relevant findings. provision of healthcare activities, i.e. plans, orders and performed activities. The decisions on the actual implementation depend on the healthcare context, as implemented within the information system. For example, the mechanism for notifications involves decisions on the quality and quantity of information that must be exchanged, as well as on the modalities for the exchange, regulated by individual user's profiles. Each notification could be sent as a message to a central repository, and then each authorized user could select within the repository the pertaining messages. Or a notification could be sent directly to the list of professionals involved in the care of a patient, that asked for it in their profile. Mandates may be used to regulate the access of the users to the patient's information. This approach was embedded in a European standard under development in CEN/TC251 (CONTSYS--"System of concepts to support continuity of care").

Structures of clinical information in patient records.

In order to support the preparation of the European Prestandard on "Communication of Electronic Health Care Record--Part 2: Domain Termlist" we carried out an analytical study about names of clinical documents, titles of generic sections, names of data elements, according to our terminological methods. We defined three layers of structures for clinical information: i) documents and sections, ii) clinical statements, iii) systematic details within statements. We prepared in correspondence many lists suitable to develop a principled coarse-grained markup for transmission and homogeneous browsing of disparate patient records across many institutions, without any preventive agreement on existing coding systems, data elements, record organization. This achievement is the basis for federated records, in particular for the virtual life-long patient record.

Syntactic-semantic tagging as a mediator between linguistic representations and formal models: an exercise in linking SNOMED to GALEN.

Natural language understanding applications are good candidates to solve the knowledge acquisition bottleneck when designing large scale concept systems. However, a necessary condition is that systems are built that transform sentences into a meaning representation that is independent of the subtleties of linguistic structure that nevertheless underly the way language works. The Cassandra II syntactic-semantic tagging system fulfills this goal partially. Within the GALEN-IN-USE project, it is used to transform linguistic representations of surgical procedure expressions into conceptual representations. In this paper, the proctology chapter of the SNOMED V3.1 procedure axis was used as a testbed to evaluate the usefulness of this approach. A quantitative and qualitative analysis of the data obtained is presented, showing that the Cassandra system can indeed complement the manual modelling efforts being conducted in the GALEN-IN-USE project. The different requirements related to linguistic modelling versus conceptual modelling can partly be accounted for by using an interface ontology, of which the fine tuning will however remain an important effort.

Structuring clinical information in healthcare records.

We carried out an analytical study about names of clinical documents, titles of generic sections, names of data elements, to prepare the European Prestandard CEN ENV13606-2 on "Health Informatics--Communication of Electronic Health Care Record--Part 2: Domain Termlist". The goal of the standard is to facilitate transmission and/or homogeneous browsing of clinical information from disparate patient records, without any preliminary agreement on coding systems, data elements, record organization. With the assistance of the members of CEN/TC251/PT27 and under the control of CEN/TC251/WG I and WG II, we defined three layers to structure clinical information (structuring records into complexes, complexes into statements, statements into details) and we prepared many lists suitable to represent coarse-grained information about the different constructs used in the above layers.

Standards to support development of terminological systems for healthcare telematics.

The Technical Committee on "Medical Informatics" of the European Committee for Standardization (CEN/TC251) is supporting developers of terminological systems in healthcare by a series of standards. The dream of "universal" coding system was abandoned in favor of a coherent family of terminologies, diversified according to tasks; two ideas were introduced: (1) the "categorical structure", i.e. a model of semantic categories and their relations within a subject field and (2) the "cross-thesaurus", i.e. a system of descriptors to build a systematic representation (called here "dissection") for each terminological phrase, coherent across diverse terminologies on a given subject field. The goal is to assure coexistence and interoperability (and reciprocal support for development and maintenance) to three generations of systems: (1) traditional paper-based systems (first generation); (2) compositional systems built according to a categorical structure and a cross-thesaurus (second generation) and (3) formal models (third generation). Various scenarios are presented, on the exploitation of computer-based terminological systems. The idea of "operational meaning" of terminological phrases within administrative and organizational contexts and the idea of "task-oriented details" are also introduced, to justify and exploit design constraints on terminological systems.

