A new mathematical model may help predict problems with drug formulations before stability testing.

The pharmaceutical industry spends a lot of money determining and validating the shelf life of drugs.  The standard protocol involves keeping drugs in stability chambers and sampling them for chemical changes over a range of time periods—perhaps every month for a year, for example. Researchers at the University of Belgrade in Serbia are trying to develop a mathematical model that will predict the shelf life.  They used a generalized regression neural network to predict the shelf life of extended-release aspirin formulations, then validated those results against actual stability chamber measurements.

For further information visit www.ncbi.nlm.nih.gov.

• 

•