Knowing Drug Reactions before They Happen

Neural network predicts adverse drug reactions

(RxWiki News) Whether they are over-the-counter or prescription, most drugs have side effects. Predicting the most serious side effects was nearly impossible until recently. Now, researchers have created a way to spot the likelihood of severe side effects of any given drug.

Researchers from the University of Medicine and Dentistry of New Jersey have made a new computer tool that uses sets of information to predict serious reactions to drugs - known as adverse drug reactions, or ADRs. The tool is almost 100 percent accurate in predicting adverse drug reactions.

dailyRx Insight: This software tool enables health care providers to know which drugs are safest for each patient.

Researchers Peng-fang Yen, Dinesh Mital, and Shankar Srinivasan created a mathematical model of the human biological neural network. This artificial neural network is trained by feeding it data on various drugs and their associated ADRs. As more and more data are added, the artificial neural network learns to make correct predictions for adverse reactions.

This technology may calm the concerns of both patients and the medical industry. As the artificial neural network can spot adverse drug reactions earlier than current methods, it could save lives, reduce the number of drug-related complications, and help drug companies maintain the confidence of their shareholders.

The FDA officially entered 490,835 reports on adverse drug reactions in 2009. However, an estimated 2.2 million patients experience an adverse drug reaction each year. In 2009, 63,846 people died from adverse drug reactions. Adverse drug reactions describe the harm caused by taking normal doses of a medication. It is crucial to seek medical attention if you are experiencing an adverse reaction. Tell your doctor, as he or she may change your dose or take you off the drug altogether.

The article about this new technology is published in the International Journal of Medical Engineering and Informatics

Review Date: 
March 28, 2011