Part of the infodump!
Firstly, placebo is known to be quite effective in relieving pain. This means that, in some cases, the patient's perception of suffering is reduced. Because a LOT of medical demand is for symptom relief, then placebo can be quite effective in benefiting a patient.
One thing that's probably intuitive is that you can get different intensities of placebo. You can even test for what components of a placebo treatment are most effective. For example, in Kaptchuk et al (2008), patients came to receive sham acupuncture along with some TLC for irritable bowel syndrome. Now, IBS changes in intensity over time, so some patients spontaneously recover. This group endeavored to see what was more effective in providing relief: acupuncture or the acupuncture & TLC
As you can see, sham acupuncture ("limited") increased the odds of the patient perceiving relief. Sham acupuncture & TLC ("augmented") was even more effective.
This is not really a surprise. What blew me away was a recent paper is that there were different ways to inhibit placebo effect, based on how the placebo effect was being generated. It's been known for ~30 years that 'naloxone' (an opiod inhibitor) inhibits the placebo effect from morphine. But 2011 had a cool paper too.
In this experiment, they would get the subjects to resist expermental pain for as long as they could. On day 2 & 3 they gave the subjects morphine. On day 4, they
thought they were getting morphine but either got naloxone or saline. You'll see that the expectation of morphine was enough to blunt pain, but naloxone blocks the ability to derive benefit from saline after morphine
In
their more recent paper (behind a paywall, sorry) that group showed that you can generate a similar effect with non-opoid painkillers (e.g., tylenol), but naloxone doesn't inhibit this benefit. But inhibiting something else (using a different drug) will block the non-opiod placebo effect.
So, there's actual neuroanatomical correlates for placebo, and (in fact) there are multiple types!
Finally (for this post, there's actually a ton of really cool stuff), a great portion of modern analgesics get their benefit
from the placebo effect. If the patient is aware they're getting a painkiller, the painkiller is more effective at relieving suffering than if they don't know they're getting it. This
paper shows that if the patient realises that painkiller is being placed into the IV, they receive quite a bit more relief than if the IV is dosed with painkiller by an automated system that doesn't inform the patient.
The 'actual' analgesic effect is
augmented by the placebo effect.
Finally, we have some clues as to where in our neuroanatomy we generate our placebo effect. Like I said, different placebos can be blocked by inhibiting certain types of receptors. But, as well, if we 'inhibit' the dorsolateral prefrontal cortex using transcranial magnetic stimulation (another totally awesome technology) you can
prevent someone from getting the placebo benefit without increasing their pain sensitivity. They don't become sissier, they just don't get placebo benefit.