Evaluation of an Automatic Vigilance System for Detecting Adverse Drug Reactions from Electronic Medical Records
- VernacularTitle:医薬品副作用自動監視システムによる副作用検出の評価
- Author:
Toshiaki IGARASHI
1
;
Aya KONNO
1
;
Hitoshi TSUKAMOTO
1
;
Ryoichi YANO
1
;
Kyohei WATANABE
1
;
Toshiaki NAKAMURA
2
;
Mikio MASADA
3
;
Nobuyuki GOTO
3
Author Information
- Keywords: pharmacovigilance; adverse drug reactions; electronic health records; signal detection
- From:Japanese Journal of Drug Informatics 2018;20(2):66-71
- CountryJapan
- Language:Japanese
- Abstract: Objective: We have developed an automatic vigilance system (AVS) that automatically reports adverse drug reactions (ADR) based on laboratory finding abnormalities and symptom keywords in electronic medical records. In this study, we aimed to evaluate the impact of detecting ADR using AVS on medical treatment.Methods: In AVS, drugs and their ADR signals, which would be detected and reported by AVS to pharmacists, were defined. Pharmacists evaluated the severity of these signals to identify whether these signals should be discussed with the doctor, continued to be followed up, or ignored. We investigated detection of ADR at University of Fukui Hospital between April 2016 and March 2017 along with whether prescriptions were modified because of ADR and the contribution of AVS. Assuming that ADR had worsened without appropriate treatment, medical expenses needed for treating severe ADR were calculated.Results: In total, 325 signals were defined for 146 drugs. There were 9,103 ADR signals confirmed by pharmacists for 8,531 subjects. Of these, 12 and 164 signals were discussed with the doctor and continuously observed, respectively. The pharmacist's suggestions based on AVS led to prescription modifications in 10 cases, corresponding to a reduction of 2.56 million yen in medical expenses in the event that these cases become severe.Conclusion: AVS assisted prescription revisions because of ADR and is thought to contribute to the prevention of worsening of ADR and reduction of medical expenses.