Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Comparison of data mining methodologies using Japanese spontaneous reports.

Kiyoshi Kubota1, Daisuke Koide, Toshiki Hirai

  • 1Department of Pharmacoepidemiology, Faculty of Medicine, University of Tokyo, Bunkyo-ku, Tokyo, Japan. kubotape-tky@umin.ac.jp

Pharmacoepidemiology and Drug Safety
|June 2, 2004
PubMed
Summary

Comparing five methods for detecting adverse drug reactions (ADRs), this study found that different data mining techniques identify varying drug-ADR signal combinations. The choice of methodology significantly impacts ADR signal detection from spontaneous reports.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Treatment Selection Patterns and Associated Outcomes for Biguanides and SGLT2 Inhibitors in Type 2 Diabetes: A Retrospective Database Study in Japan.

Drugs - real world outcomes·2026
Same author

Signal Detection Using Change Point Analysis to Identify Safety Signals from Spontaneous Reports in Aortic Stent Grafts in Japan.

Journal of medical systems·2026
Same author

Long-term outcomes associated with istradefylline versus catechol-O-methyltransferase inhibitors in patients with Parkinson's disease: a nationwide real-world cohort study.

Parkinsonism & related disorders·2026
Same author

Press-Through Packaging versus Bold and Slim Bottle-Type Containers: A Comparative Evaluation of Drug Packaging Usability in Patients with Rheumatic Diseases.

Patient preference and adherence·2025
Same author

Comparison of the Risk of Paradoxical Psoriasis between Monoclonal Antibody and Non-monoclonal Antibody Tumor Necrosis Factor-α Inhibitors in Patients with Rheumatoid Arthritis: An Observational Study Using a Claims Database.

Biological & pharmaceutical bulletin·2025
Same author

International Trends in Opioid Prescribing by Age and Sex from 2001 to 2019: An Observational Study Using Population-Based Databases from 18 Countries and One Special Administrative Region.

CNS drugs·2025

Area of Science:

  • Pharmacovigilance
  • Data Mining
  • Drug Safety

Background:

  • Spontaneous reporting systems are crucial for post-marketing drug safety surveillance.
  • Identifying true adverse drug reactions (ADRs) from numerous reports is challenging.
  • Data mining offers potential for efficient ADR signal detection.

Purpose of the Study:

  • To compare the performance of five distinct data mining methodologies in detecting potential ADR signals.
  • To evaluate the agreement and differences in signal detection across methodologies using real-world data.

Main Methods:

  • Bayesian method using Gamma Poisson Shrinker (GPS)
  • UK Medicines Control Agency (MCA) method
  • Bayesian Confidence Propagation Neural Network (BCPNN)

Related Experiment Videos

  • 95% Confidence Interval for Reporting Odds Ratio (RORCI)
  • 95% Confidence Interval for Proportional Reporting Ratio (PRRCI)
  • Analysis of Japanese spontaneous ADR data (1998-2000)

    • Main Results:

    • Significant variation in detected drug-ADR signal combinations across the five methodologies.
    • MCA and BCPNN showed no signals for low-frequency events (counts of 1 or 2).
    • RORCI and PRRCI identified signals in over half of low-frequency events, showing high concordance (kappa > 0.9) with each other.
    • GPS showed moderate agreement with MCA and BCPNN (kappa ~ 0.6).

    Conclusions:

    • The selection of data mining methodology critically influences the identification of potential ADR signals.
    • No single methodology consistently identified the same set of drug-ADR combinations.
    • Further research is needed to optimize ADR signal detection algorithms for improved drug safety.