Statistical model to calculate potential success of new drugs

Researchers at Vienna’s Institute of Medical Statistics have presented a mathematical method which can be used to identify the characteristics necessary for the prediction of the efficacy of new drugs for particular individual.

Precision medicine refers to the tailoring of medication or therapy to the unique characteristics of a particular individual. It is known that drugs do not have the same effects on every individual. Scientists have therefore been looking to identify groups of people who respond well to the active ingredients in drugs, those who do not show side effects to them. 

Using the data obtained from these studies, statistical analyses can be carried out to accurately predict the efficacy of the drug. The researchers will use statistical techniques to filter biomarkers from the data, which will then be used to develop models to predict the subgroups of patients where the newer drugs will be more effective than current treatments. 

The study describes the statistical method as using algorithms to identify relevant biomarkers and to assess the statistical reliability of the predictions made.

The researchers mentioned how it could be possible to predict the patient subgroups where treatments will be effective and safe. The team said how it is an important step towards the improvement of the reliability of predictive models in precision medicine, and help to assist the development of medicine catered specifically for individuals.

In cancer studies, it can be predicted which patients is more likely to benefit from a new drug, extending their life expectancy. Regression models and variable selection methods are used by scientists to do this. 

The researchers mentioned how these techniques (statistical predictions) are often subject to variation – the less data that is available, the less accurate the prediction would be. At the moment, the main aim is to obtain large data sets, to minimise the range of variation, and to allow researchers, and in time, clinicians, to accurately predict a patients response to a new drug or particular therapy. This could lead to precision medicine being more readily available than it currently is.