AnnoPharma: Detection of Substances Responsible for Adverse Drug Reactions by Annotating and Extracting Information from MEDLINE Abstracts

In the 2013 study, researchers developed AnnoPharma, a system for extracting information about substances responsible for Adverse Drug Reactions (ADRs) from MEDLINE abstracts. AnnoPharma utilized semantic annotation, dictionaries, and ontologies from the Unified Medical Language System (UMLS) to identify pharmacological concepts through a three-step process. The system's performance was evaluated using precision, recall, and f-measure, with organ detection achieving high success (precision: 0.97, recall: 0.93, f-measure: 0.94), while ADR detection had lower success rates. AnnoPharma demonstrated higher precision (99%) and recall (97%) in detecting concepts compared to VigiTermes, another ADR extraction system. The study concluded that AnnoPharma could be a valuable tool in drug research and highlighted plans to improve its ontology and predictive abilities in the future.