Talk:Placebo effect

House of Lords - Science and technology - Sixth Report

I think this is a really thorough and neutral explanation of the placebo effect from the source the Gareth introduced us to. D. Matt Innis 05:33, 10 January 2009 (UTC)

Probably a little too late to absorb. Will look at it tomorrow; hopefully the forecasted blizzard doesn't get connectivity or power.

Reason for "might"

Perhaps I should copy the http://jme.bmj.com/cgi/content/full/30/6/551 reference from placebo. When I said "might", perhaps not in the most flowing way, I was thinking of their third case study, which I can make even more ethically complex. A patient presents with clinical depression, for which the clinician prescribes an appropriate antidepressant. The antidepressant is known to a 2-4 week delay before it takes effect.

The patient, however, reports immediate relief of symptoms. Now, unless there's a pharmacologic miracle, this fairly well has to be a placebo effect from a real drug. What are the ethical obligations to continue?

Take it a step further. Let's say the patient, a week later, starts complaining of known side effects from the drug. Now, what is the best ethical course? Keep the patient on a non-benign drug if it has shown benefit for the major complaint and the side effects are not intolerable? Change to a true inert medication, knowing the patient is suggestible?

Incidentally, do look at Talk:Sham treatment/Related Articles‎. Larry has some questions about the relationship among placebo, placebo effect, and sham treatment. All are related, but there's no strict hierarchy among them. Howard C. Berkowitz 05:58, 10 January 2009 (UTC)

I agree with all of that, but are you saying that, other than unconscious patients, there are times when the placebo effect does not occur? I know we aren't totally correct without the 'might', but how can we bring it out that, some believe that the placebo effect is part of all treatments based on the assumption that all human contact has a psychological impact. From the House of Lords above:

"The placebo effect has been described as the therapeutic impact of 'non-specific' or 'incidental' treatment ingredients, as opposed to the therapeutic impact that can be directly attributed to the specific, characteristic action of the treatment."

""...we all recognise the strong placebo effect in, probably, all aspects of medical treatment, whether they are conventional or not" (Q 155). "

D. Matt Innis 07:36, 10 January 2009 (UTC)

No, I don't assume placebo effect will occur in all cases. My experience and observations, and I think the literature in both pain management and psychopharmacology supports it, is that suggestion may have an effect, but frequently does not. It's very common to have to try several drugs before an effective one is found.

While the singular of data is not anecdote, I was having surgery for a pilonoidal cyst, back when night-before admission was common. The anesthesiologist visited me the night before, and I told him that while I didn't need or want preoperative sedation, not to use the then-current barbiturates as I was extremely insensitive. He laughed and said he'd just use more. The morning of surgery, the nurse came in with 300mg of pentobarbital, three times the normal dose. I was a little safety-concerned, but I took it. When I was wheeled into the OR, I was awake and alert, rather to the shock of the staff.

If the placebo effect always occurs, I should have been asleep. Since that was an enormous dose of an potent drug, if anything, there was a reverse placebo effect, but I don't remember fighting to stay awake. Howard C. Berkowitz 15:27, 10 January 2009 (UTC)

Ah, but you are assuming that it is the clinician (and I agree that is what we have written) that is the cause of the placebo effect, but the effects are not always the result of the clinician. It is whatever makes us believe that the intervention will work - the more elaborate the intervention, the more the placebo effect plays a role. You are saying that you 'believed' that you were 'extremely insensitive'. That could be an isolated anecdotal example of a placebo effect (though it could be something else as well). It's just that you suggested it to yourself. When a patient believes that a pill will cure their arthritis, it's effects are improved by the placebo effect. When they do not think it will cure their arthritis, its effects are diminished by the placebo effect. In other words, the placebo effect is always a part of any treatment, but that does not mean that the entire effect is the result of the placebo effect, only a variable part of it - depending on the patient. Most of the time the 'effect' is effective enough to overcome what amounts to 'negative beliefs', but in your case it wasn't. Even looking at it as a 'reverse placebo effect' is accepting that there is a placebo effect which is not always positive. That is what I mean when I say that some believe that the net effect always includes the placebo effect. D. Matt Innis 16:25, 10 January 2009 (UTC)

I just can't agree with that. Perhaps a Zen master could resist the effects of a general anesthetic or paralyzing agent, but quite a few drugs cannot be resisted. I have watched patients in chronic neurogenic pain despertately want relief, but the interdisciplinary pain management team might try half a dozen therapies, all with suggestion they would work, before getting any results. There was an excellent review on Medscape dealing with neurogenic pain, where a pain management specialist said that he wished he could know whether anticonvulsants, cardiac stabilizers used off-label, capsaicin, opioids, or other agents would work with a given patient, but, even with encouragement, they didn't work.

