4 ms·

Due to the way medication is developed, we basically only end up with medicine that works on humans + at least one other species. I wonder how many medications

by Aaargh20318 3y ago

Due to the way medication is developed, we basically only end up with medicine that works on humans + at least one other species. I wonder how many medications we have missed out on that would have worked on humans but never got to the human trial stage because it doesn’t work in any of the animal models we use.

haldujai 3y ago
This isn’t a medication it’s a mechanism study. I would expect not many and certainly far less than the number of medications that worked in animals but did not work or harmed humans.
PaulKeeble 3y ago
And a condition we can induce in another animal. If we can't cause it in another animal then we can't test a drug against it. There are a lot of diseases that can make no progress under this model which we can not induce in animals or humans.
- patapong 3y ago
  Fascinating! Reminds me how machine learning as a field is often focused on tasks that can be quantitatively evaluated using large datasets. There are a lot of tasks that do not fit this model, especially interactive tasks.
ekianjo 3y ago
Aspirin is famous for having teratogenic effects on rabbits and would never have been approved for humans if it were to be tested nowadays
- haldujai 3y ago
  Why? Plenty of teratogenic drugs are approved and in routine use in non-pregnant patients.
  - ekianjo 3y ago
    Any recent example?
    - haldujai 3y ago
      Not really sure what you're getting at, teratogenicity only matters if a person is pregnant and lack of teratogenicity is in no way a requirement for FDA approval. There are many commonly used drugs with teratogenic risks that we discontinue in pregnancy or counsel the patient on, picking randomly from a list of medications that were approved in 2023: sparsentan I'm not sure if I'm missing something but why are you thinking teratogenicity matters when the vast majority of patients are not pregnant at a given moment?

haldujai 3y ago

This isn’t a medication it’s a mechanism study. I would expect not many and certainly far less than the number of medications that worked in animals but did not work or harmed humans.

PaulKeeble 3y ago

And a condition we can induce in another animal. If we can't cause it in another animal then we can't test a drug against it. There are a lot of diseases that can make no progress under this model which we can not induce in animals or humans.

patapong 3y ago
Fascinating! Reminds me how machine learning as a field is often focused on tasks that can be quantitatively evaluated using large datasets. There are a lot of tasks that do not fit this model, especially interactive tasks.

ekianjo 3y ago

Aspirin is famous for having teratogenic effects on rabbits and would never have been approved for humans if it were to be tested nowadays

haldujai 3y ago
Why? Plenty of teratogenic drugs are approved and in routine use in non-pregnant patients.
- ekianjo 3y ago
  Any recent example?
  - haldujai 3y ago
    Not really sure what you're getting at, teratogenicity only matters if a person is pregnant and lack of teratogenicity is in no way a requirement for FDA approval. There are many commonly used drugs with teratogenic risks that we discontinue in pregnancy or counsel the patient on, picking randomly from a list of medications that were approved in 2023: sparsentan I'm not sure if I'm missing something but why are you thinking teratogenicity matters when the vast majority of patients are not pregnant at a given moment?