King Arthur: Now stand aside, worthy adversary.
Black Knight: ‘Tis but a scratch.
King Arthur: A scratch? Your arm’s off.
Black Knight: No it isn’t.
King Arthur: What’s that, then?
Black Knight: [after a pause] I’ve had worse.
King Arthur: You liar.
Black Knight: Come on ya pansy.King Arthur: [after Arthur’s cut off both of the Black Knight’s arms] Look, you stupid Bastard. You’ve got no arms left.
Black Knight: Yes I have.
King Arthur: Look!
Black Knight: It’s just a flesh wound.Monty Python and the Holy Grail

I am, I think, in a minority on this blog, in that I do not think there is a placebo effect. Period. None. Zip. Zero. Nada. Zilch.

For analysis purposes, I divide the lack of placebo effect into outcomes that do not occur with objective measurement and those that do not occur with subjective measurement.

Why the dichotomy? Those studies where there have been an active treatment, a placebo treatment and an observation group, have demonstrated no difference between observation and placebo (1). To summarize from the conclusion of the compelling NEJM review:

“We found little evidence in general that placebos had powerful clinical effects. Although placebos had no significant effects on objective or binary outcomes, they had possible small benefits in studies with continuous subjective outcomes and for the treatment of pain. Outside the setting of clinical trials, there is no justification for the use of placebos.”

Which makes sense to my reductionist materialistic way of thinking: why would actively doing nothing have any measurable physiologic effect? It shouldn’t and it doesn’t. Mind over matter? Bah, humbug.

How about subjective endpoints like pain? Isn’t there good data to support a placebo effect for pain? I saw the MASH episode where they ran out of morphine and relieved the post operative pain with placebo and the healing powers of caring and compassionate nurses and doctors (9).

Stick a knife in your finger. Pain receptors fire. Nerve impulses travel to the spinal cord for a quick withdrawal of your hand before you know you are hurt. The pain impulse travels up the spinal cord to the brain where you go ouch. Then you decide whether it hurts and if so how much (15). Ever watch a child injure themselves and look around to see if a parent is watching before deciding, at some level, that since Mom is watching, the injury is worth a good cry, some sympathy and a kiss?

There are several levels where pain can be blocked. Having had a crown this week, I am always grateful that a peripheral nerve can be blocked with some novocaine. Pain can be blocked with various medications that work at various levels in the CNS.

And then there is the appreciation of pain. How you react to the pain is different from the pain itself. Attitude can make the same amount of pain better or worse (6)

I had a herniated disc for about 9 months with chronic nerve root pain that was surgically repaired when all conservative therapies had failed (10). When I woke from the surgery my nerve pain was completely gone, and it was such a relief that in comparison the incision pain was nothing. I never bothered to take pain medications for the incision pain; I was so giddy to be pain free that the incision was nothing. Someone may argue that I was flooded with endorphins and that is why I had so little pain, but I have had enough other surgeries to wonder, if so, where were the endorphins during those other incisions where I was glad for the invention of the PCA pump.

I think that the placebo effect with pain is a mild example of cognitive behavioral therapy; the pain stays the same, it is the emotional response that is altered.

There was a recent study that sheds some light into the placebo myth.

“Sham device v inert pill: randomised controlled trial of two placebo treatments ” from the BMJ (11)

This was part of a larger study that demonstrated that sham acupuncture is better than real acupuncture (12). They took 270 patients with arm pain due to repetitive use injury of at least 3 months duration and pain rated as >= 3 on a 10 point scale.

“Participants could have any of a range of clinical diagnoses involving the tendons, soft tissues, and nerves of the arm or non-specific symptoms related to repetitive movement or overuse. We excluded people with systemic connective tissue or muscular diseases, neurological disorders, or acute trauma to the arm…This condition is the modern equivalent of “weaver’s hand,” “sprout picker’s thumb,” and “scrivener’s palsy.” Most patients had pain due to computer keyboard and mouse use.

The patients were randomized to sham acupuncture 2x per week with a blunt device that felt like acupuncture but did not penetrate the skin. They had 5 to 10 sham needles in the arm and one in the foot. The other group received a placebo pill one time a day, which they were told to take at night as they were warned it may cause drowsiness and dry mouth.

