But Bill..... the FDA's rulings on VIoxx and other drugs are irrelevant. Merck is a respectable company. But they misbehaved and concealed data that harmed patients, for money. Why are you astonished they landed in trouble? Their behavior--muzzle the drug reps on key information, fight the earliest release of heart concerns--would have left them blushing if a patient confronted them. And that's all there is to it.
First, I enjoyed the eloquent discussion of what you have to go through.
Now... Walk a mile in my shoes. You're a smart guy Ian (big understatement) so I think I can convey this to you.
I wanted to tell you what I am doing at work this week to tweak a product we sell. Then I realized that the folks may get mad if I spill the recipe for the secret sauce. Damn... It was a great example.
The whole issue of cardiac events and analgesics is a big problem. If you have a population taking a brand new analgesic and you note that they have a higher-than-normal incidence of a cardiac event, what does that tell you? Well, it could be any of the following:
* The new drug is causing cardiac events
* Those who are at high risk for a number of different clinical conditions such as stomach bleeding are being steered away from the traditional analgesics (name your NSAID) and towards this COX II inhibitor.
Think of the kind of patient who should be taking a COX II instead of ibuprofen. That person probably has a chronic issue such as osteoarthritis. We know that long-term use of NSAIDs can cause gastric bleeding, and that's not a good thing.
Now, split out the folks who need long-term analgesics and/or anti-inflammatories from those who just need a few rounds of aspirin or ibuprofen. Are they identical cohorts? Of course not. I would bet you a dinner at the finest restaurant that those people who protocol would dictate should be getting a COX II (before anyone knew of the cardiac risk) are at higher risk of a cardiac event. You have a natural selection bias towards those who are sicker and will be having their "final event" sooner than later. (Sooner or later, one out of three of us is going to bite the bullet in this fashion. Medicine just delays the inevitable.)
Today I'm having to come in and do a multivariate analysis where someone who didn't know any better did a simple comparison of two cohorts where there was a selection bias. They drew inappropriate conclusions, and wanted to screw up a perfectly good predictive model. If you do epidemiologic studies, this is the way you have to do it. You have to account for everything going on that contributes to overall variability, and remove these extraneous sources so you can separate out the effect you want to measure.
Or...
You can run a very expensive but important randomized, double-blinded, controlled trial. And it must be large enough, run well enough, and survive the peer-review process.
I am telling people not to draw inappropriate conclusions today so someone won't screw up our product because they don't understand some basic principles of causality and experimental design. This is no different from folks at Merck telling people not to draw inappropriate conclusions from epidemiologic data, or RTC trials too small to show statistically significant odds ratios. I completely understand where the folks from Merck were coming from, and I would be one of those folks telling people to wait until the data appropriately tell us God's truth.
Yes, you can look at raw data and get a hunch. I do this all the time. I suspect these "hunches" were what drove people to do the published study that caused Merck to pull their product from the market.
That's the way the system works today. If you want it to work better, maybe you slow the approval process down, and demand that bigger (and MUCH more expensive) clinical trials be done first. But that's not what the system asks for today. The FDA approved the drug by today's approval standards. Merck met that standard.
Don't like it? Change the process!! And don't complain when it takes longer to get a drug or product to market, and lives are lost waiting for life-saving treatments. And those drugs, treatments, and medical devices that do make it are going to be more expensive. Ultimately I suspect this may happen, and that's not an altogether bad thing. Either that, or some people are going to need to understand the higher risks involved in getting the latest and greatest. Fortunately some healthcare providers I know preach this to their patients.
Believe me, if the insurance industry knew the risks associated with COX II inhibitors they would have been screaming from the top of the mountain. Go to your average insurance plan and pick out the top ten drugs by dollar volume. I bet you that the Average Joe on the street would recognize the name of 8 out of the 10. They are the newest, and most heavily advertised.
But no... Insurance companies took a BIG beating saying "No" to their patients who wanted Vioxx, Celebrex, or Bextra, or from making them pay 3rd tier copays that heavily discouraged the use.
IF they only knew, they would have saved beaucoup bucks. But they didn't. And collectively they had more (and better) data than the drug companies.
And that, BTW, is one reason the company I work for started a special relationship with the government to troll data for such issues. It's about time!
- Bill