I am currently reading Daniel Moerman’s “Meaning, medicine and the ‘placebo effect’”. As well as containing many interesting asides, the book discusses what is at the heart of the so-called placebo effect: patients’ response to the meaning of their treatment. Moerman calls this the ‘meaning response’. This response to meaning explains why two inert pills produce more cures than one inert pill, and why inert injections are even more effective (because “everybody knows” that injections are more powerful than pills). But importantly, it is possible to show that doctors are as important in producing the meaning response as patients. Gracely et al (1985) looked at the effect of placebo on pain in patients having their wisdom teeth extracted. The study was set up as a standard double-blind (neither the doctor nor the patient knows if the patient is getting a real medicine or an inert placebo), with the possibilities being a placebo, fentanyl (which usually reduces pain) and naloxone (which usually blocks reduction in pain, so could be expected to increase the pain of the procedure). The twist was that for the first half of the experiment the doctors, but not the patients, were told that a supply problem meant that no patient would be getting the pain-relieving fentanyl. In the second half the doctors were told that the problem had been resolved, so that now the patients might receive fentanyl. By comparing levels of patient pain in the placebo condition is possible to gauge the effect of doctor expectations on the meaning response of the patients. In this condition patients are all receiving inert substances, and they all ‘know’ the same thing: they might receive a placebo, pain-relief or ‘pain-enhancement’. The doctors don’t tell them about the supply problem and, for that matter, they don’t know themselves for definite what the patient is given. The only difference is that for the patients in the first half, the doctors think they know that pain-relief is not a possibility, whereas in the second half it is. The graph of the results, copied from Moerman’s book is below:
As you can see, patients in the PNF group — those whose doctors thought they might receive pain-relief had a large pain-relieving placebo effect compared to those in the PN group — those whose doctors thought they couldn’t receive pain-relief (update in the original edit of this post I had these labels the other way around, incorrectly)
What I think is interesting about this study is, firstly, it confirms the need for rigorous double-blind controls in studies of medicine and, secondly, just how significant an effect this subtle manipulation has. The doctors don’t know anything definite, and they certainly aren’t telling the patients what they suspect or guess, but somehow — a look? a slightly brighter smile? a slightly lowered tone? — they communicate their knowledge of the probabilities to the patients who then experience a real change in their levels of pain because of it.
A striking aspect of the meaning response is that one could suppose that patients have control over their experience of different levels of pain. After all, we know that the pills are inert. Could we just imagine ourselves a ‘placebo effect’ in all situations where we have unnecessary pain? Sadly, normally we can’t do this — the meaning response doesn’t work like that. Doctors are required to give patients permission to feel less pain. Perhaps a fundamental part of the creation of meaning is that it requires other people.
Update: A great recent post by Vaughan ‘placebo is not what you think’, which deserves to be linked up with this post
Gracely, R. H., Dubner, R., Deeter, W. R., & Wolskee, P. J. (1985). Clinicians’ expectations influence placebo analgesia. Lancet, 1(8419), 43.
Moerman, D. E. (2002). Meaning, medicine, and the “placebo effect”. Cambridge University Press: New York.