Beliefs rarely have much to do with infectious disease. If cellulitis is due to a resistant organism, until the right antibiotic is found or there is surgical intervention, I doubt very much that belief has much to do with outcome. Again speaking from anecdotal experience, I myself suggested cephalexin for an infection in my leg, and was surprised when it didn't work. A fluoroquinolone was then selected and did work. If staphylococci are methicillin-resistant, the approach isn't to reassure the patient and use more, but to get cultures and sensitivity, give vancomycin, and change again if the sensitivity gives different answers or there's no clinical response. Howard C. Berkowitz 17:07, 10 January 2009 (UTC)

Bibliography

Have to say I'm with Matt on this one; I think we have to assume that with just about any medical intervention for any condition there are likely to be at least effects of stress, anxiety, expectation and trust on the overall outcome, for many reasons, and possibly conditioned responses also. Psychological studies show all kinds of things have an effect on many symptoms - both objective (heart rate etc) and reported - ritual, authority, even the colour of pills - and their price and packaging. Doesn't mean that conventional meds aren't effective - just that a (quite large) part of their effect is attributable to placebo - and because in some conditions the placebo effect is so large there has to be some doubt about whether some conventionat treatments are any more than placebos. But let's not go there here, so to speak confusingly.

Susceptibility to infection is certainly affected by the HPA axis, through its effects on the immune system. Central defences to infection are also powerful, and include pyrogenesis - and these hypothalamic responses can be conditioned so they are at least capable of being modulated by expectation. Few medicines are effective in all cases - there are usually "responders" and "non-responders"; the same is so with the placebo effect. However, attempts to profile placebo "responders" gives, as Ben Goldacre put it, an astrology-type portrait that could pretty well be anyone. Anyway, to start maybe a next phase on this article I've been mining the literature, and you'll see on the Bibliography page the fruits so far. Please add anything that you think is germane there, and maybe we can expand the article from the evidence base? - and Howard - maybe you should have taken a placebo first time - maybe that would have worked :-) Gareth Leng 18:56, 10 January 2009 (UTC)

- I'm startled by how much has come out on the placebo effect in the last couple of years. I know there's been a sea-change underway in medical attitudes here though - not long ago in academic medicine the placebo effect was rather sneeringly dismissed as equivalent to no effect. Now it is realised just how important the patient-doctor relationship is to outcomes - and medical students are being taught how to speak to patients - and about time too. If this is all we learn from CAM, and it's probably not, it's an important lesson. Overprescribing is a massive problem, and we need to address it. The Lancet study in homeopathy has an important moral for conventional medicine - small trials are unreliable; it's a weak evidence base, and it seems likely that many treatments based only on small trials are placebo effects.Gareth Leng 19:21, 10 January 2009 (UTC)

RE:the unreliability of small trials as learned from homeopathy. I saw that in the article, but hadn't put two and two together that medicines might also fit this scenerio. D. Matt Innis 21:29, 10 January 2009 (UTC)

Did you notice the discussion about 'natural history of a condition'. This needs a mention and discussion allits own. D. Matt Innis 00:13, 11 January 2009 (UTC)

Sorry, did you mean "natural history of a condition" as in the Lancet study, or elsewhere (e.g., a discussion that explicitly includes individual genetics and environmental (but not treatment) factors [1]], or one that does not, as in the more general epidemiology text at hand (Medical Epidemiology, Greenberg, Daniels et al.)?Howard C. Berkowitz 00:32, 11 January 2009 (UTC)

My fault, no, I meant in the bibliography that Gareth is building there is at least one article discussing the use of 'natural history of a condition' and comparing that to the treatment and then to the placebo. It talks about stomach conditions that limit themselves tend to have a spike in pain then decrease in intensity over a perod of time. Treatments that tend to lower the peak pain and lessen the duration are considered feasible. They discuss that as long as the placebo is as effective as the treatment, why not consider it. That is what I meant by comparing treatments to the natural history of the condition rather than to placebo. But that has nothing to do with this article, so I'll stop with that. D. Matt Innis 01:26, 11 January 2009 (UTC)

I'm still looking for the specific article I want, but here are a couple of studies, one a pharmacologic review, , and another for nurse practitioners, on the range of pharmacologic options available for neurogenic pain. Now, let me make a few observations of interdisciplinary pain management practices. First, they tend to be friendly in terms of CAM, especially mind-body medicine. I don't know of any that don't consider relaxation and visualization methods part of the first-line approach.

Most of the pain management physicians, and all the nurses I know in the field, do use drugs with a fairly strong suggestion that an intervention is likely to be beneficial. Time, especially advanced nursing, is spent with the patient; there are usually support groups; nonpharmacologic measures are common as well. Yet it's a pleasant surprise when the first or second drug works.

