Evaluating the Adverse Drug Reactions Learning Framework Using Novel Text Mining

摘要

Drugs are used to cure disease, but most of the drugs in the markets produce adverse drug reaction commonly known as side effects. This Adverse Drug commonly led to injury or death during medical treatments for patients. Only very few Adverse Drug Reactions are identified before a drug is marketed. Most drug adverse reactions are identified only after its intervention in the market. The post-marketing drug observation methods are based on the reports provided by the history of patient records and recommendations of doctors, this result in loss of identification of the earliest indications necessary to prevent the occurrence of such injuries or deaths. There are few online healthcare forums which provide some prior information to medications. The overall objective of this research is to extract reports of adverse drug side-effects from the information's available in online healthcare forums and use them as early indicators to assist in post-marketing drug observation. In this research we follow a novel method of extracting adverse side-effects of drugs from health care forum messages as a sequence labeling problem and present a Novel Hidden Markov Model (NHMM) based Text Mining system that can be used to classify a message as containing drug side-effect information. This is further extracted with the adverse side-effect mentions from it. This common forum is used in the training and validation of the NHMM based Text Mining system. The outputs of the system are consolidated and finally interpretation from the results is useful by gathering the pre medication results.