3 ms·

> The idea that "the body keeps it around" just completely flies in the face of reality. The fact is, the body doesn't really have a way of preventing it from s

by blix 5y ago

> The idea that "the body keeps it around" just completely flies in the face of reality. The fact is, the body doesn't really have a way of preventing it from staying there. This is a difference in semantics, not reality. The body also doesn't have a way of spontaneously ejecting the appendix. I would describe this as the body being happy to let it stay there. Of course, sometimes the body becomes unhappy with the existence of the appendix. I think the interesting question is why this happens and the proper treatment is to remove or mitigate these triggers. This is harder than removing the appendix. We thought there were no serious consequences of antibiotics for a century. We were wrong. As someone living with autoimmune disease, I can assure you that we continue to misunderstand these diseases. I wish I could be so confident that we would never be wrong about anything else ever again. A few days ago there was an article where a virus was used to fight a bacterial infection [1]. How can you earnestly say it's impossible that any virus that makes its home in the immune system could play a role in the immune response? Proving negatives is really hard. Maybe if I saw a 10000 studies that showed no correlation for a wide range of different infections I could be convinced, but our system doesn't encourage publication of negative results so I'm not holding my breath. How many have you seen? The problem with statistical medical studies is that in general samples are non-random (as you note in another comment) and cannot easily control for any of the 1M (lower bound) confounding variables. Our current methods here are mainly good for p-hacking publications, not uncovering causal relationships or complex effects. This is a discussion for somewhere else though probably. [1] https://english.elpais.com/usa/2022-01-27/how-a-virus-helped-a-woman-survive-a-dangerous-antibiotic-resistant-infection.html https://english.elpais.com/usa/2022-01-27/how-a-virus-helped...

sudosysgen 5y ago
>We thought there were no serious consequences of antibiotics for a century. That's not true. We knew there were serious consequences and that microbiome was important from before antibiotics were even a thing. >The body also doesn't have a way of spontaneously ejecting the appendix. I would describe this as the body being happy to let it stay there. It has a way of doing so over evolutionary timescales, which is not true for EBV. >A few days ago there was an article where a virus was used to fight a bacterial infection [1]. That is a bacteriophage. It is a virus that does not infect human cells, but only bacteria. We've known about them for decades. > How can you earnestly say it's impossible that any virus that makes its home in the immune system could play a role in the immune response? Proving negatives is really hard. I don't need to prove a negative, you have the burden of proof. It's absolutely certain that has an impact on immune response, we know that, it's just that every single effect we have seen so far has been negative. > The problem with statistical medical studies is that in general samples are non-random (as you note in another comment) and cannot easily control for any of the 1M (lower bound) confounding variables. There aren't 1 million confounding variables as far as sampling for EBV. You could feasibly account for all remaining variance, or enough of it to know that it doesn't have a sufficient impact.
- blix 5y ago
  > We knew there were serious consequences ... We had some minor inklings of the microbiome, but we were not aware of serious consequences. Unless you are accusing the all the governments and medical bodies of the world of gross malpractice. > It has a way of doing so over evolutionary timescales, which is not true for EBV. Do you have evidence for this claim? > That is a bacteriophage. Do you really think nature has drawn a bright line between viruses than can affect humans and viruses that can affect bacteria? I'm not sure this lines up with reality. People are even seriously starting to talk about the "Human Virome" [1] > I don't need to prove a negative, you have the burden of proof. This is what they said about tobacco causing cancer and pesticides killing bees and carbon emissions causing global warming. How did that work out? I'd prefer a more proactive scientific framework than a reactive one. Maybe we could stop the next crisis before it happens. > There aren't 1 million confounding variables as far as sampling for EBV. There are 1M confounding factors in sampling outcomes. Not controlling for confounding variables is a great way to get weak conclusions. [1] https://www.scientificamerican.com/article/viruses-can-help-us-as-well-as-harm-us/ https://www.scientificamerican.com/article/viruses-can-help-...
  - sudosysgen 5y ago
    > We had some minor inklings of the microbiome, but we were not aware of serious consequences. Unless you are accusing the all the governments and medical bodies of the world of gross malpractice. That is not true. Microbiome treatments were proposed at the beginning of the 20th century. I don't see how it would be malpractice. Pathogenic bacteria are a much worse problem than damage to microbiome in most cases. If your microbiome was damaged by evidently useless antibacterial treatments, it could be malpractice. > Do you really think nature has drawn a bright line between viruses than can affect humans and viruses that can affect bacteria? I'm not sure this lines up with reality. People are even seriously starting to talk about the "Human Virome" Yes, there is bright biological line between viruses that can infect humans and bacteria. And sure, the human virome could be thing, and certainly natural bacteriophages are important and so are many benign, transient infections important for the human system. > This is what they said about tobacco causing cancer and pesticides killing bees and carbon emissions causing global warming. How did that work out? I'd prefer a more proactive scientific framework than a reactive one. Certainly not. Tobacco causing cancer was the default because it is a foreign contaminant in the human body. Insecticides killing bees (insects) is also the default position. Global warming was theorized and predicted since the late 19th century, see Arrhenius, so yes energy companies had the burden of proof. We already have the priors of EBV causing massive damage to the human body. You're saying that we must prove the negative that it causes nothing beneficial at all. That's akin to saying that despite we know that global warming has many negative effects, we have to show it has no positive effects. > There are 1M confounding factors in sampling outcomes. Not controlling for confounding variables is a great way to get weak conclusions. That isn't how statistics work. You can calculate a distribution of impact of every confounding factor onto the measured variance, and from then on you can know if you have controlled enough confounders or not. If you have 999,990 factors that alltogether account for 0.1% of variance, and 10 members that account for 99.9% of variance, you know it's enough.