At first they were randomized to sham acupuncture or placebo as part of a run up to the study where they were then further randomized to real or sham acupuncture and amitriptyline or placebo, and they were followed after the second randomization.

Result? The pain scores (subjective findings) decreased more in the sham acupuncture than the placebo group. The magnitude of the pain relief was not impressive as admitted in the discussion: “In this trial, the magnitude of this effect was small.” Objective findings, no surprise, did not change in either group (grip strength or arm function).

More patients in the sham acupuncture (75%) thought they were receiving active therapy than placebo (48%) and the believers had the better subjective, but not objective scores.

Two different placebos decreased pain, one placebo was better than the other. Why?

The answer may be in another interesting result of the study: side effects. Patients were told in the informed consent what the side effects of the active therapies were, even though initially they all were getting placebo. And the informed consent worked to ‘cause’ side effects: three of the placebo subjects dropped out from dry mouth and fatigue and 10% of the sham had increased pain after the needle was ‘removed’.

Also, the sham acupuncture group had more interaction with their providers. So more chance of developing expectations, important in placebo effects

My take: Aint no such thing as a placebo effect, only a change in perception.

Take a different example of how expectation can alter the perception of a complex sensory experience. The example is taste.

I know that the same bottle of Bordeaux tastes far better in a 2 star restaurant on a summers day in the south of France with friends than it does on a January day alone with a bowl of top raman.

The wine will taste better if the bottle has a higher price tag. There was a recent study where people were given the same wine, but were told it cost 10 dollars or 90 dollars. They thought the 90 dollar wine tasted better (2). They did functional MRI’s on the participants as well and when they had the 90 dollar wine the part of the brain that deals with pleasure lit up but the same area did not light up with the 10 dollar bottle of wine.

I would argue the molecules in the wine did not change, the same receptors were activated in the mouth and nose, and information conveyed to the brain did not change, but the perception of the wine did. The wine became better, even as it stayed the same.

Same thing happens, I wager, with the so called placebo effect. No physical change. Pain is unchanged but the perception of the pain is changed. People have the result they expect to have and the side effects they are told they will have.

Outside of alt med, perception does not alter reality (13). People can talk themselves into believing anything. You may think that is nonsense, but in medicine you get to see the extremes of human experience. People will see things that are not there and not see things that are (16).

I remember the first time I saw a patient with this issue. I was a 3rd year medical student in the ER and a patient came in feeling lightheaded. I went through the extensive review of systems med students do, and found nothing, so I moved on to the exam. He had a blood saturated bandage under his gown, and beneath the bandage was a bleeding melanoma the size and shape of large mushroom that had oozed blood to the point where his red cell counts were 25% of normal. When asked about the tumor, he said it was nothing. He denied the huge bleeding tumor was a problem.

That is not all that an unusual experience: patients with horrible pathology that they deny exists. Patients present with huge tumors or endstage AIDS and completley deny they have anything wrong. At the other end of the spectrum are the patients who have nothing wrong with them but are convinced that every bump or twinge is a manifestation of some horrible, soon to be fatal, illness. After seeing these extreme examples, someone thinking their pain is decreased, when in fact it isn’t, requires no stretch of the imagination.

In humans there is no ‘real’ effect from placebo.

It was also mentioned in this blog that animals experience a placebo effect. As is so often the case, opinions go unreferenced, but my first response was disbelief. How could a dog have a non existent experience?

So I went looking.

It would be critical in determining if there is a true placebo response in animals that there was an observation group, a placebo group and a treatment group. Even more important would be that those collecting the data have to be blinded to the intervention. It keeps coming back to N Rays. People see what they want to see, what they expect to see, not necessarily what is there (3).

I admit up front that I do not read the veterinary literature and am limited in my access to this information compared to human medicine. So I may have missed some important literature.

That said, my start was a 1999 review of placebo in animals (4). It is a good discussion as to why it is difficult to assess the placebo effect in animals and its potential mechanisms. Then it moves on to clinical studies.

“Identifying placebo effects in veterinary medical studies is problematic. To the authors knowledge, studies specificly examining the placebo effect in therapeutic trials have not been reported. Placebo data for treatment trials is available only for control group results in studies investigating specific treatments. Most importantly, in trials in which placebo is selected as the control method, untreated groups serve as second controls to distinguish placebo effects other causes of disease resolution are not used. Even if a beneficial response is detected in the placebo group, such improvement cannot be conclusively attributed to the placebo, because non treated individuals are not included in the study. Accordingly, specific placebo data in studies of therapeutic effects is sparse.”

