Chargement de votre page…
JMIR medical informatics
Par Jang B, Park E, Lee S et 8 autres.
Contenu en anglais
With the rapid expansion of health care data, multiple international standardized terminology systems have been developed and refined to ensure semantic interoperability, which is a crucial factor in facilitating collaborative use.
Ward réunit ces articles dans la même app que vos cartes et vos questions. Un seul compte gratuit pour Ward, Suture, Osler et Prognia.
This study aimed to evaluate the real-world use and applicability of standardized terminology systems by mapping clinical terms from routine cancer care and to propose guidelines that optimize mapping.
Clinical terms generated during cancer care at the National Cancer Center Korea were extracted and mapped to multiple target terminology systems: the Korean Standard Classification of Diseases; International Classification of Diseases, 9th Revision, Clinical Modification; Logical Observation Identifiers Names and Codes; RxNorm; Anatomical Therapeutic Chemical classification; and Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT). A multidisciplinary working group of health care information managers, data experts, and oncologists verified the mapping outcomes, calculated the cross-verification discordance rate, and compared the coverage across systems.
A total of 3472 source terms were identified across 7 domains: chief complaint, diagnosis, laboratory test, imaging test, surgery/procedure, medication, and radiation therapy. With postcoordination and predefined mapping guidelines, SNOMED CT achieved high overall mapping rates (98.3%-100%) across domains. Applying postcoordination as needed substantially improved exact semantic match rates, which exceeded 90% in most domains. However, more than 40% of source terms were broadly mapped in the imaging test and radiation therapy domains because procedural details such as imaging technique and radiation field configuration could not be represented in SNOMED CT. The cross-verification discordance rate between mappers, defined as the proportion of source terms with disagreement, remained low (0.3%-1.3%). RxNorm and Logical Observation Identifiers Names and Codes showed complementary strengths within their specialized domains, adding detail or specificity where SNOMED CT was less complete.
Standardized terminology systems are feasible and effective in encoding clinical terms for cancer care. SNOMED CT was the most comprehensive, and predefined mapping guidelines provide a structured framework for mapping and consensus resolution. Domain-specific systems should be used alongside SNOMED CT, and all systems should be updated regularly to reflect evolving clinical practice.
Publié dans JMIR medical informatics. Le résumé est reproduit tel qu’il a été publié et reste l’œuvre de ses auteurs ; lisez l’article complet à la source.
Prognia référence cet article à des fins de documentation et de formation. Il ne constitue pas un avis médical. Lisez toujours l’original et suivez les recommandations en vigueur et les protocoles locaux avant d’agir.
