The Rest of the Story

I've been reading Scientific American since I was about 14.  I haven't keep up a subscription for years, but I check their website fairly often.  It is rare that they publish anything about the governement.  In the last issue, though, they have an article about the FDA.  Seems as though everyone is getting into the act.  Their article, though, is worthy of attention, because it has an idea that I have not seen published anywhere else. 

Avoiding Another Vioxx
Guarding against unsafe drugs means major changes
By Sara Beardsley

"Too cozy" is how Senator Charles Grassley of Iowa referred to the relationship between the U.S. Food and Drug Administration and pharmaceutical companies in the aftermath of the Vioxx debacle. Although the FDA admits no substantive lapse in vigilance, congressional pressure and consumer outrage are forcing officials to rethink the agency's role and perhaps even the drug approval process itself.

Vioxx, of course, is the cox-2 inhibitor that was taken off the market, voluntarily, by Merck, after having been shown to increase the risk of vascular  problems.  Although this is only one of many such cases, it has generated a lot of media attention.  the author goes on to review the impact of the 1992 Prescription Drug User Fee Act, which I've discussed previously.  She mentions the suggestion that the government split the functions of the FDA: keep the approval process where it is now, but set up a separate agency for monitoring drug safety.  Personally, I think this is a great idea, although there will be considerable political resistance to the establishment of a new agency. 

In case any readers are thinking that it would be a waste of taxpayer money to set up a new agency, let's try to put things in perspective.  Currently, the FDA employs about 10,000 persons, and has a budget of $1.7 billion.  They are responsible for regulating products that account for about 20% of all consumer spending in the USA.  The Vioxx problem has been estimated to have resulted in 55,000 premature deaths.  The terrorist attacks on 9/11/2001 resulted in about 3,000 premature deaths.  So far, we have spend hundreds of billions of dollars in response (most of it wasted in Iraq, but that's another story).  

If we set up a separate drug safety agency, and, let's say, it employs 1,000 people and has a budget of $170 million, and it prevents just one incident like the Vioxx problem every ten years, it would be an incredible bargain.   And that's just looking at the economic side of things.  From an humanitarian perspective, it is just the right thing to do, never mind the economics.

Earlier, I mentioned that the SA article contained a new idea, or at least, an idea that's new to me. 

The Vioxx withdrawal may also mark a turning point for the industry. Instead of seeking blockbuster drugs aimed at large numbers of consumers, companies may focus on niche medications designed for particular groups of ailing people. Garret A. FitzGerald of the University of Pennsylvania School of Medicine, an expert on COX-2 inhibitors, notes that Vioxx might have been able to stay on the market if its use had been restricted to people who have a low risk of cardiovascular disease but a history of gastrointestinal problems. (Vioxx is less likely to irritate the digestive tract than nonprescription painkillers.) "Here are drugs that are very useful in a segmented market," he says, "and their usefulness is being put at risk by the pursuit of a blockbuster strategy. I think, inevitably, the trend is going to be toward more segmented indications."

I'm not sure what to make of this.  Presumably, Merck was not directly responsible for the wide use of Vioxx.  After all, they cannot sell a single pill without a physician writing a prescription.  Yes, they have a marketing machine, and yes, that machine is quite powerful.  But the physician has a responsibility to weight out the risks and benefits carefully. 

Hindsight shows that too many people jumped on the cox-2 inhibitor bandwagon too quickly.  One could argue endlessly about where the fault lies for that phenomenon.  My argument is that the medical profession has to accept some of the blame; it is not merely a case of a drug company going for a "blockbuster" drug.  We need to be more cautious about accepting new classes of drugs, take more time before expanding their use to wider patient populations, and pay careful attention to adverse effects that occur at a low frequency.  If a drug is given to tens of millions of people, even a rare adverse effect can have a big impact.

All of this reinforces the need for a large, national database to track prescriptions and outcomes.  There would be considerable potential for abuse of such a database, so it would have to be managed in isolation from industry influence (not to mention law enforcement and intelligence agencies.)  It seems clear that the only way to do that would be to establish a separate agency. 

As for the idea that drug companies would start to move away from the mindless pursuit of blockbuster drugs, that does not seem too likely.  The allure of an easy buck is pretty hard to resist; drug development cycles are long (12 years, on average) and development costs are high.  What we probably will see is a more prudent, limited approach to the marketing of drugs.  If the companies have any sense, they will instruct their sales force to avoid the temptation to boost sales by hinting at off-label use of the products.  Likewise, physicians will have to be cautious about expanding the use of drugs to new, untested indications. 

Or, How Brain Chemistry Tempts Us To Do Dumb Things

Warning: this is a rambling-armchair-musing kind of post.

