If placebo controls work even when the patient is aware that the treatment is a placebo. Why are deceptive placebo controls used?

Question

Placebos are “fake” treatments prescribed for the psychological benefit to the patient rather than for any physiological effect. Placebos are used for conditions defined by “self-observation” symptoms (E.g. pain, nausea, or fatigue). But a new area of research has extended toward 'open-label placebos' or "educated placebos".

While it was once assumed that deception was necessary for placebos to have any effect, there is now evidence that placebos can have effects even when the patient is aware that the treatment is a placebo.

Source (2nd paragraph)

Even though placebos contain no real treatment, researchers have found they can have a variety of both physical and psychological effects

If placebo controls can have effects even when the patient is aware that the treatment is a placebo, then why not just always tell clinical trial participants they are using a placebo?

Benefits

• Cost-effective
• Easier to recruit participants (Thus, potentially larger sample sizes)
• Less stress on participants (Wondering if they are taking a placebo or not)
• Speeding up the clinical research time-frame

Source

Using 'open-label placebos' you can still control for researcher bias (Single-blind study - only the researchers blinded, it cannot be double-blinded because the participants know.)

It is reducing the quality of the research? Because there is no control of the physical and psychological effects? But that isn't right because again:

Placebos can have effects even when the patient is aware that the treatment is a placebo.

Answer 1

The entire point of having a placebo control from a statistics perspective is that the only thing different between the groups is the treatment. You cannot say this is true if you also tell people which group they are in: if you do this, there are now two differences between groups, 1) Treatment vs placebo, 2) Knowing they are in the placebo group vs. knowing they are in the treatment group.

Your logic does not make sense: It doesn't matter if an "open placebo" also has some treatment effect in some situation. You would instead need to verify that there is no difference of effect between knowingly getting a placebo and knowingly getting an ultimately ineffective treatment.

Furthermore, you would need to show this for every single condition you might want to investigate, not just one, and not just no statistically significant difference, you would actually need exactly the same effect, which you cannot ever demonstrate (and which will reasonably never be true, anyways). Otherwise, you will have unknown bias in your study that could cause you to falsely think a treatment is better than it actually is.

There are other good reasons to have a placebo control, besides statistics: doctors and other study staff might be hoping to find a positive result, so they may (intentionally or unintentionally) bias their approach to the placebo and treatment groups. Ideal placebo controls blind the researchers and staff, too, not just the patient (of course with some things like a surgical intervention this is not entirely possible, but you still try to do it to the extent possible).

Patients may also report things differently if they know what group they are in. If they are concerned about side effects, maybe they report more side effects on the treatment than they would have on the placebo. If they are hopeful for the treatment, they might mask side effects if they know they are in the treatment group.

Placebo controls aren't possible in all circumstances, and there is lots of clinical research that does not use a placebo, especially exploratory research. There are also some studies where it's not really possible to completely hide that someone is in a certain group. But the randomized controlled trial is the gold standard for important statistical and methodological reasons, and a double-blind placebo control is the gold standard comparison when comparing to no treatment.