Country / Region
APAC
Tags
Clinical Practice, Mapping
This study investigates the comparative applicability and semantic precision of ICD-10 and SNOMED CT in the context of unstructured clinical narratives derived from abdominal region CT imaging reports. A dataset of approximately 100 cases, performed over one week at a tertiary medical center, was analyzed. The source terms extracted from the CT reports were mapped to both terminology systems (ICD-10 and SNOMED CT) in their raw form and after clinical interpretation. Results demonstrate that SNOMED CT consistently provided higher coverage and more granular, clinically meaningful representations than ICD-10, particularly when clinically interpreted terms were used. The study categorized semantic alignment into exact matches, broad matches, and no matches, and used a Venn diagram to illustrate the conceptual overlap and divergence between the two terminologies. These findings highlight the potential of SNOMED CT to facilitate more comprehensive and clinically meaningful standardization of healthcare data, extending beyond traditional diagnosis-based coding frameworks, particularly in the context of multi-center research and real-time clinical data analysis. Furthermore, the results emphasize the need for trained terminology experts in the mapping process to ensure accurate representation of clinical content. Overall, the study offers meaningful insights into the comparative advantages and limitations of ICD-10 and SNOMED CT, reinforcing the value of SNOMED CT as a more suitable framework for the standardized representation of clinical information in both research and practice.
Description
The purpose of this study is to compare the applicability and accuracy of standardized terminology systems‚ specifically ICD-10 and SNOMED CT‚ in multi-center research and clinical data-driven studies. While ICD codes have traditionally been used for classification, SNOMED CT offers a more granular and clinically meaningful framework, enabling more precise representation and management of patient data. Distinct from prior structured code-based research, this study leverages unstructured clinical data to evaluate the coverage and applicability of the two terminology systems. In addition, mapping was performed by interpreting unstructured data to capture clinical meaning. It is anticipated that these findings will contribute to the development of more robust and clinically relevant standardization methodologies that extend beyond diagnosis-centric paradigms.
Scope
SNOMED CT encompasses a wide range of clinical concepts, including diseases, symptoms, treatments, test results, medical procedures, and medications. This enables healthcare professionals to handle patient data from various perspectives in clinical practice.
SNOMED CT provides greater granularity and clinical relevance than ICD-10, making it more suitable for real-time clinical applications, multi-center research, and comprehensive data standardization. Granularity, in this context, refers to the capacity to express the same clinical concept with detailed distinctions such as severity, anatomical location, underlying cause, clinical context, and timing.
SNOMED CT is applied in electronic health record (EHR) systems in healthcare institutions, allowing for more accurate and consistent documentation of patient information. This provides standardized data that can be used for multi-center research or national-level data analysis.
How SNOMED CT will be used
The SNOMED CT International Edition (version 2025-01-01) was used for the mapping process. Post-coordination was not applied in this study.
This study utilized approximately 100 abdominal region CT records performed over the course of one week at a tertiary medical center. The dataset consisted of unstructured clinical narratives from various abdominal imaging modalities, including abdominopelvic CT, biliary-pancreatic CT, appendix CT, genitourinary/gynecologic CT, and abdominal trauma CT. Mapping to SNOMED CT and ICD-10 was conducted on both the raw (uninterpreted) and clinically interpreted data. Raw data refers to the source terms extracted directly from the CT reports, whereas clinically interpreted data refers to source terms refined based on clinical knowledge. For example, "Spleen size, 13 cm" represents raw data, while "splenomegaly" represents the clinically interpreted version of the same information. Standard terminology browsers were employed to identify relevant codes, and the degree of semantic alignment was classified into three categories: exact match, broad match, and no match. Additionally, a Venn diagram was used to visualize the conceptual overlap and divergence between the two terminology systems, highlighting shared and unique mappings.
Why SNOMED CT will be used
Contact
