Putting odds on FDA approval
KAI RYSSDAL: You've heard the pharmaceutical industry joke about the costs of getting new drugs to market, right? The first pill costs half a billion dollars to make. Everything after that's pure profit.
Part of the cost is developing drugs that never get approved. Researchers in Boston said today they've figured out a way to get around that problem. Helen Palmer has more now from the Marketplace Health Desk at WGBH.
HELEN PALMER: I tell you, this sounds like the drug industry's dream.
ASHER SCHACHTER: This is a model that can predict accurately which early development drugs will be safe, effective and profitable.
's a doctor and informatics expert at Boston's Children's Hospital. He uses a complicated computer probability tool called a Bayesian
He feeds in lots of clinical trial data from both successful and failed drugs. Hey presto! Schachter says he can indeed predict future triumphs or disasters.
That knowledge could help a pharmaceutical company cut its losses on a drug doomed to fail.
SCHACHTER: However, that drug has also had tremendous time and money invested into it. And that investment would be lost if we didn't learn from that information.
So that failure could help the company design a better drug. Schachter says it's vital drug companies share all their trial information.
That would make Rodney Hayward
happy. He's a professor of public health at the University of Michigan. He says that currently, drug companies tend to bury their failures and hate to share trial data.
RODNEY HAYWARD: What is different though about this is that although there's still risk to the company in terms of sharing information and finding bad things, it's before they've put in most of the money into the drug.
And on the subject of money, Schachter says they also did some cost modelling of this new approach.
SCHACHTER: We could reduce drug development costs by $283 million per approved drug.
Not only did they cut development cost, the model actually managed to increase a drug's profitability by $160 million.
So the next question is, what are the odds of the drug companies buying into this model?
In Boston, I'm Helen Palmer for Marketplace.