6 ms·

Hats off to brave Elizabeth Parrish who put her life at risk to push humanity forward. People were dismissing her as a lunatic at the time of treatment administ

by zavi 10y ago

Hats off to brave Elizabeth Parrish who put her life at risk to push humanity forward. People were dismissing her as a lunatic at the time of treatment administration. Now independently verified results speak for themselves.

patall 10y ago
If it is true what is written in this pr articel, she is an even bigger lunatic then previously thought. You do not treat someone with two experimental treatments at the same time, that is unscientific to the power of five. How would you determine which treatment was successfull and which was not (well, Phase I is about side effects, but you still need to distinguish)? Highly controversal.
- CuriouslyC 10y ago
  It doesn't matter how many hypotheses you test at one time, as long as they are all orthogonal. The problem arises when attempting to test multiple hypotheses that have some co-linearity in the solution.
  - SCHiM 10y ago
    But you don't know that going in do you? If you knew exactly what was going to happen you wouldn't need to experiment in the first place.
  - aidenn0 10y ago
    And it's not always possible to tell which hypotheses are orthogonal; if you test two hypotheses at once that you believe to be orthogonal, but it turns out that there is some underlying connection between them, you will get bogus results.
  - lunula 10y ago
    Here the co-linearity is that of them being applied to the same body. Now, I think it is still useful. At the first you always have N of one, and obviously there is no control. We at least know you don't die right away when your WBCs are upgraded.
  - agumonkey 10y ago
    Orthogonality and biology aren't friends.
    - kpil 10y ago
      In this case, I guess that they aim to make a whole cocktail of therapies to counteract age related effects, so in that light it kind of makes sense, at least as a PR trick although not as a scientific study. Anyway, it reminds me that someone (that I have forgot) said that the first human that will never die is already born.
      - agumonkey 10y ago
        Unless a drunkard drive over him on in a lonely area.
      - Beltiras 10y ago
        Aubrey de Grey has stated that the first human to reach 1000 years old is born today. He has also stated that most won't reach those livespans even with available treatments and methods due to the suboptimal behaviour we primates engage in.
        semi-extrinsic 10y ago
        Ray Kurzweil has stated (as recently as 2008) that in eleven years (in 2028), all of mankind's energy needs will be supplied by solar power. This will obviously not be true. Predictions are difficult, especially about the future. As for Aubrey de Grey, he can state whatever he wants to, but that doesn't make it true. It is, however, noteworthy that a scientific article published in the fairly prestigious EMBO Reports [1] concluded that "In his writings, de Grey fails to mention that none of these approaches has ever been shown to extend the lifespan of any organism, let alone humans." [1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1371037/#__ffn_sectitle https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1371037/#__ffn_...
        Beltiras 10y ago
  - patall 10y ago
    No. Pharmaceutical research does not and cannot work this way. Over the last few years, there has been much reasearch on molecular interactions. The current consensus is that everthing interacts with pretty much everthing at some point, it is just the degree of these interaction that varies. Molecular pathways become more and more vague (which does not mean that they disappear, its just very complex). For example, we do not even know how many existing drugs interact with each other, because who would fund that? Some reasearch hospital now do it but it is tedious when you have hundreds of potential interacting partners. For clinical trials, this leads to very strict guidelines. One of my university professors once told us about a drug they had to test in dogs for some side effects. Nothing happend to the dogs but afterwards they would have been eutanized because it was regulated that these exact dogs cannot enter another study. They finally adopted them, but still. Similar, people entering clinical trials are usually not allowed to take certain drugs, may need to abstinent from alcohol etc. Keep in mind. This is not an experiment, this is a clinical trial in a human.
    - Dylan16807 10y ago
      Giving a drug to one person to see what happens is definitely an experiment.
    - gumby 10y ago
