The Function of Placebos
- Admin
- Nov 26, 2022
- 7 min read
I thought I’d give a quick and simplified primer on what placebos actually are, and what their function, purpose, and use is in science. If I’m to explain how ‘the placebo effect’ is not a viable form of treatment and is actually pop culture and media hype, then it’s important to understand what a placebo is first of all.
Arguably a difficult topic and a great lesson in critical thinking, to understand this holds countless more value when assessing claims about health interventions and the placebo effect in day to day life.
There are plenty of legitimate sounding ‘scientists’ pushing the hype to sell something, of course. These are still not representative of the scientific consensus, but a vocal minority. If you hear ‘power of the mind’, you should be skeptical immediately!
When/why do we need placebos?
Imagine that 15 years of lab work before trials/drug development is done. Possible effects of a drug have been observed at early stages and at least a level of plausibility has been established, to warrant development of a drug which is now ready to move to human trials.
You have a group of people participating in the study of the effects of a new drug. Some will be given the drug, and some will take nothing, to see if the drug makes a difference to their condition.
The patients will report back on their symptoms and experiences, and at the same time a physical examination will be done to see if there is any difference in their condition now that the drug was trialled.
For strictly physical outcomes, it is easier to test. They take the same measurements as before the patients got the drug and see if there are any significant differences in the results after the drug. Next compare the groups. If there is a significant difference between the groups, there’s some plausibility to the drug working or having an effect. From there, careful and honest interpretation of the results, independent replication, and the scrutiny of peer review (constructive criticism) is all to come.
But for subjective, self-reported things that we can’t measure, like pain, irritation, drowsiness, aches, etc, how do we know the people taking the drug:
Weren’t in a peak of pain when they went to be part of the trial, and the symptoms were going to wane in the next week anyway – making the drug look effective, even though it may not be?
Naturally got better by coincidence?
Had the expectancy of being given a drug to get better, affecting their reporting of the severity of symptoms afterwards?
That their perception (and therefore reporting) of symptoms changed because they felt more at ease, or relaxed, at being given the hope that a drug may help them?
Weren’t experiencing confirmation bias – something else changed during the trial (maybe diet, sleep, other medication, naturally got better) and was not reported to the researchers, but the patient attributes their getting better to the trial drug because that’s what they remember taking?
Weren’t misremembering the experience as worse, better, or unchanging, but influenced by the pressure to say something meaningful for the study?
The short answer is, we don’t know, with the way this study is set up. There are a lot of other factors not listed above which could also give inaccurate results. The solution to the above problems is that we introduce a third group – the placebo group.
Now the study is significantly stronger, as it controls for the above list of confounding factors.
What is ‘The Placebo Effect’ and what is a Placebo?
Combined, the above non-exhaustive list of confounding factors is what amounts to the placebo ‘effect’. Things that would appear promising to the new drug working, but which don’t have anything to do with the treatment and would vary to a degree even if nothing was done at all.
In short, the placebo effect is simply a list of biases which can skew the data and deceive the trial participants and the researchers by distorting and affecting self-reported outcomes. It makes study results less reliable and more difficult to interpret.
Thus, a placebo itself is a decoy or dummy-intervention designed to lure out and highlight the human biases/errors associated with testing treatments – it controls for biases.
How to interpret the placebo effect – what does it mean?
Now we have the drug group, the do nothing group, and the placebo control group.
The nothing group produces results which show variations in the condition over time, with no intervention.
The placebo group shows the extent that all above biases (and more) might affect the people reporting the outcome – and therefore to what extent we subtract this level of ‘effect’ from the real drug outcomes.
Finally, by comparing the do nothing group and the placebo group to the drug group, you are able to see to see how much the results vary if you do nothing, plus to what extent the placebo biases affect the results compared to nothing, and then subtract all of this noise (or perceived ‘effect’) from the results of the drug group.
If the drug group still shows a significant improvement beyond the amount of variation in the placebo group, it is a promising and likely real effect. If the drug performance is much the same as the placebo group, then the drug (like the placebo) does not work.