I learned in a seminar on psychoanalysis that the original meaning of the word seduce was: to traduce from duty, meaning that something is seductive if it pulls you away from what you ought to be doing. 

I learned in high school Humanities class about the concept of Platonic Ideals.  Of course, since most ideas are connected to most other ideas, in some way; and most ideas have been written aobut somewhere on the Internet, there is a paper about the Historical Context of Cognitive Therapy (135 Kb PDF) that you can read if you are snowed in or otherwise inclined to burn up a few clock cycles in your brain.  The paper explains how Plato's concept of Ideal Forms is part of the foundation of psychotherapy. 

The fundamentals of cognitive therapy have been around since the ancient Greeks. The idea that our perceptions determine the reality that we experience was clearly a focus of Plato' idealism. In a classic story, known as "Plato's Cave", Socrates describes how a group of men who are chained facing a wall observe shadows dancing across the wall in front of them. They have never known that these shadows are due to figures near the entrance to the cave moving behind them in front of a candle. To these men, the shadows are reality. One day one of the men turns around and sees that there are figures moving behind him casting their shadows across the wall. From that day on, the "reality" of the shadows no longer exists. Reality is now defined as the figures that one sees outside the cave.

We might view cognitive therapy as the attempt to get the patient to unchain himself and see outside the cave. Plato's philosophy was the search for "ideal forms" or qualities---whether Socrates spoke of geometry, love, justice, or political structure. The Platonists believed that these ideal forms were innate to the human mind and that one only needed to "educate" (that is, draw it out of the mind) through questioning. Truth and reality were entirely determined by the Platonic ideals. Socrates attempted to demonstrate this in one of the dialogues by asking a series of directed questions of an uneducated man about the principles of geometry. From these questions Socrates demonstrated that the "ideals" of geometry were already innate, but only had to be extracted through questioning (Cornford, 1957)..

Another explication, somewhat more succinct, is here:

Quite early on in life, we learn to abstract from individual instances and so reach a general idea of concept. Thus from our experience of observing pairs of objects (socks, gloves, parents, whatever), we come to form a notion of the number 2. Once we have done this, 2 becomes quite real for us; 2 exists, although not of course in the same straightforward way that the socks, the gloves, the parents, etc. do. This is an example of what is called a platonic ideal.

Razib, writing on the blog, Gene Expression, explains how the tendency to think in terms of ideal forms, as opposed to thinking about the objects themselves, can lead to false generalizations.  He focuses on the subject of racism:

To Tom, I was "black." This makes sense, my skin is brown, and West Asian Muslims and Europeans who came to Southern Asia often referred to the natives as "black." Nevertheless, one time Tom referred to a Cambodian girl we both knew as "black." She was somewhat lighter in complexion than I was, but her East Asian features seemed to be a sign that she wasn't black. But OK, she was brown, so perhaps that was close enough. Then one day, Tom referred to Patty Chen (last name changed, but close enough) as black. As you can tell, Patty Chen was Chinese. When I pointed this out to Tom, he agreed she was Chinese, "a Chinese-type-black."

What I'm trying to get across is in Tom's world there were two high level racial classifications: black and white. Or more properly, non-white and white. This is a common theme in American culture, shaped as it has been by the tension between the white majority and the black minority. This dichotomy continues to loom large, resulting in the The New York Times series "How Race Is Lived in America" only profiling black and white Americans, ignoring the 15% who don't easily fall into either category, or the ABC special "America in Black and White," as if that is the totality of America (in large parts of American Latinos or Asians are the dominant minority).

This, in fact is what I was going to write about.  However, since Razib already has done so, there is not much point in me echoing it.  Instead, I will examine the more general impact of the kinds of cognitive errors that can result from the formation of incorrect generalizations. 

The human mind is designed to form generalizations.  In general, this is useful; it allows us to simplify complex problems to the point that they become tractable.  An example is the (enormous) problem of figuring out whom you can trust.  Trusting the wrong person can lead to serious consequences.  Yet, there is no simple, objective test that one can apply to another person, to determine if that person is trustworthy. 

Simply put, it is impossible to predict specific instances of human behavior.  To solve that problem, we rely upon generalizations.  We classify people according to ethnic divisions, class divisions, political allegiances, religious practices, etc.  We tend to  place more trust in a person, if that person is sufficiently similar to ourselves.  There is some sense to this.  The more two people have in common, the more likely they are to exhibit similar behavior in similar circumstances.  Thus, it is easier to anticipate the behavior of persons who are like ourselves.  

