AI can increase clinical trial probability of success by finding the best trial population

Clinical trials are designed to compare new drug candidates to the existing standard of care, to see if the new drug is better for patients. But patient populations are heterogeneous - a new drug may work better in one subgroup than another. 

AI can be used to identify the patient subgroups that will respond best to a drug planned for clinical trials. Then trials can be designed to test the drug on this population, to maximize the chance of seeing the beneficial action of the drug, reducing enrollment times and minimizing the risk of trial failure.