Libertarians argue that the FDA prevents helpful drugs getting to market. Pro-regulation types tend to argue that the FDA rushes drugs through too quickly allowing too many dangerous drugs on the market. In the past few years, Phen-Fen, Rezulin and Baycol are just three drugs that have been pulled from the market because of adverse effects discovered well after Phase III trials were completed and the drugs approved for sale.
Besides the hidden human costs of restricting potentially helpful drugs from the market and the very visible human costs of not stopping potentially dangerous drugs getting to the market, there are real financial costs for the industry in not getting this right. Bayer has already paid out over $450 million for damages caused by Baycol and is looking at something between $1bn and $5bn more to come–not to mention the loss of more than $1 billion in annual sales. Just yesterday news started to surface that Avandia and Actos, both drugs for type-II diabetics may cause heart failure in some patients. There have also been serious suggestions that the other statins, Lipitor, Provachol and Zocor, don’t work and also have nasty side effects like Baycol. You can be sure that the attack-dog lawyers are just hoping that they can get their teeth into Pfizer, Merck, GSK and the rest over those issues. Yet we will see many, many more drugs coming to market over the next few years even before the flood of products from the genome revolution heads our way.
So what can we do to get out of this bind in which everyone loses but the lawyers? One of the keys to this issue is that drugs work differently for different people. Clinical trials, even the big ones required for phase III, often exclude too many types of people who end up taking the drugs in the real world, or are not long enough to discover some of the long-term effects (the Baycol example). Phase IV clinical trials (those that happen after the drug is on the market) are expensive and really only used when the FDA demands them.
Some of the brighter people in the industry have been talking about a combination of silicon and genomics eventually solving this problem. For one view, that of Kim Slocum at AstraZeneca, look at slides 75-95 of this long talk. I summarize Kim’s concept here (hopefully accurately!):
in the future we will understand the impacts of drugs on patients by recording their information electronically, matching it with their genetics, their treatment and their outcomes, and doing consistent long term monitoring and reporting of all of this information. Aggregated information from real care will then be collected, analyzed and delivered back to clinicians. This will eventually create a massive feedback loop that should show which drugs work best for which people over the long term. Inevitably this information will be incorporated into the drug development process, and the medical care process.
In other words drugs and their uses on different types of patients in the real world will become another facet of clinical trial research and of course another venue for FDA attention. In a meeting last year in direct response to my question Kim told me that he believed most of the pharma industry was on board with this new view of how pharmaceuticals should be used and monitored, despite its potential for showing up warts. Much like the related controversy over mistake-reporting, it would surely be better if we could see these types of systems in place, and find out the effects of pharmaceuticals before more drugs have to be pulled off the market in the face of human tragedies with only lawyers benefiting from the 20-20 hindsight.