Ontological Organization and Bioinformatic Analysis of Adverse Drug Reactions From Package Inserts: Development and...

Area of Science:

• Pharmacovigilance and Drug Safety
• Bioinformatics and Computational Biology
• Medical Informatics

Background:

• Licensed drugs can cause unexpected adverse drug reactions (ADRs), leading to significant patient morbidity and healthcare costs.
• Drug package inserts contain crucial ADR information from clinical trials and postmarketing surveillance.
• Formalizing ADR data from package inserts can improve understanding of side effects and promote rational drug use.

Purpose of the Study:

• To develop an ontological organization of traceable ADR information extracted from licensed drug package inserts.
• To provide machine-understandable knowledge for bioinformatics analysis, semantic retrieval, and intelligent clinical applications.
• To enhance the identification and analysis of adverse drug reactions for improved patient safety.

Main Methods:

• A generic ADR ontology model was proposed based on package insert content, covering ADR information and medication instructions.
• A custom Python-based natural language processing method was developed to extract relevant ADR information from package inserts.
• An ADR ontology was automatically built through biocuration and identification of retrieved data for further bioinformatic analysis.

Main Results:

• A specialized ADR ontology was constructed from 165 quinolone drug package inserts, containing 2879 classes and 15,711 semantic relations.
• ADR information and medication instructions for quinolone drugs were logically represented and formally organized within the ontology.
• Bioinformatic analysis identified frequent ADRs, contraindications, and mitigation methods for quinolone drugs, demonstrating the ontology's utility.

Conclusions:

• Ontological representation of ADR information from drug package inserts facilitates identification and bioinformatic analysis of drug-specific adverse reactions.
• The resulting ontology-based ADR knowledge source enhances the understanding of ADRs and supports safer medication prescription.
• This approach offers a structured framework for leveraging drug label data to improve pharmacovigilance and clinical decision-making.