      > Keep in mind. This is not an experiment, this is a clinical trial in a human. (Note: I have one drug in clinical trials, for which I designed the phase I and phase II protocols, and have appeared before the FDA in support of clinical trials of new medications.) Actually this is not a trial, it is considered an experiment. Trials are controlled (not always blinded or tasted against placebo, as many people think). Strictly speaking you cannot give an unlicensed medication without getting in trouble -- in fact permission to do a trial is actually agreement by the FDA that they won't prosecute you for breaking this law, basically because you followed one of the procedures spelled out in the law and the FDA agreed with your proposal. You are completely right on interactions. And yes, the animals are pretty much always sac'ed at the end of the experiment, even those in placebo arms :-( There are a few quasi-informal exceptions to the no unapproved administration. One is if you develop the drug yourself -- hard to prosecute. In a modern corporation however you can't really do that -- if you're a research scientist developing a drug, lawyers could claim that you were forced into it by your boss (another example is that the FDA requires that QC be an independent function). So in practice the only person who can try such a drug is the CEO. In my case I was both the CEO as well as one of the developers so was confident that the in vivo data we had showed that I would be fine. As a result I was a pincushion (these were true experiments in PK, diffusion etc -- I didn't have the indication we were treating for). The drug hasn't yet gone for approval, but if it does, somewhere in the approval application there will be a description under "previous human experience". You can do other experiments and trials someplace not subject to the FDA (i.e. other countries). Unless they are also formally part of your FDA process you can't use them in support of approval -- but if you paid for them you have to include them as previous human experience as well. And there are also various "compassionate use" cases though they are also somewhat-formalized now because some people try to stretch them into trial territory. By definition they are not adequately powered. There are amusing examples of previous human experience. My favorite is someone who shall go nameless who argued in support of an erectile disfunction treatment in front of his board by dropping his trousers and injecting his penis in front of them. I have only heard of this second hand but I bet there was quite a reaction! NB for anyone else in pharma: yes I glossed a few terms like calling the NDA an "approval application" in the interest of clarity for those not in the industry.
- jldugger 10y ago
  You have a sample size of one, and no control group. The patient dies. Are either of the drugs safe individually? Can't tell. Maybe one caused death and the other was fine. Maybe the two interacted in the human body somehow. Okay, now consider the scenario where the patient lives. Are either of the drugs safe individually? Can't tell. Again, maybe the two interacted in the human body in such a way that a potentially lethal treatment was rendered ineffective at curing or killing.
  - jjoonathan 10y ago
    Nope, Bayes makes the second situation (and to a lesser extent the first) informative since the probability of a life-saving second-order interaction between otherwise dangerous drugs is small compared to the probability of the toxicity having a more or less cumulative effect. Now, do I expect regulatory recognition of that fact in proportion to what you would get from a proper phase 0/1 trial? Probably not, and I'm OK with that because of the moral hazard involved, but it strikes me as more than a little ungrateful to pretend that the experiment was scientifically worthless. Someone put their life on the line for all of us, we should thank them and move on.
    - patall 10y ago
      Again, no. Just because the result turned out okay, the way how it was done does not become good (or scientific). There is no reason why it could not have been done in two persons. And btw: applying it to herself is a PR stunt. Just like politicians drinking water from Flint or Fukushima. Though iIt may even be that she had to do it because she would not get FDA permission. If not, still I would not thank her. She put a life (or health) unneccessary at risk, if it is hers or someoneelses does barely matter at that point.
      - darawk 10y ago
        No one is claiming the result is scientific or in any way equivalent to a proper study. But that doesn't make it worthless either. It demonstrates that it is at least possible to do this successfully in one human subject, which lends substantial credibility to the pursuit of the endeavor. An enormous amount of science has been built on scientists testing things on themselves. Here's another excellent example: http://discovermagazine.com/2010/mar/07-dr-drank-broth-gave-ulcer-solved-medical-mystery http://discovermagazine.com/2010/mar/07-dr-drank-broth-gave-...

