Sunday, January 1, 2012

Tell Me Why

Another tasty treat allowing me to get my geek on can be found here in the online edition of Wired magazine, dealing with some of the problems that scientists have when studying very complex systems like the human body.

Jonah Lehrer outlines a drug study and the history of treatment for back pain as two ways to look at how gathering more information about a problem doesn't always lead to its solution. A drug company studied what's called the "cholesterol pathway" to figure out a new way to treat people who have high levels of the so-called "bad cholesterol" in their systems. This pathway is the way that the body and its enzymes use and then break down both good and bad cholesterol. It is pretty widely understood, and the drug company had a compound that would assist the good cholesterol in removing the bad cholesterol and thus improve the health of people whose cholesterol levels were too high. The drug did exactly what researchers thought it would do and was in its final trials before being approved for public prescriptions. Except that when it did exactly what the researchers thought it would do, it actually endangered people's health instead of improving it. Although all of the chemicals acted exactly as they had in experiments, the effect of those actions was the opposite of what was intended.

Lehrer's article is long but is worth the read as he explores one of the problems the drug company's situation demonstrates: Complex systems are not easily understood. Well duh, we might say, but the truth is that a whole lot of the dietary and medical advice given out today is based on assumptions about causes and effects that themselves might overlook crucial elements.

He mentions an experiment done by a Belgian psychologist in the 1940s in which people saw short films with a moving red ball and a moving blue ball. The red ball would move across the screen and touch the blue ball and then stop. The blue ball would then move in the direction the red ball had been traveling. When people described it, they said the red ball hit the blue ball and made it move -- they almost automatically spoke in terms of causation, even though the film showed nothing that supported that idea. That interpretation matched most people's experience of watching what happens when a moving object hit a stationary object, of course. But there was no evidence for saying what happened in the film was the same thing that happened in those other cases. People shortened the process and supplied their own understanding.

Lehrer points out that's the way the brain works and if it doesn't work that way we can't operate in the world. It's sort of like we're hard-wired to produce a narrative explanation for things we see, whether there's any kind of explanation like that or not. We simplify a process by removing steps. Instead of giving the most complete and accurate description:  "The red ball moved until it touched the blue ball. The red ball stopped. The blue ball then started moving in the same direction that the red ball had been moving," we take the shortcut and say "The red ball hit the blue ball and knocked it away."

Most of the time that works. In the experiment, we don't lose much information by using the shortcut version. But when systems become very complex, we don't know what kind of impact removing steps can have, and that's why the drug company's compound didn't do what they thought it would do.

Scientists, of course, usually expect this sort of thing. They're used to not knowing things, and even not really knowing what it is they don't yet know. But a lot of goofballs who insist on a rigid cause-and-effect relationship in systems where it may not even be possible to know all the causes mostly overlook it and make all sorts of wild claims about the universe and knowledge that they can't really back up. Your Friar, mired as he is in his traditional Christian theism, understands this because he takes a lot of things on faith as well.

Of course, he admits it.

No comments: