Dr. Gupta:

The sport of cycling is really dirty with doping. They know how to fake the tests out. Certain drugs used for performance enhancing will not be picked up in the tests. When seconds ( or smaller ) can determine the outcome some will use anything to get that edge.

Susan
Phoenixville,PA

Doesn’t science already have a method for this? Usually a 95% confidence level is used as a cutoff. In the given example, there is only 66% confidence that the results were a true positive.

To nitpick a little, my understanding is that it’s actually only on the condition if Landis were truly negative, there would be a 66% chance that all 8 samples would test negative. If Landis were truly positive, there would be a much smaller chance that all 8 samples would test negative.

To evaluate the meaningfulness of a positive test result, you also need to know the sensitivity of the test and do a calculation with sensitivity. My understanding is that specificity measures how well the test classifies true negatives as negative. Sensitivity measures how well the test classifies true positives as positive. If the test is fairly sensitive, the chance that a person who is truly positive tests positive only 1 out of 8 times is probably very low.

For example, let’s pretend the sensitivity of the test is 95%. Then the chance that a person who is truly positive tests positive only 1 out of 8 times is 0.05^7 * 0.95, the seventh power of 0.05 multiplied by 0.95, which is something so low like 0.0000000…. Then you can easily see that if the test is sensitive, using 1 out of 8 as positive is quite wrong.

It seems that it’s the interpretation and understanding process that’s flawed. No test can make up for how humans understand its meaning. If everyone involved in the drug testing process, or any medical testing process, had to take a math and statistics test to test their actual understanding of what the results and numbers mean, it sounds like we would see a lot of cause for concern.

This type of skill is basically mathematical and computational logic. Do we need more training in that in medical schools and health professions?

It’s not quite a fair characterization to say that his conclusion is that Landis has a 34% chance of a falsely positive test. The real conclusion is that without knowing what the false positive and false negative rates are, Floyd’s chances COULD be that high or possibly higher.

The point I think is that it’s unconscionable to run an anti-doping program when you apparently have no idea what the odds are on the validity of your tests, for Landis or anyone else.

I think I was most struck by the editorial that was linked to Berry’s article. I was surprised that the editorial staff of Nature would so strongly agree with Berry, and so soundly thrash the current anti-doping regime.

tom

I don’t understand the big deal about doping. Millions have been spent developing swim suits that shape the body to reduce a swimmers performance by milli-seconds. Divers shave their body hair to reduce their splash. Bike riders shave body hair and wear special helmets to improve aerodynamics. Gymnasts wear special gloves and apply talc to improve their grip. Runners use special shoes with better traction, etc. Boxers trim weight to complete in a lower class. Just like doping, none of these actually make the competitor stronger, faster, or better than their competition.

Whatever happened to competing based on skill? What is the value of a competition that can be won by simply developing a piece of equipment competitors cannot afford? What’s the difference between doping and slimming the body’s aerodynamics if they both result in a reduction in lap time without an improvement in the competitors actual performance? Isn’t a technological advantage just legal cheating?

The problem isn’t doping. The problem is competitors who care more about winning than competing fairly. The ancient Greeks may have had the right idea about competing in the nude – at least their contests had more to do with strength, speed and endurence, than bio-engineering.

Grace said it best with, “No test can make up for how humans understand its meaning.” That 34 percent represents false positives for eight samples, but not the whole sampling itself. One positive result comes hand-in-hand with seven negative results — hardly a convincing reason to accuse someone of doping.

My concern is that most of the drug testing is conducted by the drug companies.

There are no third-party testing services that test drugs. Clearly with little FDA oversight and test studies that are conducted and controlled by the pharmaceutical companies–who is really looking out for the interests of the end user with respect to: adverse reactions, costs.