2009 (v.17 no.2)
The Journal of Korean Society for Clinical Pharmacology and Therapeutics
Korean Society for Clinical Pharmacology and Therapeutics
Development of Auto Coding System on the Basis of MedDRA to Analyze Adverse Events for Clinical Trial
J Korean Soc Clin Pharmacol Ther | v.17, no.2, pp.164-173, Dec, 2009
Introduction: Currently, the adverse events are being
inspected, aiming to prove the safety from phase 1 up to phase 4 of clinical
study to Post Marketing Surveillance (PMS). Along with the movement of the
changing international society, an auto coding system that is developed based
on the existing MedDRA was devised to provide medically consistent and accurate
as well as commonly applicable terminologies in the field of medicine while
aiming to share the clinical safety of the medical supplies oriented by the ICH
Material and Methods: The auto coding system developed in this study basically connected the 67,159 LLT terminologies from MedDRA 12.0 and 5,583 terminologies from WHO-ART 2006, while improving the coding efficiency by utilizing the existing coding data. As for the comparison between the Copy & Paste Method (hereinafter called as ‘CPM’) and AE Mapper (hereinafter called as ‘AEM’), which was an auto coding system, the assessment was made in terms of efficiency, accuracy, and consistency. In addition, the difference depending on the level of medical background among the coders’ skill was measured when comparing CPM and AEM.
Result: In case of comparing CPM and AEM, the time consumed for CPM was 4.5 times greater than AEM. When comparing the accuracy, the file of the experiment 1 did not display a significant difference resulting CPM 86.7% and AEM 94.9% of the total average; however, the file of experiment 2 showed a significant difference as CPM was 62.0% and AEM was 92.4% in terms of the total average. When comparing the consistency, the file of the experiment 1 did not display a significant difference resulting CPM 89.0% and AEM 99.3% of the total average; however, the file of experiment 2 showed a significant difference as CPM was 79.6% and AEM was 98.9% in terms of the total average.
Conclusion: Based on the result derived by comparing CPM that copied and pasted AE and AEM (AE Mapper) that was an auto coding system for coding AE, it was known that the use of the auto coding system was superior in terms of efficiency, accuracy, and consistency. Furthermore, when using the auto coding system, there was no significant difference depending on the users’ medical background and past experience in terms of accuracy and agreement compared to the CPM. Therefore, it is suggested to seek a method to improve the mapping of the auto coding system and conduct a further study that applies such system.