The DSRU’s Second Conference on Big Data for Pharmacovigilance has been a great success. Delegates from regulatory bodies, the pharmaceutical industry and academia have heard presentations from high calibre speakers, examining the challenges and opportunities for use of big data sources for the benefit of pharmacovigilance.
Gianluca Trifirò from the University of Messina set the scene with an overview of the current status of big data in pharmacovigilance, including developments since our first Conference on Big Data for Pharmacovigilance in 2018. Niklas Noren from the Uppsala Monitoring Centre detailed some lessons learned in developing and deploying machine learning for pharmacovigilance, including the observation that it is possible to make use of even very raw data, and the importance of detecting duplicates in spontaneous reports.
Michael Hughes from SciBite explained how drug labels can be mined for safety risks related to genetics and how semantic enrichment can help differentiate between a term used for both an indication and a side effect (such as “headache”). John Rigg from IQVIA presented the ‘what’, ‘how’ and ‘where’ of artificial intelligence and machine learning in healthcare and provided a list of conditions that must be satisfied in order to ensure the greatest chance of success in using machine learning.
Will Dixon from the University of Manchester detailed some recent advances in Digital Epidemiology. Now that smart phones are owned by 80% of people they provide opportunities for drug safety studies with large enrolment and repeated measurements. David Martin from the FDA outlined the development of the MyStudies app, an open source app that is secure to regulatory standards which can be used to gather data from patients via their smart phones.
Pantelis Natsavias from the Centre for Research & Technology Hellas, Thessaloniki outlined the development of OpenPVSignal, a tool for electronic exploration of signal reports. Eva-Lisa Meldau from the Uppsala Monitoring Centre described the use of natural language processing to remove patient identifiers from case narratives, emphasising the importance of context and the need to reserve the relations between words that are normally used together.
An engaging panel discussion with audience participation provided ideas to explore further in our next conference on Big Data for Pharmacovigilance.
For further information about our conferences click here.