Mining Adverse Drug Side-Effects from Online Medical Forums

Pharmaceutical drugs prescribed for the prevention, treatment or cure of diseases can have adverse reactions or side-effects that lead to further health complications or sometimes even death. Most of the common side-effects of drugs, reported by their manufacturer, are based on clinical trials. However, not all possible side-effects are identified, as their detection is limited by the extent of the number and diversity of the participants in the trials. Online medical help forums where patients voluntarily provide feedback on the drugs they take, provide an excellent source for identifying the unreported side-effects of drugs. Mining for these side-effects would help patients make informed decisions about the suitability of a drug for their treatment and also for health authorities to take appropriate action against drug manufacturers. In this paper we present a Hidden Markov Model based text mining system that can be used to extract adverse side-effects of drugs from online medical forums.