We have often talked about the signaling effect of hubristic utterances like “We are the best”. “We are first in class, best in class, and I believe we will own this class for as long as it exists” said John La Mattina, Pfizer’s senior vice president for global research, in a November 30th, meeting with analysts. In retrospect, Mr. La Mattina must have known about the problems brewing with Torcetrapib, and this must have been troubling.One wonders about the statistical significance of 82 deaths out of 7,500 for the combined regime of Lipitor and Torcetrapib, versus 51 deaths among 7,500 people for the Lipitor regime alone. This is a statistic that should really be computed by re-sampling. One would take a 1.1% probability and a 0.7% probability and run 7,500 trials for each, noting the difference in proportions random numbers. Then repeat 10,000 times to come up with the probability of observing a difference of 31 in 2 groups of size 7,500. Not having the artful simulator around, one uses the formula below.

The standard error of difference between proportions is the square root of: p1 q1 / n1 + p2q2/n2, where p1 is close to 0.01 and q1 =0.99. Thus, the standard error is the square root of (1/100x 7500 + 99/100 x 7500), or 1/86, so the difference is some 22 standard errors away from expectations. This is a very big number, and re-sampling would not change it that much … but this is also an artifact of small standard errors for small proportions.

Despite this, this kind of exercise shows what is wrong with decision making based on the chances that something can do harm rather than based on the expected value of costs and benefits. One has seen many lines of evidence that these drugs have much life enhancing value. And is a difference of 31 deaths per 15,000 — let us say 150 years of total life expectancy — really comparable to the benefits that such a drug might have vis a vis reduced heart attacks, and enhanced life expectancy. Probably not even close.

Multiply this by the hundreds of drugs not approved, the thousands that do not get tested at all because they have risks, and the other tens of thousands that do not get invented because only the billion dollar companies can afford tests like this at all (consider the approximately 1 billion dollar cost to test a drug like this), and you see the incredible loss to life.

It would be like not using a system because it loses big 30 times out of 15,000 without taking into account how many times it makes big.

Roger Longman adds:

You heard, I suppose, that they have now completely abandoned Torcetrapib? No one knows the data, but the independent monitoring board killed it… apparently independent of the hypertension issue.

In any event, I completely concur about the predictive use of hubristic utterances, as we’ve seen with Rumsfeld, Cheney, Bush & Co. And it is particularly ridiculous with drugs.

Dan Grossman comments:

While I want to think further about the implications, my initial reaction centers on the massive misallocation to one fairly marginal drug category (anti-cholesterol) caused by the combination of dinosaur drug companies like Pfizer and overly expensive, overly time-consuming, mandated statistical testing and approval procedures, (This appears a fairly equal alliance and I am not blaming only the government.)

Anti-cholesterol (anti-lipids) is one of the only, actually the only, drug regimen where the favorable circumstance of government and medical recommendation, easy and understandable measurement through an annual blood test, and effective drugs and marketing, has been able to convince tens of millions of healthy consumers to take a drug every day at an annual cost of some $25 billion.

But this category has now succeeded. Cholesterol levels have fallen and, perhaps partially as a result although it is far from clear, so have heart attacks. And now there are generics (and Pfizer’s $12 billion Lipitor will in a couple of years also be a generic) that can bring about the same result for pennies a day. So the therapeutic problem is now taken care of at low cost, the $25 billion category can be cut to a couple of billion, with great cost saving and benefit to the public.

But Pfizer has not read Schumpeter, and with its $12 billion Lipitor grandly insists on “owning” this category “as long as it exists”, the way Proctor & Gamble or somebody might “own” the multi-billion dollar toothpaste category. It cannot give it up, even though there are probably only marginal advances in cholesterol lowering to be made in return for increasing billions of dollars chasing hoped-for slight advances, and any molecule with a slightly different mode of action that can desperately be combined with Lipitor and thus perhaps extend Lipitor’s patent life.

So Pfizer, and Merck, and Astra-Zeneca, and now Abbott with its multibillion dollar acquisition of KOS, will now spend a hundred billions dollars (literally, look at their research budgets) on this drug category that has already succeeded, instead of in a truly free, Adam Smith market where the natural incentives would be to spend the money in drug categories far more needed for the cure of disease and the advancement of human health.

John Tierney mentions:

What I would also like to know is the expected mortality rate for a study of this kind, a study involving subjects and drugs of these kinds. Clearly the Lipitor-only mortality rate of 0.68% must be well-below the threshold, otherwise there would be calls for it to come off the market as well. But there are none. And apparently the combination mortality rate of 1.09% is well over the expected mortality rate, and hence the rush to end the trials. Casino operations, I am told, make or lose money based on small differences in the percentage that the house is favored in games of chance. Pharma shouldn’t be reduced to such a state. For an interesting article on this …





Speak your mind


Resources & Links