[protomyth]

I worked on one in the 90's that was $50,000 a day for three years. The drug company, assuming the trial was successful, had about 4 years to make all its money back before the patent expired. Testing blood, urine, and stool for multiple things is expensive.

Intellectually, I can understand the need for control groups, but I still think it's immoral. When you stare at a spreadsheet and see 70% of the control group is dead because some random number generator sorted them there like hell's own sorting hat, and the 98% of the people getting the drug are alive, you have no business talking about statistics. That graph will haunt me til my dying day.

[s1artibartfast]

You didn't kill 70% of the control group, you saved 98% of the experimental group.

Also, with the appropriate trial design, you can stop the control and transfer patients once you see these big differences. Same thing should happen if you see the opposite, e.g. killing 70% of your experimental group.

[protomyth]

Everyone knew the results of "standard treatment", that's why they were researching the drug. There were years and years of statistics.

We killed 70% of the control group. Doomed by a random number generator.

[disgruntledphd2]

Yeah if it's a potentially life saving treatment you really need that design that allows peeking at the results.

[Khelavaster]

Peek-and-do-more is compatible with double-blindness, as long as the peeker-and-decider is different from the dispenser and patient.

[refurb]

It's not unusual for trials to be stopped earlier for that exact reason. However, you can't just glance at a spreadsheet and say "it's working!".

Statistical analyses need to be pre-defined and powered to measure an effect.

I've been on clinical trials where the innovator drug showed a fantastic effect early on, then two overall survival curves crossed a year later.

[Slaminerag]

It's possible to stop a trial because the new drug is much more effective than the old drug. It's also possible to stop a trial because the new drug's significantly worse than the old. We've done both. Also, I hope that no one ever tells you that you have to do a drug schema change because "too many kids are dying."