Exploiting the terminological approach from CEN/TC251 and GALEN to support semantic interoperability of healthcare record systems.

We apply the principles included in two CEN standards (ENV 12265, ENV 12264) to the analysis of the semantic structure of health record systems, to support their semantic interoperability. This result was made possible by dramatic methodological progress in the field of terminological systems--due to a worldwide evolution towards a new generation--and by the experience we acquired in the GALEN-IN-USE (formerly GALEN) project. The meaning behind names, content and context of record items and record item complexes can be considered as a 'semantic continuum'. This continuum is made explicit, by building a suitable paraphrase in a controlled language. We can then apply the principles we previously elaborated for the second generation of terminological system. Methodology and tools for generating a controlled language and a second-generation terminological system were developed and successfully used in the GALEN-IN-USE project and promising experiments were performed on elements of record structure listed in LOINC and SDM. In this way, the semantic structures of different record systems can be expressed by the resulting common formalism and thus, information units can be faithfully exchanged among different structures.

Building cross-thesauri with the support of UMLS.

The absence of a robust system of descriptors (cross-thesaurus) hampers the development of combinatorial terminological systems. We developed a tool (I-BROWSE) to produce a cross-thesaurus by analyzing terminological corpora. To facilitate the work of experts and to produce re-usable results, our application interacts via the Internet with the UMLS Knowledge Sources Server. We applied our tool on 2999 dissections on surgical procedures produced in the project GALEN-IN-USE, as a part of the internal Quality Assurance program. Support from UMLS seems mostly promising about descriptors on , , and . Additional assistance can be given to domain experts on less frequent descriptors on pervasive modifiers. We plan to apply our tool also to production of terminological standards in CEN, as a part of a worldwide process of gradual convergence and transformation of coding systems into second-generation systems and terminological services.

An ontological perspective on surgical procedures.

We studied terminological phrases on surgical procedures-from coding systems, controlled vocabularies, textbooks, and medical records-by an ontological point of view. A surgical procedure can be accurately described only by a set of sentences, in textbooks or surgical reports; a terminological phrase is just a short synthesis of that description. We outline three points of view actually used to construct a phrase, based on i) relevant phases and variants; ii) focus on structures, functions and pathologies; iii) evolution of information and decisions during the process of care. For each of them we discuss potential principles and mechanisms, with the aim of deriving guidelines to generate homogeneous systematic names, to organize regularities in classifications and nomenclatures, to normalize expressions in formal languages.

Coding systems and controlled vocabularies for hospital information systems.

Modern healthcare information systems are requested to support an increasing interaction among professionals (inside and across the borders of the hospital) and a growing integration of specialised tasks (provision of care, reimbursement, document retrieval, optimisation of resource use, clinical audit, etc.). Coding systems were conceived and optimised independently for various specific purposes. They are now facing each other and thus conflicting into this new environment; the solution will be in a more application-independent representation of concepts. Developments are going towards three complementary directions: (i) to separate different functions about the management of terms and concepts, and thus to produce more specialised software components; (ii) to develop a new class of software which is able to manage terminological diversity without imposing uniformity; and (iii) to enhance reusability of concepts, and facilitate a spontaneous convergence among controlled vocabularies.

Medical-concept models and medical records: an approach based on GALEN and PEN&PAD.

OBJECTIVES: To investigate the issues raised in applying a preliminary version of the GALEN compositional concept reference (CORE) model to a series of radiographic reports, and to demonstrate that the same underlying concept model could be used in conjunction with both a detailed, fine-grained model of medical records based on that used in the PEN&PAD project and with other more conventional medical-record models. DESIGN: Following analysis and representation of concepts from a set of reports, a single report was taken as a "case study." This report was analyzed in detail in its entirety and represented using each of the medical-record models. RESULTS: The reports were successfully represented within the limits of the study, but a number of significant issues were raised. CONCLUSION: The compositional approach plus the PEN&PAD medical-record model allowed detailed information in the radiographic report to be represented, including information about the inferences and the clinical process. The resulting representation was large, and more compact representations may be necessary for some systems. Alternative encapsulations of the information as might be used in such systems were successfully prepared. The compositional approach avoided many issues that often cause controversy in the design of traditional coding and classification systems, but it raised other issues, including the handling of ambiguity and underspecification, linkage to information not explicitly present in the report, and questions concerning the focus of individual concepts. All work is preliminary and definitive conclusions await further studies and systematic evaluation.

