3 ms·

I had to read the paper to understand that this means each of the 24 drugs was approved by a single study- that is, one drug, one study. Not 24 drugs, one stud

by dekhn 3y ago

I had to read the paper to understand that this means each of the 24 drugs was approved by a single study- that is, one drug, one study. Not 24 drugs, one study. Running clinical trials is absurdly expensive and challenging. It's often not possible to run multiple trials.

duskwuff 3y ago
Precisely. It's particularly difficult to run multiple trials of drugs for rare diseases -- there's a limited pool of patients available for these studies, and splitting them up between multiple studies makes it much more difficult for each of those studies to reach a statistically significant conclusion.
jfengel 3y ago
Would adding more studies even do anything more than one proper study? More data is always better, to be sure, and it would be great to run multiple designs. But one "gold standard" study should suffice, and demanding more just feels like moving the goalposts.
- dekhn 3y ago
  The largest studies are composed of multiple sites of participants. This has multiple advantages; for example, it makes it easier to recruit a diverse population, reducing trial failures due to population-specific effects. Also, it helps randomize some aspects of treatment, so you don't depend too much on one site's special magic. If you can't recruit a large enough N of patients in one site, you might be able to recruit your N across multiple sites. There are downsides- overhead (each site has its own IRB that must approve the study) and additional complexity (the statistical modelling needs to be carefully managed). It's not uncommon in Cancer trials where there are multiple regional centers that specialize in cancer treatments, who see all the patients with specific types of cancer in a geographical region, and the prevalence of the cancer is high in the population. Also, if you want to change the usage for a drug, IE you were approved for lung cancer but now you want to use it for liver cancer- you might need to run another study. Not multi-site, but just a completely new study with the same medication, but new patients, new protocol (dosage and timing), new site location.
_Microft 3y ago
The title made me think that there was one study that included 24 drugs which I found concerning. In the end I have flagged the submission now because of its editorialized and misleading title.

by dekhn 3y ago

duskwuff 3y ago

Precisely. It's particularly difficult to run multiple trials of drugs for rare diseases -- there's a limited pool of patients available for these studies, and splitting them up between multiple studies makes it much more difficult for each of those studies to reach a statistically significant conclusion.

jfengel 3y ago

Would adding more studies even do anything more than one proper study? More data is always better, to be sure, and it would be great to run multiple designs. But one "gold standard" study should suffice, and demanding more just feels like moving the goalposts.

dekhn 3y ago
The largest studies are composed of multiple sites of participants. This has multiple advantages; for example, it makes it easier to recruit a diverse population, reducing trial failures due to population-specific effects. Also, it helps randomize some aspects of treatment, so you don't depend too much on one site's special magic. If you can't recruit a large enough N of patients in one site, you might be able to recruit your N across multiple sites. There are downsides- overhead (each site has its own IRB that must approve the study) and additional complexity (the statistical modelling needs to be carefully managed). It's not uncommon in Cancer trials where there are multiple regional centers that specialize in cancer treatments, who see all the patients with specific types of cancer in a geographical region, and the prevalence of the cancer is high in the population. Also, if you want to change the usage for a drug, IE you were approved for lung cancer but now you want to use it for liver cancer- you might need to run another study. Not multi-site, but just a completely new study with the same medication, but new patients, new protocol (dosage and timing), new site location.

_Microft 3y ago

The title made me think that there was one study that included 24 drugs which I found concerning. In the end I have flagged the submission now because of its editorialized and misleading title.