OntoStudyEdit: a new approach for ontology-based representation and management of metadata in clinical and epidemiological research.

Alexandr Uciteli,Heinrich Herre

doi:10.1186/s13326-015-0042-0

Abstract

BackgroundThe specification of metadata in clinical and epidemiological study projects absorbs significant expense. The validity and quality of the collected data depend heavily on the precise and semantical correct representation of their metadata.In various research organizations, which are planning and coordinating studies, the required metadata are specified differently, depending on many conditions, e.g., on the used study management software. The latter does not always meet the needs of a particular research organization, e.g., with respect to the relevant metadata attributes and structuring possibilities.MethodsThe objective of the research, set forth in this paper, is the development of a new approach for ontology-based representation and management of metadata. The basic features of this approach are demonstrated by the software tool OntoStudyEdit (OSE). The OSE is designed and developed according to the three ontology method. This method for developing software is based on the interactions of three different kinds of ontologies: a task ontology, a domain ontology and a top-level ontology.ResultsThe OSE can be easily adapted to different requirements, and it supports an ontologically founded representation and efficient management of metadata. The metadata specifications can by imported from various sources; they can be edited with the OSE, and they can be exported in/to several formats, which are used, e.g., by different study management software.ConclusionsAdvantages of this approach are the adaptability of the OSE by integrating suitable domain ontologies, the ontological specification of mappings between the import/export formats and the DO, the specification of the study metadata in a uniform manner and its reuse in different research projects, and an intuitive data entry for non-expert users.

Highlights

There is a large variety of particular clinical and epidemiological research projects, which typically produce a large amount of data
In this paper we present and discuss a new approach for the ontology-based representation and management of metadata in clinical and epidemiological research, which is demonstrated by the software tool OntoStudyEdit (OSE)
Advantages of this approach are: 1. the adaptability of the OSE to intended aims and given needs by integrating suitable domain ontologies in a modular way; 2. the ontological specification of mappings between the import/export formats and the domain ontology (DO), such that no changes of the source code are needed by the replacement of the DO; 3. the specification of the study metadata in a uniform manner and reuse of which in different research projects; 4. an intuitive data entry for non-expert users

Summary

Introduction

There is a large variety of particular clinical and epidemiological research projects, which typically produce a large amount of data. The metadata include, in the context of this paper: The description of all data elements/items (e.g., questions of a questionnaire, measurements of an investigation) by particular attributes (e.g., question text, description of a measurement, unit of measure, data type, codelist); The description of a study structure, i.e., the grouping of the items and the description of the corresponding groups by suitable attributes (e.g., title, commentary) These groups may be modules within an assessment, complete assessments (e.g., questionnaires, interviews, physical examinations, laboratory analyses of taken specimen), or assessment groups (e.g., according to Uciteli and Herre Journal of Biomedical Semantics (2015) 6:41 study cohorts or to particular data acquisition time points), as well as all items of the study. The latter does not always meet the needs of a particular research organization, e.g., with respect to the relevant metadata attributes and structuring possibilities

Methods

Results

Conclusion