PV-OWL — Pharmacovigilance surveillance through semantic web-based platform for continuous and integrated monitoring of drug-related adverse effects in open data sources and social media

Abstract

The recent EU regulation on Pharmacovigilance [Regulation (EU) 1235/2010, Directive 2010/84/EU] imposes both to Pharmaceutical companies and Public health agencies to maintain updated safety information of drugs, monitoring all available data sources. Here, we present our project aiming to develop a web platform for continuous monitoring of adverse effects of medicines (pharmacovigilance), by integrating information from public databases, scientific literature and social media. The project will start by scanning all available data sources concerning drug adverse events, both open (e.g., FAERS — FDA Adverse Event Reporting Systems, medical literature, social media, etc.) and proprietary data (e.g., discharge hospital records, drug prescription archives, electronic health records), that require agreement with respective data owners. Subsequent, pharmacovigilance experts will perform a semi-automatic mapping of codes identifying drugs and adverse events, to build the thesaurus of the web based platform. After these preliminary activities, signal generation and prioritization will be the core of the project. This task will result in risk confidence scores for each included data source and a comprehensive global score, indicating the possible association between a specific drug and an adverse event. The software framework MOMIS, an open source data integration system, will allow semi-automatic virtual integration of heterogeneous and distributed data sources. A web platform, based on MOMIS, able to merge many heterogeneous data sets concerning adverse events will be developed. The platform will be tested by external specialized subjects (clinical researchers, public or private employees in pharmacovigilance field). The project will provide a) an innovative way to link, for the first time in Italy, different databases to obtain novel safety indicators; b) a web platform for a fast and easy integration of all available data, useful to verify and validate hypothesis generated in signal detection. Finally, the development of the unified safety indicator (global risk score) will result in a compelling, easy-to-understand, visual format for a broad range of professional and not professional users like patients, regulatory authorities, clinicians, lawyers, human scientists.