by blix 5y ago

sudosysgen 5y ago

>We thought there were no serious consequences of antibiotics for a century. That's not true. We knew there were serious consequences and that microbiome was important from before antibiotics were even a thing. >The body also doesn't have a way of spontaneously ejecting the appendix. I would describe this as the body being happy to let it stay there. It has a way of doing so over evolutionary timescales, which is not true for EBV. >A few days ago there was an article where a virus was used to fight a bacterial infection [1]. That is a bacteriophage. It is a virus that does not infect human cells, but only bacteria. We've known about them for decades. > How can you earnestly say it's impossible that any virus that makes its home in the immune system could play a role in the immune response? Proving negatives is really hard. I don't need to prove a negative, you have the burden of proof. It's absolutely certain that has an impact on immune response, we know that, it's just that every single effect we have seen so far has been negative. > The problem with statistical medical studies is that in general samples are non-random (as you note in another comment) and cannot easily control for any of the 1M (lower bound) confounding variables. There aren't 1 million confounding variables as far as sampling for EBV. You could feasibly account for all remaining variance, or enough of it to know that it doesn't have a sufficient impact.

blix 5y ago
> We knew there were serious consequences ... We had some minor inklings of the microbiome, but we were not aware of serious consequences. Unless you are accusing the all the governments and medical bodies of the world of gross malpractice. > It has a way of doing so over evolutionary timescales, which is not true for EBV. Do you have evidence for this claim? > That is a bacteriophage. Do you really think nature has drawn a bright line between viruses than can affect humans and viruses that can affect bacteria? I'm not sure this lines up with reality. People are even seriously starting to talk about the "Human Virome" [1] > I don't need to prove a negative, you have the burden of proof. This is what they said about tobacco causing cancer and pesticides killing bees and carbon emissions causing global warming. How did that work out? I'd prefer a more proactive scientific framework than a reactive one. Maybe we could stop the next crisis before it happens. > There aren't 1 million confounding variables as far as sampling for EBV. There are 1M confounding factors in sampling outcomes. Not controlling for confounding variables is a great way to get weak conclusions. [1] https://www.scientificamerican.com/article/viruses-can-help-us-as-well-as-harm-us/ https://www.scientificamerican.com/article/viruses-can-help-...
- sudosysgen 5y ago
  > We had some minor inklings of the microbiome, but we were not aware of serious consequences. Unless you are accusing the all the governments and medical bodies of the world of gross malpractice. That is not true. Microbiome treatments were proposed at the beginning of the 20th century. I don't see how it would be malpractice. Pathogenic bacteria are a much worse problem than damage to microbiome in most cases. If your microbiome was damaged by evidently useless antibacterial treatments, it could be malpractice. > Do you really think nature has drawn a bright line between viruses than can affect humans and viruses that can affect bacteria? I'm not sure this lines up with reality. People are even seriously starting to talk about the "Human Virome" Yes, there is bright biological line between viruses that can infect humans and bacteria. And sure, the human virome could be thing, and certainly natural bacteriophages are important and so are many benign, transient infections important for the human system. > This is what they said about tobacco causing cancer and pesticides killing bees and carbon emissions causing global warming. How did that work out? I'd prefer a more proactive scientific framework than a reactive one. Certainly not. Tobacco causing cancer was the default because it is a foreign contaminant in the human body. Insecticides killing bees (insects) is also the default position. Global warming was theorized and predicted since the late 19th century, see Arrhenius, so yes energy companies had the burden of proof. We already have the priors of EBV causing massive damage to the human body. You're saying that we must prove the negative that it causes nothing beneficial at all. That's akin to saying that despite we know that global warming has many negative effects, we have to show it has no positive effects. > There are 1M confounding factors in sampling outcomes. Not controlling for confounding variables is a great way to get weak conclusions. That isn't how statistics work. You can calculate a distribution of impact of every confounding factor onto the measured variance, and from then on you can know if you have controlled enough confounders or not. If you have 999,990 factors that alltogether account for 0.1% of variance, and 10 members that account for 99.9% of variance, you know it's enough.