当前位置： SCI文献检索 > INTERNATIONAL JOURNAL OF MEDICAL INFORMATICS期刊下所有文献 > Semantic similarity-based alignment between clinical archetypes and SNOMED CT: an application to observations.

Semantic similarity-based alignment between clinical archetypes and SNOMED CT: an application to observations.

Abstract：

PURPOSE:One of the main challenges of eHealth is semantic interoperability of health systems. But, this will only be possible if the capture, representation and access of patient data is standardized. Clinical data models, such as OpenEHR Archetypes, define data structures that are agreed by experts to ensure the accuracy of health information. In addition, they provide an option to normalize clinical data by means of binding terms used in the model definition to standard medical vocabularies. Nevertheless, the effort needed to establish the association between archetype terms and standard terminology concepts is considerable. Therefore, the purpose of this study is to provide an automated approach to bind OpenEHR archetypes terms to the external terminology SNOMED CT, with the capability to do it at a semantic level. METHODS:This research uses lexical techniques and external terminological tools in combination with context-based techniques, which use information about structural and semantic proximity to identify similarities between terms and so, to find alignments between them. The proposed approach exploits both the structural context of archetypes and the terminology context, in which concepts are logically defined through the relationships (hierarchical and definitional) to other concepts. RESULTS:A set of 25 OBSERVATION archetypes with 477 bound terms was used to test the method. Of these, 342 terms (74.6%) were linked with 96.1% precision, 71.7% recall and 1.23 SNOMED CT concepts on average for each mapping. It has been detected that about one third of the archetype clinical information is grouped logically. Context-based techniques take advantage of this to increase the recall and to validate a 30.4% of the bindings produced by lexical techniques. CONCLUSIONS:This research shows that it is possible to automatically map archetype terms to a standard terminology with a high precision and recall, with the help of appropriate contextual and semantic information of both models. Moreover, the semantic-based methods provide a means of validating and disambiguating the resulting bindings. Therefore, this work is a step forward to reduce the human participation in the mapping process.

journal_name

Int J Med Inform

journal_title

International journal of medical informatics

authors

Meizoso García M,Iglesias Allones JL,Martínez Hernández D,Taboada Iglesias MJ

doi

10.1016/j.ijmedinf.2012.02.007

subject

Has Abstract

pub_date

2012-08-01 00:00:00

pages

566-78

issue

eissn

1386-5056

issn

1872-8243

pii

S1386-5056(12)00039-1

journal_volume

pub_type

杂志文章

在线工具

Semantic similarity-based alignment between clinical archetypes and SNOMED CT: an application to observations.