Because of the fact that such generalizations make it easier to "solve" problems -- even if the solution turns out to be incorrect sometimes -- it is rewarding, psychologically, to make use of them.  Sometimes, the reliance upon such generalizations leads to massive, systematic, errors.  Such errors can afflict entire populations, or disciplines. 

For example, ancient astronomers had a notion that the perfect Universe would have the Earth in the center.  They surmised, accordingly, that the planets had to circle the Earth.  Never mind that objective reality did not mesh with this theory: they came up with complex geometrical patterns that could describe their observations of the motion of the planets.  They had a problem, which was that they assumed that the orbits must be circular, since circles are perfect, and all things in heaven must be perfect.  To explain the back-and-forth motion of the planets, they came up with the idea that the planets moved in large circles around the Earth, and that while following these circular orbits, they repeatedly moved in smaller circles, or epicycles.  The idea of circles upon circles is appealing, but it turns out to be incorrect. 

Another example is the belief, commonly held by biologists, that the perfect adaptation for a parasite was a form of commensialism or mutualism.  That is, the parasite are host would evolve to the point that either the parasite had no deleterious effect upon the host, or that each benefited the other.  This is an appealing -- even seductive -- notion.  But if you do the math, it turns out to be incorrect. 
From Paul Ewalds book, Evolution of Infectious Disease:

Few ideas have been so ingrained in the literature of medicine and parasitology as the idea that parasites should evolve toward benign coexistence with their hosts. 

It is possible to mathematically model the evolution of populations, and the changes that occur in the genome over time, under various circumstances.  If you do that, and make sure that your models correspond with experimental and naturalistic observations, you find that virulence sometimes is the best strategy.  Dr. Ewald wrote an entire book about this, and concludes that the notion of mutualism being the perfect adaptation is false, despite its apparent concordance with some kind of Platonic Ideal. 

When a person is trying to make sense of a series of observations, and suddenly realizes that the observations all seem to fit some kind of pattern, it is tempting for that person to conclude that --Aha! --all similar situations fit that pattern. 

Like racism, or the geocentric theory, or the notice that parasites inevitably evolve to a benign form, ideas that mesh with a beloved Platonic Ideal are seductive. 

I remember reading a book, sometime back when I was in high school, in which there was an exercise: construct a proof of the fact that all numbers of the form ABCABC are evenly divisible by 7.  Of course, the first thing to do is to get a calculator and try a few:

123123/7=17589
456456/7=65208
789789/7=112827

The proposition seems correct, after three decidedly nonrandom trials.  When I was in high school, the proof was not too difficult for me to come up with.  I recall the Aha! sensation that came from finding the proof.  However, I also recall many instances in my life in which I had an Aha! experience, only to realize later that I had made an inaccurate generalization. 

What is it about the concordance with a Platonic Ideal that makes a false conclusion seem appealing?  I suspect that, in part, it is the gratification that comes from the Aha! experience, when one realizes that a bunch of observations all seem to fit into some kind of organized pattern.  At the moment that experience occurs, it seems that the problem is solved. 

Of course, one could use the calculator to run through all possible instances of ABCABC, but that would take a lot of work.  If you come up with a proof that the answers always will fit a particular pattern, it eliminates the need for all that work.  That, in psychological terms, is an outcome that is better than what one initially expected. 

In nonhuman vertebrate  animals, we know that when an animal anticipates the outcome of a behavior, and the outcome is better than expected, the animal gets a little burst of dopamine in the brain.  That little burst of dopamine is a reward.  It prompts the animal to remember the behavior, and to try it again.  This mechanism has been implicated in the pathophysiology of addiction.

Thus, when we see something that would be difficult to understand if we were to try to analyze all aspects of the issue, but which, at least superficially, seems to fit some kind of Ideal pattern, it is extremely tempting to mentally fit the thing into that pattern.  Certainly, it is much easier to do that, that actually to do the math. 

I suspect that, not only is it easier, it actually is pleasurable to do so.  It makes us feel smart.  Ironically, some of the stupidest things that people do are caused by this phenomenon.  Razib, in his Gene Expression post,  gives examples of dumb things racists think, falling into this neurochemical trap. 

There was a time in my life that I thought that the occurrence of an Aha! moment was a clue that I was on the right track in solving a problem.  Since then, I have come to be suspicious.  I find it useful to be introspective, and notice when that seductive feeling occurs.  When I am just so sure that something is right, I try to step back and ask myself what kind of mistake I am making. 

Seduction comes in many forms.  The pleasures of the flesh are perhaps the most obvious, but it is important to realize that intellectual seduction is just as dangerous; perhaps more so.  The dopamine tweak they provide can, very well, pull us away from our rightful duty.