Further study accuracy is achieved by:
How not to interpret the placebo effect – Ted Kaptchuk
The way some researchers like Ted Kaptchuk wrongly spin it, is that the placebo is performing ‘as well as’ a good drug. As demonstrated above, this is incorrect. This is how poorly-conducted science can misrepresent trial results for ineffective treatments.
If you look at studies which the authors make such claims about the placebo effect being of any use itself other than to test real treatments against, the presumed ‘effect’ shrinks and disappears the higher quality the study is.
The logic of some researchers like Ted Kaptchuk is flawed. He suggests that because people can ‘benefit from’ (aka misinterpret) the placebo ‘effect’ even if they know it’s a placebo, we should study that as a form of treatment.
He’s been quoted before as effectively saying that to ‘feel better’ is of more importance than to receive a real and effective treatment. This is of course an absurd position for a medical researcher (in which he is not trained) to hold – because they are meant to cure people, not trick them and ignore the issue.
But all this does is invoke a weaker (than medicine), temporary set of biases - knowingly fooling oneself in plain sight. It frequently leads into the pseudoscientific notion that the power of the mind is such that it can heal the body as well, and it's unsurprising that Ted flirts with this hyped nonsense, given his background and history of pseudoscience.
At that point it becomes more dangerous, and leads to other flawed habits of thinking – where usually someone else profits while it becomes the victim’s fault if they are not thinking positively enough to ‘heal’ themselves.
He characterises the biases that studies should control for to be more accurate, as being the solution. This is flawed, circular reasoning. You can’t have an accurate study of something which is itself the inaccuracy of a study. You can’t control for ‘biases which affect themselves’. This logic leads you in circles without ever addressing the root/cause of an issue.
The resulting disconnect is that ‘The Placebo Effect’ when used by media or lay people then takes on the meaning of a mysterious, little-known, effective intervention, or the mind healing itself, when in reality it is merely a list of variables that can lead to misinterpreted results. Irony at its finest!
Conclusions
Much of the popular hype is “but the mind blowing thing is that the placebo ‘works’ (as discussed, it doesn’t) even when the patient knows it’s a placebo!”. All this means is that while controlling for the biases associated with believing you are getting a real drug, there are other biases which exist whether you know it is a placebo or not.
These still include:
Regression to the mean (getting better without doing anything, pain subsiding for a period)
Confirmation bias (you attribute improvement of symptoms to the placebo because you wanted it to be true – when really it would’ve happened without placebo or for another reason)
Expectancy bias (because if you’ve been told the placebo effect still works even if you know it, then you still have expectancy of an effect)
and still many more.
In conclusion, this is all the more reason to use well controlled, double-blinded studies. Placebos might temporarily be able to modulate subjective symptoms such as pain, but they do not cure, heal, or fix anything.
If there is a discernible underlying reason that is causing the pain, placebos may change a person’s perception of that pain, but they don’t affect the cause; therefore, real medicine would be the only fix of the cause; and especially where the cause is known, real medicine is the only ethical treatment.
Pain management is something you could much more effectively and reliably take a pain killer for because that is physical medicine with actual mechanisms – but despite the hype, the evidence that placebos perform anywhere near as well as drugs is lacking and weak.
Beyond this, pain is usually there to indicate that something is wrong. It perhaps wouldn’t be wise to ignore it by encouraging placebos to mask it (and discourage medical intervention), rather than treat it at the cause – again, placebos do not ‘trick’ the mind into ‘healing’.
Administering a placebo is not a viable treatment or medicine. The placebo effect does not unlock the ‘healing power of the mind’. Finally, while there are some unanswered questions about exactly how these biases can alter the perception of symptoms, they do not work as often or reliably as medicine, even for pain relief. It is not as powerful a mechanism as is hyped.
Placebos are a brilliant tool for improving the validity of scientific studies, when used correctly. This is where they belong.
Comments