Monday, November 27, 2006
Beer and Pizza Diet
A fat man goes to his doctor for a new diet to help him lose weight. The doctor advises him to eat pizza and drink beer for every meal. One month later the fat man returns to the doctor weighing an additional ten pounds. "What kind of beer did you drink? Coors? Well why'd ya do that. You need to drink a microbrew."
You can imagine the endless game the doctor is trying to engage in. One of the things most white lab coat lab techs know is the Beer and Pizza diet. You are given an experiment to do. Most often the instructions are verbal. Sometimes they are sketched out on a chem wipe. Never, never, never (in R&D) are they fully written out. Material and methods sections are written only after the data that was desired has been obtained. It may have taken 10 tries but on the tenth try the data worked out in the cubicle scientists favor.
Is it wrong then to talk only about the tenth experiment? With a Beer and Pizza diet you can easily explain away the first 9. In experiment 1 the lab tech used the wrong brand of sodium chloride. In experiment 2 the cells were used at the wrong confluency. When you get the data that you want, the Beer and Pizza questions needn't be asked. If the data fits, you did the right thing. If it doesn't, you did something wrong.
I once worked with a supervisor who was charged with screening antibodies. When the antibodies failed the western blot test, the questions began. One lab tech was told to repeat a western blot because she had only boiled the protein samples 4 minutes instead of 5. She rolled her eyes and repeated the western. Same result. The supervisor threw up her arms stating that if this tech was foolish enough to boil protein samples for 4 minutes (instead of 5) that there could be a million other things she was doing wrong. One day a good antibody showed up and passed the western blot test. We boiled the samples 4 and 5 minutes and asked the supervisor to point out which were which. She explained that good antibodies sometimes work so well that you cannot see the subtle difference. "Could you show us an example of one that can show the subtle difference." She didn't respond.
Moving higher up the big pharma food chain we have the latest success of Pfizer and Celebrex. Celebrex was the first of a class of new painkillers, called cox-2 inhibitors, approved in December 1998, and it is the last one to remain on the market. Two years after Vioxx was taken off the market and Merck has spent millions defending themselves and their own Beer and Pizza approach to clinical trials, Pfizer is asking the FDA to expand the use of Celebrex. They want to give it to kids as young as 2 who have arthritis. FDA advisers will meet Wednesday to consider the company's request. Why would one cox-2 inhibitor be the cause of so many lawsuits yet another still be on the market and expanding its applications?