As far as clinician-patient interaction, a number of years ago, I was involved with a study, which started at Mass General and we were modifying at Georgetown, which compared OB/GYN history taking by residents and by computer. The computer was preferred, and the major reason given was that it was more patient and humane. :-(

Pain management and palliative care programs are among the most interdisciplinary in medicine, and make significant use of support and suggestion. Still, they often have much trial and error in treating chronic pain. If the placebo effect were that omnipresent, I would think they'd be a lot less frustrated. Yes, I recognize that a true chronic pain patient may tend to believe that nothing will help, and may be clinically depressed. Still, I don't think they depend that much on placebo effects. Howard C. Berkowitz 19:52, 10 January 2009 (UTC)

Nobody depends on the placeo effect, but that does not mean it is not there. You are familiar with 'the white coat' effect, right? D. Matt Innis 21:40, 10 January 2009 (UTC)

Laughing...you remind me of a physician I fired during a first visit in April. White coat would have been much better than her outfit, which I would have delighted in seeing socially, perhaps even to the black leather and matching boots. She didn't actually have a whip, but otherwise that sort of authority demand. I'm afraid I got to the point of asking her which part of NO on a particular request that she failed to understand...was it the N part, the O part, or the HIPAA part. It was, if anything, a sideways placebo effect; she annoyed me enough that I went back to my primary care physician and decided we really could do without the consult, and thought out what to do on our on.

Nobody depends on the placebo effect...are you a Monty Python fan, as in No One ever Suspects the Spammish Inquisition?

Suggestion is valuable, as is the relationship between clinician and patient. The literal laying on of hands definitely has its place. The low-responder problem is real, although I wonder how much pharmacogenomics will predict. That obviously won't happen overnight and still won't determine everything.

What were you thinking about with an article on natural course of disease? Untreated vs. placebo? Howard C. Berkowitz 03:58, 11 January 2009 (UTC)

Related topics

Did you see Larry's questions on Talk: Sham treatment/Related Articles? I'm not sure of the best approach, but the commonalities and differences among placebo, placebo effect, and sham treatment aren't coming across well. Do we want one article to make better compare-and-contrast, or more transition and crosslinks? Howard C. Berkowitz 19:52, 10 January 2009 (UTC)

I saw that discussion and its too early for me to say.  You know how I like to write then split topics off.  This drives people crazy, I know. I'll leave that up to you guys to determine the best way to handle it if you must do it now. D. Matt Innis 21:44, 10 January 2009 (UTC)

I think we need to define the scope of the articles clearly. Placebo is a well developed article on placebos as used in clinical trials to help establish the specific efficacy of an intervention, and the associated ethical concerns::This is intended as an article on the placebo effect per se; what it is and how it might work - not specifically in a clinical context, informed by clinical studies but also by psychological studies on suggestion, and by preclinical and basic work about what brain processes are involved. But an article for the general reader. We need to explain what it is, the history, explain what it appears to be capable of achieving, and discuss what is known of the underlying mechanisms. Sham treatment, I haven't looked at yet; this is different again; - sham treatments incur other confounding variables that are not part of the placebo effect.

I think we look to the audiences - placebo effect has to be general interest  - things like hypnotism/mesmerism should be mentioned; placebo has a defined medical trial context, sham treatment a specialist technical audience maybe? What do you think? I think that placebo effect should aim at a wide lay readership; I've loaded the bibliography page to provide (some) raw material, there's a very interesting history too, around mesmerism for instance, and the origins with Cullen - Cullen developed his ideas in the context of a concept of "sympathy" - very closely alligned with the use of the same word by two of his friends - David Hume (any comments here Larry?) and Adam Smith.

Yes, we don't want people to go rushing to stock up on expensively packaged placebo pills, even though they work best when they're expensive and well packaged. We shouldn't overstate the effectiveness nor understate it; there are well documented examples of people in particular psychological states being apparently insensitive/impervious to pain that would normally be incapacitating; the capacity of the CNS to suppress pain seems to be almost unlimited. So I don't find it surprising that in some circumstances, suggestion alone might suppress the perception of even severe pain.  As for organ conditions; not very long ago, organs were assumed to be largely autonomously functioning; now we know that virtually every organ and tissue is extensively innervated - even fat stores are directly innervated by neural pathways descending from the hypothalamus. The capacity of the CNS to influence selectively peripheral function is apparently considerable. This is not under conscious control - but autonomic reflexes can be conditioned - so there is an element of subconscious control. Control of the immune system - neuroimmunology - is a rapidly expanding field; it's not just HPA axis, important though that is. I'm not saying we should get into these things -  but we should be aware, in the background, that when we talk about the power of suggestion - there are lots of effector mechanisms that might be involved in translating that woolly "vitalistic" notion into mechanism.Gareth Leng 11:17, 11 January 2009 (UTC)