He then discusses a double blind study of carprofen in dogs (5). While the drug was better than the placebo, 56% of the dogs objectively improved on placebo (as judged by force plate measurement), better than the subjective improvement as judged by the owner or the vet. I do not know, as I cannot yet access the article, what the activity of the animals was during the study. If I had a dog that was diagnosed with hip problems, I would exercise it less, which would also lead to improvement that could be attributed to placebo.

A Pubmed search using “animals and pain” yielded 277 references that were clinical trials and had abstracts. I found few studies with objective end points that demonstrated an improvement attributed to placebo (7).

One study, which had a 60% placebo effect for gold wires injected for hip dysplasia, makes no mention of animal activity, so making judgements as to response to placebo is problematic (8).

I tried other combinations of search terms using placebo and animals and could not find well done blinded studies that demonstrate any placebo effect in animals. If there is compelling data that demonstrates placebo effects in animals where confounding issues are controlled (14), I cannot find it. I am sure the readers will send me the references to let me know I am wrong.

I conclude placebo effect is a myth: for humans and for animals.

References and Snotty Comments

(1) IS THE PLACEBO POWERLESS? An randomized of Clinical Trials Comparing Placebo with No Treatment N Engl J Med, Vol. 344, No. 21 • May 24, 2001

(2) (

(3) This is a huge issue in clinical medicine. If you think you have the right diagnosis, you may not see the information that contradicts your diagnosis or suggests an alternative. Good docs are aware of this.

(4) J Am Vet Med Assoc. 1999 Oct 1;215(7):992-9.
The placebo effect in animals. McMillan FD. PMID: 10511866

(5) Randomized, controlled trial of the efficacy of carprofen, a nonsteroidal anti-inflammatory drug, in the treatment of osteoarthritis in dogs.
J Am Vet Med Assoc. 1995 Mar 15;206(6):807-11.
PMID: 7759332

(6) Curr Opin Anaesthesiol. 2007 Oct;20(5):435-9.
Nocebo hyperalgesia: how anxiety is turned into pain.

(7) J Vet Intern Med. 2007 May-Jun;21(3):410-6.
Evaluation of client-specific outcome measures and activity monitoring to measure pain relief in cats with osteoarthritis.

(8) Acta Vet Scand. 2005; 46(2): 57–68. Stratification, Blinding and Placebo Effect in a Randomized, Double Blind Placebo-controlled Clinical Trial of Gold Bead Implantation in Dogs with Hip Dysplasia

(9) Season 6, episode 24. It’s on TV, it must be so.

(10) No, I did not use acupuncture.

(11) Sham device v inert pill: randomised controlled trial of two placebo
treatments. 2006;332;391-397. BMJ

(12) Not their intent, I am sure. As if there is a difference between real and sham acupuncture. Like the difference between real and sham ghosts.

(13) And the reality of WMD’s. And don’t give me the “What the Bleep Do We Know?” bs.

(14) Thanks to Dr Hall:
“animals ought not to have a placebo effect”
Maybe they “ought not” but they clearly DO. The veterinary literature recognizes that.
“I wonder how this can be explained.” Here’s how:
(1) They can develop a learned physiologic response to a drug and then respond similarly when a placebo is substituted.
(2) They respond to attention and care from humans.
(3) Human owners can experience the placebo effect for their pets by perceiving a response where there really is none.
(4) Since animals can’t talk, we have to interpret an animal’s behavior as indicating relief of pain; this may not always be accurate.

(15) Yes, I know, simplistic. Dr. Novella is probably squirming.

(16) Sometimes like essays in blogs, which are like Rorschach tests. Or maybe Horshack tests. Welcome back.

Posted by Mark Crislip

Mark Crislip, MD has been a practicing Infectious Disease specialist in Portland, Oregon, since 1990. He is a founder and  the President of the Society for Science-Based Medicine where he blogs under the name sbmsdictator. He has been voted a US News and World Report best US doctor, best ID doctor in Portland Magazine multiple times, has multiple teaching awards and, most importantly,  the ‘Attending Most Likely To Tell It Like It Is’ by the medical residents at his hospital. His growing multi-media empire can be found at