[trilbyglens]

I think only very simple mechanisms can be fully described and reproduced by modern methods. Is it not too hard to image that our grasp on reality is limited, and that the tools and language we possess right nowight simply not be up to the task? Why be so fast to dismiss things that are not currently reproducible? I bet if you took 100 people you'd have a hard time reproducing an Olympic athlete even if all 100 were subjected to the same fitness regimen. But we know Olympic athletes exist. Maybe you should also be skeptical of your skepticism.

[vacuity]

Also, we know placebos exist and in fact are expected to exist in medical trials, but...how? Why? Can we control this? Evidently we're seeing something, and we've incorporated it into scientific understanding, but only at a high level so far.

[tsimionescu]

The placebo effect is much less mysterious than it's made out to be.

It is almost entirely explained by three things:

1. Some ability of the mind to interact with the effect of the medication - such as heart rate being tied to mood (so that being administered a heart-rate reducing drug in a menacing way could have the effects of the drug masked by your heart rate increasing because of your anxiety), or when the effect itself is tied to mental phenomena (depression, pain, nausea, etc). Apart from psychoactive drugs, this only significantly impacts a relatively small amount of drugs, since only a small amount of biomarkers are very directly tied to mental states - heart rate, blood pressure, maybe a handful of others. There may also be a smaller effect on longer term treatments of other kinds due to the myriad effects of stress/anxiety on various other systems in the body, that may have interactions with more complex biomarkers as well, especially in the longer term.

2. The experimental setup being impacted more or less knowingly by people based on their pre-conceived notions. This can take many forms, from outright doctoring the recorded data to fit the desired outcome, to much more subtle effects like differences in the level of care, differences in how well the patients follow the prescribed protocol, and many others. None of these are "real" effects, they only affect the relevance and quality of the data being collected. For example, if patients have a way of finding out they are in the placebo wing, they may drop out at a higher rate, and thus the data may show that the patients who stuck out had a higher benefit than the ones on the control wing simply because of that skew.

3. Noise in the disease itself - sometimes, people spontaneously heal or improve from various conditions on various fronts without any intervention whatsoever, and any medical intervention has to be compared to this baseline level. This is similar to the problem of checking if a code fix actually addresses a hard to reproduce bug.

There is no proof whatsoever of the placebo effect being some ability to heal the body through belief (again, beyond certain mental phenomena, either related to mental illness or pain or nausea relief). It is simply a matter of noise and measurement artifacts for the vast majority of studies.

[vacuity]

I think your points are mainly issues with determining that an experimental treatment actually has a significant effect. They aren't as relevant to explaining how the control group experiences an impact. I've heard that patients and doctors often correctly guess if a placebo was administered, so if a placebo actually does have some mind-driven impact then I would expect it to disappear when known. Then your point 2 reduces to point 3. But noise in how the disease naturally progresses is also split amongst control and experimental groups, and so the important test is the control group compared to people with no treatment at all, or even unwitting of their observation. Although you are likely right that the placebo effect is fairly simple: something along the lines of lowered stress and anxiety from a potential treatment, even if it could be fake.

[tsimionescu]

> I think your points are mainly issues with determining that an experimental treatment actually has a significant effect.

Yes, this is the place where the placebo (and nocebo) effect was observed and where it has an impact.

> But noise in how the disease naturally progresses is also split amongst control and experimental groups, and so the important test is the control group compared to people with no treatment at all, or even unwitting of their observation.

It's basically never possible to do a study comparing a new medication to no treatment, mostly on ethical grounds (beyond the complicated logistics of studying people outside of clinical settings). There are very few diseases for which we have no treatment whatsoever, so it's very rarely ethically acceptable to compare between offering a medication vs no treatment.