医療用医薬品の使用成績調査における収集データ項目からみた診療情報とのデータ自動連携の可能性

抄録

Objective: This study evaluates the potential for automated data linkages between survey items collected during Japanese postmarketing surveillance (PMS) of pharmaceuticals and the medical health data stored in hospital information systems (HIS) by automatically relating the PMS survey items to the structured data (SD) in HIS. This relationship has not been explored previously; therefore, our findings offer fundamental insights for exploring automated data linkages between PMS survey items and medical health data.
Methods: The PMS survey items from 107 case report forms (CRFs) currently used at Kagawa University Hospital as of April 1, 2022, were analyzed. The survey items were categorized into major and detailed items based on the CRF units, and the frequency of occurrence for each detailed item was calculated. We considered the SD used in MID-NET^®, which is a Japanese national medical information database, as being analogous to the SD used in HIS and determined whether each detailed item had a one-to-one relationship with the SD in HIS.
Results: Twelve major items were identified, including a total of 83 detailed items. Among them, 25 items (30%) showed a one-to-one simple relationship with SD in HIS, while 28 items (33%) did not show a relationship. The remaining 30 items (37%) did not show any correspondence to the SD in HIS.
Conclusion: The results demonstrate that approximately 60% of the detailed survey items could be collected from SD. However, a physician’s medical judgment was needed for approximately half of the items. These findings will contribute to the realization of automated data linkages between the PMS survey items and the medical health data in HIS, thereby improving the efficiency of information acquisition required for PMS of pharmaceuticals.