Can Machine Learning fix the Pharmaceutical Industry's productivity crisis? Print
Written by Abraham Heifets | LinkedIn   
Tuesday, 30 April 2019 13:29
 
Very few people would claim that we have all of the medicines we need, whether we're discussing untreated chronic diseases (e.g. Alzheimer's), neglected tropical diseases (e.g., Chagas), or re-emergent infectious diseases (e.g., antibiotic-resistant tuberculosis). Unfortunately, drug discovery is hard. Most people don't appreciate that there's a 66% failure rate before a drug candidate even gets to the clinic, and a 90% failure rate after that point, and that it's getting exponentially harder over time. Even achieving these rates requires large teams of smart, careful, dedicated, extensively-trained scientists who have spent many hundreds of millions of dollars on a wide diversity of experiments to prove - to other equally careful and skeptical scientists, regulators, insurance providers, doctors, and patients - that the new drug is safe, effective, and provides some advantage over existing standard of care. Drug hunters are not lacking in motivation, focus, expertise, drive, or skill, and yet we want better medicines faster. But what can be done?


Last Updated on Monday, 20 May 2019 19:56