by zavi 10y ago

patall 10y ago

If it is true what is written in this pr articel, she is an even bigger lunatic then previously thought. You do not treat someone with two experimental treatments at the same time, that is unscientific to the power of five. How would you determine which treatment was successfull and which was not (well, Phase I is about side effects, but you still need to distinguish)? Highly controversal.

CuriouslyC 10y ago
It doesn't matter how many hypotheses you test at one time, as long as they are all orthogonal. The problem arises when attempting to test multiple hypotheses that have some co-linearity in the solution.
- SCHiM 10y ago
  But you don't know that going in do you? If you knew exactly what was going to happen you wouldn't need to experiment in the first place.
- aidenn0 10y ago
  And it's not always possible to tell which hypotheses are orthogonal; if you test two hypotheses at once that you believe to be orthogonal, but it turns out that there is some underlying connection between them, you will get bogus results.
- lunula 10y ago
  Here the co-linearity is that of them being applied to the same body. Now, I think it is still useful. At the first you always have N of one, and obviously there is no control. We at least know you don't die right away when your WBCs are upgraded.
- agumonkey 10y ago
  Orthogonality and biology aren't friends.
  - kpil 10y ago
    In this case, I guess that they aim to make a whole cocktail of therapies to counteract age related effects, so in that light it kind of makes sense, at least as a PR trick although not as a scientific study. Anyway, it reminds me that someone (that I have forgot) said that the first human that will never die is already born.
    - agumonkey 10y ago
      Unless a drunkard drive over him on in a lonely area.
    - Beltiras 10y ago
      Aubrey de Grey has stated that the first human to reach 1000 years old is born today. He has also stated that most won't reach those livespans even with available treatments and methods due to the suboptimal behaviour we primates engage in.
      - semi-extrinsic 10y ago
        Ray Kurzweil has stated (as recently as 2008) that in eleven years (in 2028), all of mankind's energy needs will be supplied by solar power. This will obviously not be true. Predictions are difficult, especially about the future. As for Aubrey de Grey, he can state whatever he wants to, but that doesn't make it true. It is, however, noteworthy that a scientific article published in the fairly prestigious EMBO Reports [1] concluded that "In his writings, de Grey fails to mention that none of these approaches has ever been shown to extend the lifespan of any organism, let alone humans." [1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1371037/#__ffn_sectitle https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1371037/#__ffn_...
        Beltiras 10y ago
- patall 10y ago
  No. Pharmaceutical research does not and cannot work this way. Over the last few years, there has been much reasearch on molecular interactions. The current consensus is that everthing interacts with pretty much everthing at some point, it is just the degree of these interaction that varies. Molecular pathways become more and more vague (which does not mean that they disappear, its just very complex). For example, we do not even know how many existing drugs interact with each other, because who would fund that? Some reasearch hospital now do it but it is tedious when you have hundreds of potential interacting partners. For clinical trials, this leads to very strict guidelines. One of my university professors once told us about a drug they had to test in dogs for some side effects. Nothing happend to the dogs but afterwards they would have been eutanized because it was regulated that these exact dogs cannot enter another study. They finally adopted them, but still. Similar, people entering clinical trials are usually not allowed to take certain drugs, may need to abstinent from alcohol etc. Keep in mind. This is not an experiment, this is a clinical trial in a human.
  - Dylan16807 10y ago
    Giving a drug to one person to see what happens is definitely an experiment.
  - gumby 10y ago
    > Keep in mind. This is not an experiment, this is a clinical trial in a human. (Note: I have one drug in clinical trials, for which I designed the phase I and phase II protocols, and have appeared before the FDA in support of clinical trials of new medications.) Actually this is not a trial, it is considered an experiment. Trials are controlled (not always blinded or tasted against placebo, as many people think). Strictly speaking you cannot give an unlicensed medication without getting in trouble -- in fact permission to do a trial is actually agreement by the FDA that they won't prosecute you for breaking this law, basically because you followed one of the procedures spelled out in the law and the FDA agreed with your proposal. You are completely right on interactions. And yes, the animals are pretty much always sac'ed at the end of the experiment, even those in placebo arms :-( There are a few quasi-informal exceptions to the no unapproved administration. One is if you develop the drug yourself -- hard to prosecute. In a modern corporation however you can't really do that -- if you're a research scientist developing a drug, lawyers could claim that you were forced into it by your boss (another example is that the FDA requires that QC be an independent function). So in practice the only person who can try such a drug is the CEO. In my case I was both the CEO as well as one of the developers so was confident that the in vivo data we had showed that I would be fine. As a result I was a pincushion (these were true experiments in PK, diffusion etc -- I didn't have the indication we were treating for). The drug hasn't yet gone for approval, but if it does, somewhere in the approval application there will be a description under "previous human experience". You can do other experiments and trials someplace not subject to the FDA (i.e. other countries). Unless they are also formally part of your FDA process you can't use them in support of approval -- but if you paid for them you have to include them as previous human experience as well. And there are also various "compassionate use" cases though they are also somewhat-formalized now because some people try to stretch them into trial territory. By definition they are not adequately powered. There are amusing examples of previous human experience. My favorite is someone who shall go nameless who argued in support of an erectile disfunction treatment in front of his board by dropping his trousers and injecting his penis in front of them. I have only heard of this second hand but I bet there was quite a reaction! NB for anyone else in pharma: yes I glossed a few terms like calling the NDA an "approval application" in the interest of clarity for those not in the industry.
jldugger 10y ago
You have a sample size of one, and no control group. The patient dies. Are either of the drugs safe individually? Can't tell. Maybe one caused death and the other was fine. Maybe the two interacted in the human body somehow. Okay, now consider the scenario where the patient lives. Are either of the drugs safe individually? Can't tell. Again, maybe the two interacted in the human body in such a way that a potentially lethal treatment was rendered ineffective at curing or killing.
- jjoonathan 10y ago
  Nope, Bayes makes the second situation (and to a lesser extent the first) informative since the probability of a life-saving second-order interaction between otherwise dangerous drugs is small compared to the probability of the toxicity having a more or less cumulative effect. Now, do I expect regulatory recognition of that fact in proportion to what you would get from a proper phase 0/1 trial? Probably not, and I'm OK with that because of the moral hazard involved, but it strikes me as more than a little ungrateful to pretend that the experiment was scientifically worthless. Someone put their life on the line for all of us, we should thank them and move on.
  - patall 10y ago
    Again, no. Just because the result turned out okay, the way how it was done does not become good (or scientific). There is no reason why it could not have been done in two persons. And btw: applying it to herself is a PR stunt. Just like politicians drinking water from Flint or Fukushima. Though iIt may even be that she had to do it because she would not get FDA permission. If not, still I would not thank her. She put a life (or health) unneccessary at risk, if it is hers or someoneelses does barely matter at that point.
    - darawk 10y ago
      No one is claiming the result is scientific or in any way equivalent to a proper study. But that doesn't make it worthless either. It demonstrates that it is at least possible to do this successfully in one human subject, which lends substantial credibility to the pursuit of the endeavor. An enormous amount of science has been built on scientists testing things on themselves. Here's another excellent example: http://discovermagazine.com/2010/mar/07-dr-drank-broth-gave-ulcer-solved-medical-mystery http://discovermagazine.com/2010/mar/07-dr-drank-broth-gave-...