24 Feb 2011

Equipoise

“Doctor, am I on active drug or placebo?” asked my patient one day. She was enrolled in a randomized controlled trial in liver disease.

“No, we deliberately make sure that neither you nor I know the treatment assignment so that we can assess the treatment effect fairly. Otherwise, I may treat you differently just because I know what treatment you are on,” I tried to explain.

“But if I am on placebo right now, are you delaying my treatment?”

Was I delaying treatment?

In this case, I was quite comfortable in explaining that although the mechanism of action of the new drug looked promising, there really was not any clinical data to show that it worked. That was precisely why the pilot study had to be performed. If she was not in that trial, the unproven drug would not be offered anyway. Meanwhile, she was receiving all other optimal treatments. This is the principle of equipoise. Some people hold the belief that clinical trials are only ethical when there is no existing evidence that one of the study arms is worse than others.

However, equipoise often does not happen in clinical trials. In medical jargons, we move forward to large phase 3 trials because the results of early phase 2 trials look good. If equipoise is essential, most clinical trials should be abandoned and drugs would be registered with preliminary data. This is obviously problematic. If new drugs are not fully evaluated, the efficacy may be inaccurately determined and important side effects may be missed.

In this month’s New England Journal of Medicine (2011;364:476-80), Franklin Miller and Steven Joffe argued that the concept of equipoise was flawed. Especially because new drugs are often very expensive nowadays, it is important to confirm it works before its introduction to the market.

I would not bore readers with the arguments. Suffice it to say, we are not doing favor to our patients by enrolling them to clinical trials. The aim, however, is to develop better treatments to benefit future patients. Participants in clinical trials may potentially be receiving less effective treatment or treatment with more side effects. This should be explained clearly to our patients, and we should be grateful.

17 Feb 2011

Valentine

Two years ago, seeing a pair of secondary school students kissing each other at a mall, my wife sighed, “What should we do when this happens to our girl?”

“This is simple,” I did not need to think. “I will come up to the boy, break his leg, and say, ‘This will teach you a lesson!’”

Last month, however, the mother of Angelina’s classmate told us how much her son liked her. “Your daughter is really lovely,” the grandma added. Meanwhile, that boy buried his face in his mother’s dress, trying to hide himself.

Did I break his leg?

While the mothers were laughing heartily, I just told myself, “I can understand that.”

Oh, I was too soft. Let me practise on Ken later.

10 Feb 2011

Relativity

One day, Angela asked Angelina at a restaurant, “Do you want one of these kid’s meals?”

Angelina glanced through the menu and was not interested. “Mummy,” she said, “I am not a kid. I am a big girl.”

A few days later, Angela asked her to tidy up the toys. Seeing no response from her daughter, she said, “I thought you were a big girl already.”

“Mummy,” Angelina did not even raise her eyes from the book, “you call a five-year-old a big girl?”

3 Feb 2011

Fake

Football fans can easily recall a few faked falls during important matches. Quite often a faked fall is sufficient to change the result of a match.

At the recent meeting of the Society for Integrative and Comparative Biology, Robbie Wilson reported a study on faked falls at football matches.[Science 2011;331:280] Among 2800 falls, only 6% were definitely deceptive dives, i.e. no contact between the players. At first sight, this figure appears surprisingly small. However, it makes perfect sense. According to game theory, if faked falls occur more commonly, the general scrutiny of dives would change, making the cost-benefit ratio less attractive.

Secondly, the closer to the goal, the more likely a fall was deceptive. This is because of the high payoff of creating a scoring opportunity. Besides, when it was closer to the goal, the success rate of faked falls was high. Free kicks or penalty kicks were offered to the attacking side in 80% of the time.

The observation can be extrapolated to other situations such as scientific publications.

1. Although faked works do occur, thank God they should be rare events.
2. Faked works are stimulated by payoffs. Since the payoff from a single scientific publication is usually small and the result of being caught is devastating, young researchers are at highest risk of crossing the line.

What if there is a sudden increase in the number of less established researchers? This is the worrying bit.