A model for the structured description of healthcare activities and related data.

Rehabilitation involves long-term, interdisciplinary processes. A model was developed, for the structured description of typical healthcare activities. Telematic services based on this model can support accurate data acquisition and communication among healthcare teams. Presentation of data within their context and according to the specific user's view is envisaged, based on deviations from the typical behaviours. Benefits are also expected in better understanding of the care processes themselves, easier comparison of different approaches, and diffusion of consensus-based knowledge.

KAVAS-2: Knowledge Acquisition, Visualization and Assessment System.

The objective of KAVAS-2 is the development of a tool, named KAVIAR, with which domain experts can make their knowledge explicit. It contains components for (computer assisted) knowledge elicitation and for machine learning. A key issue in KAVAS is the assessment of the quality of the classification and domain models built. Various quality measures are available and implemented in KAVIAR to assess the quality of models, specifically those developed from data bases by machine learning techniques.

Semantic standards for the representation of medical records.

Physicians developed their sublanguage (a system to represent medical concepts and their relations) to store and transmit general medical knowledge and patient-related information. Adequate formalisms are needed to obtain a standard representation of semantics of medical expressions for computer use. Comparison of the semantic contents of two expressions is possible only if a unique canonical form is defined; the transmission of medical facts or patient-related information is really meaningful only by defining a set of primitives (semantic categories and links) and the domains of values (concepts). These primitives must be harmonized to yield a "common core subset" of semantic categories and links. This subset provides a common basis; a procedure to register extension sets of primitives must also be defined, to comply with specific representation needs of specialties and classes of application software.

Converting the representation of medical data: criteria to code the underlying cause of death.

This study deals with a set of coding directives that were conceived for trained coding clerks and rely upon knowledge of their cultural background. These directives were formalized and adapted for computer use and in this form must rely upon a background of explicit medical knowledge. Medical data on death certificates are an invaluable source of information regarding prevention of major causes of death. These causes are coded and tabulated worldwide by means of the International Classification of Diseases (ICD). The ICD manual issues directives to achieve uniformity of coding throughout the world. The coder is required to trace back the flow of events which caused death and to single out the most significant concept from the statistical point of view. After emphasizing the problems encountered in the formalization, the methodological contribution of this work to the identification of a modular architecture for a system which represents and "reshapes" knowledge from medical documents is presented. Therefore we focus on the features of the two kinds of knowledge that must be supplied to a knowledge-based system, in order to enable it to perform semantic conversions on given medical data, namely: i) generic guidelines; ii) detailed medical knowledge.

Evaluation stages and design steps for knowledge-based systems in medicine.

After the early experiments in artificial intelligence a methodology is emerging around advanced systems for the management of medical knowledge. The stress is moving away from the implementation of prototypes to the evaluation. It is possible to adapt and to apply this to field evaluation techniques already developed in similar contexts of knowledge management (books, drugs, epidemiology, consultants, etc.). The time is ready for a further step: to envisage a methodology for the design of real systems that cope with the 'knowledge environment' of the user. Every stage of the evaluation process is re-examined here, and considered as a framework to define goals and criteria about a step of design: (1) the impact of the system on the progress of health care provision (priorities, cost-benefit analysis, share of tasks among different media); (2) effectiveness in the end-user's environment and long-term effects on his behaviour (changes in people's role and responsibilities, improvements in the quality of data, acceptance of the system); (3) the intrinsic efficiency of the system apart from the operational context (correctness of the knowledge base, appropriateness of the reasoning). The need to differentiate the test sample into three classes (obvious, typical, atypical) is emphasized, discussing the influence on both evaluation and design. In particular the difficulty of having 'gold standards' on atypical cases, due to the disagreement among the experts, leads to the definition of two alternative attitudes: the 'standardization mode' and the 'brain-storming mode'.