Apropos of pain or pain medicine, it's painful to discover that my copy of Melzack & Wall isn't on the shelf, but I'm afraid is in storage. I'm not sure where it should go, and want to find some good references, but there's been an autonomic pain response described that isn't always addressed by pharmacologic means. The theory is that even if higher brain centers are blocking pain perception, there are autonomic signals sent to the site of certain trauma, which causes at least prostaglandin release in the immediate area of trauma. This triggered local muscle spasm and inflammtion, which was postulated as a survival benefit, if rigid muscles acted as an "internal splint" to help you run away from the cave bear or whatever primitive threat had to be escaped. That resultant inflammation, however, produces pain of its own. The therapeutic implication was that even if the pain required opioids, it was still appropriate to use NSAIDs or even corticosteroids, and certainly heat, manipulation, etc., to reduce that secondary pain generation by the autonomous nervous system.

Yes we should link to an article on the Gate Theory of pain - don't know if we have one yet.Gareth Leng 10:39, 13 January 2009 (UTC)

I think I'm responding to Gareth, but there may be multiple posts. Just a couple of clarifying bullets, not implying anything is right or wrong. I'm just not sure where things should go, and possibly in more than one place. I may try to draw a picture, such as I did in the CZ Talk: Healing Arts material, not intended to be definitive but a discussion aid. Howard C. Berkowitz 18:40, 11 January 2009 (UTC)

• Should placebo be limited to research uses?

• Where is the best place for the clinical use of placebo and associated ethical issues? (There may be an interesting spinoff into direct-to-consumer advertising of prescription drugs, a pernicious American practice of often-misleading TV advertising. What does the clinician do when a patient demands a drug based on advertising and no other information, yet preserve the doctor-patient relationship?)

Let's leave those in Placebo for now - may be better later to have a separate article on the ethics of placebosGareth Leng 10:39, 13 January 2009 (UTC)

• sham treatment, I believe, is the place where the administration/performance of the control arm may, if not being risky, still has real effects, such as the mild trauma of injection. I'm trying to remember a very good article that showed sloppy design: it was a wound cleansing system, with a very predictable mechanically-driven irrigator that was precisely placed, while the control was a handheld 10ml syringe of saline.

there are other more dramatic examples - sham ECT especially.Gareth Leng 10:39, 13 January 2009 (UTC)

If we take the broader issue of suggestion in all contexts, not just from a clinician, do we want something on advertising and not-always-authoritative lay sites? If I can find it from my mail system crash, I have some material that I sent as a comment to an author and editor of a Hastings Center article on informed consent, but, to my surprise, they said they may publish on a space available basis. At the moment, I'm being very cautious about starting new CZ articles, so I'm not going to take that on in the near future. Howard C. Berkowitz 18:40, 11 January 2009 (UTC)

O yes, there are great studies on packaging effects and effects of pricing.Gareth Leng 10:39, 13 January 2009 (UTC)

Mutters about escorting fair ladies to the upscale cosmetics counter, actually reading the ingredients in the incredibly expensive "face conditioner", and thinking what that would cost if made with reagent grade chemicals and stored in high-grade laboratory glass -- as opposed to a work of artistic packaging that slides out of fingers lubricated with beauty unguents and smashes on the floor, creating yet more revenue stream. Howard C. Berkowitz 16:24, 13 January 2009 (UTC)

You described the local reaction to injury at a specific site.  The higher centers of the brain react something like this Neuroimaging of Acupuncture Specific versus Placebo effects in the Brain.  You ought to be interested in that one, Howard, it is cutting edge on acupuncture.  Norm Kettner was my radiology instructor at Logan when I was young and much more viral ;)

My understanding is that two mechanisms exist. Obviously, there's the local reaction with the kinin-kallikrein system. As I remember, and cursing I don't have a medical library anywhere close and even can't find my pain textbooks, there is a more newly discovered mechanism, which still may operate even if the gate stops propagation of pain perception. It is in the autonomic system, is related to the signals coming into the dorsal horn, but has an additional pathway that can trigger local prostaglandin (at least) release independent of inflammatory substances separate from local kinin-kallikrein. The therapeutic argument was made that while pain perception might be blocked by opioids or some anesthesia, healing and comfort could still be improved by concomitant use of anti-inflammatory drugs (with due caution) to reduce the brain-mediated as well as local response. Howard C. Berkowitz 16:24, 13 January 2009 (UTC)

I'm game for whatever you guys decide concerning article scope. D. Matt Innis 19:00, 11 January 2009 (UTC)