Social Media Mining for Pharmacovigilance

点击翻译按钮获取中文摘要

Project Summary Drugs undergo extensive testing in animals and clinical trials in humans before they are marketed for widespread use. Pre-market testing produces reasonably high quality information about the efficacy of the drug as a treatment for the condition for which it was approved, but gives a very incomplete picture of the drug's safety. It is only after a drug is marketed and used on a more widespread basis over longer periods of time that it is possible to identify other effects, such as rare but serious adverse effects, or those that are more common in the special subgroups excluded from the trial (such as pregnant women), or effects of long-term use of the drug, among others. Despite the increase in research in the past years exploring social media data for pharmacovigilance, and the evidence that it indeed can bring forward the patient perspective, there is no systematic approach to collect and annotate such data for research purposes. This renewal builds on our prior research and natural language processing (NLP) methods for social media mining in pharmacovigilance to make the collection of social media data about medication use precise and systematic enough to be useful to researchers and the public, alongside established sources such as the FDA's data and other public collections of drug adverse event data. It presents innovative methods to automatically collect and analyze longitudinal health data, piloting methods for interventions through the same media that can inform the public and help validate the automatic methods. As validation, we include a comparison to an existing reference standard for adverse effects that integrates FDA's data and HER data, as well as specific case studies focused on (Aim 3.1) the use of NSAIDs and anti-depressants in pregnancy and (Aim 3.2) factors for non-adherence.