Placebo vs drug – The power of positive and negative information

Most people have heard of the “placebo effect”. It basically refers to the fact that people receiving a treatment for a given condition may report an improvement in how they are feeling even though what they are actually taking is an inert substance. The placebo effect is typically observed when what is measured is a subjective symptom, for example pain. By contrast, it is virtually non-existent when looking at an objective marker, for example clearance of a bacterial infection. The placebo effect is therefore an important factor to be aware of when assessing the efficacy of a drug in the treatment of conditions for which the symptoms are evaluated subjectively, such as pain, depression or chronic fatigue.

The placebo effect means that if you think you are being treated, you may report a positive outcome even if what you are actually taking is a placebo. But what about the reverse situation? If you think you are taking a placebo, will that change the benefit you may derive from taking an actual drug? That’s the question researchers at Harvard Medical School addressed in a study published recently in Science Translational Medicine.

  • Study design

The individuals participating in the trial suffered from recurrent migraine attacks. They were asked to evaluate the level of pain they experienced for a total of 7 migraine attacks for which they received different treatments. Two pain scores were recorded for each migraine attack: the first score was taken 30 minutes after the beginning of the migraine attack (baseline) and the second one two hours later. After that, regardless of what treatment they had already received, the participants were offered rescue medication to relieve pain as needed.

For the first migraine attack that occurred during the study period, the participants did not take any medication for the first 2.5 hours. For the six subsequent migraine attacks, they took a pill 30 minutes after the onset of the migraine (just after having recorded the first pain score). This pill was either a placebo or an actual drug (Maxalt, 10mg), but each was given to the patient in an envelope labeled with either “placebo”, “placebo or Maxalt”, or “Maxalt” (click on schematic figure below for trial design). Two hours after taking the pill, the participants re-evaluated their pain score. The order in which the participants took the placebo or the drug, with the different labels, was randomized, so that it would be different across the study participants and any bias that a particular order might introduce would be avoided.

Study designThe effects of the different kinds of “treatment” were measured in terms of pain reduction, that is, by comparing the pain score at 2.5h (so 2h after taking a pill) with the pain score at 0.5h (30minutes after the beginning of the migraine attack, just before taking the pill).

The advantage of doing such a trial with patients suffering from migraine attacks was that each patient was its own “control”: the efficacy of drug and placebo in different conditions of information (namely the labels on the envelopes containing the pills) was evaluated for each individual subject, which means that differences arising from the fact that different people do not respond in the same way to treatment or might not report pain in the same way would not influence the results.

  • Main findings

When the researchers looked at the reduction in pain scores only by taking into account what the patients actually took, regardless of what they thought they were taking, they found that the actual drug was better at reducing pain than the placebo (48% decrease in pain versus 21% decrease 2h after taking the drug or the placebo, respectively). Nothing surprising there, though one might already notice that people reported a reduction of the pain they were feeling even if they had taken a placebo.

Now, here comes the more interesting part: when the researchers also took into account what the patients thought they were taking, they saw that the different ways of labeling the pills had an effect on the level of pain reduction experienced. This was true when the patients received a placebo (as expected: placebo effect) but also when they received the actual drug.

Here is then the answer to our initial question: If you think you are taking a placebo, does it change the benefit you may derive from taking an actual drug? Well, according to this study, yes, it does.

1. The efficacy of the drug in reducing pain was actually lower when the patients thought they were taking a placebo compared to when they thought they were taking the drug or when they did not know what they were taking.

2. The drug labeled as placebo turned out to be about as effective in reducing pain as the placebo labeled as drug. The drug tended to induce a bit more of pain reduction, but the difference was not large enough to be significant from a statistical point of view (meaning that this small difference could have happened by chance).

3. On the whole, the pain relief obtained from the pill was larger when the pill was labeled as being the active drug, and smaller when the pill was labeled as placebo, regardless of whether it was the placebo or the active drug.

The main takeaway message from the study is therefore that what a patient thinks he/she is taking (placebo, drug, or unknown) is important, regardless of whether the substance administered is neutral or pharmacologically active.

  • The power of the placebo effect

When comparing the efficacy of the placebo to that of the drug when both pills were labeled identically, the researchers found that the effect of the placebo was always 50-60% as large as that of the actual drug.

Importantly, even when the placebo was given openly as placebo (open-label placebo) the patients reported a reduction in their pain level: the pain score decreased by about 15% compared to baseline. By contrast, during the first migraine attack for which the patients had not taken any pill, the pain score had increased by about 15% compared to baseline. So it seems, at least in this study, that the simple act of taking a pill is by itself a non-negligible part of care.

However, there may be a limit to the power of placebo. Even though the open-label placebo seemed to reduce pain to some extent, when the researchers looked at a somewhat more clear-cut measure (being pain-free or not), they found that the proportion of people reporting being pain-free 2h after taking the open-label placebo was not statistically different from that observed when participants did not take any pill at all.

Another detail at which I have been wondering, and that was actually not discussed by the authors of the study, is that it was always the migraine attack that occurred first during the period of study that was used as the “control” one, the one for which no pill of any kind was taken. While this detail does not affect the findings of the study regarding how the labeling of the pills influences the pain reduction experienced by patients (which was the primary question addressed by the study), it might affect the finding that the open-label placebo pill has a beneficial effect compared to no pill at all. Indeed, it turns out that the average pain score reported at 30 minutes for the first migraine attack is lower than the average pain score reported at 30 minutes for all subsequent migraine attacks. Since the pain scores at 2.5h are similar between the first migraine (no pill) and the migraine for which the open-label placebo was taken, the net result is that there is a 15% increase in pain in the case of no treatment, but a 15% decrease in pain in the case of open-label placebo. Is it then possible that the beneficial effect of placebo observed even when it is identified as placebo is actually due to the participants consistently underestimating their pain level the very first time they were asked to do so?

On the whole, I thought the study particularly interesting, not just because it showed the existence of a placebo effect (that has already been shown in other studies), but more importantly because it revealed that the efficacy of an actual active drug could be diminished if the patients thought they were receiving an inactive substance. Of course, these findings are limited to the precise setting of this study (migraine attack, measured outcome, drug used), for now at least.

Altered placebo and drug labeling changes the outcome of episodic migraine attacks. Kam-Hansen S, Jakubowski M, Kelley JM, Kirsch I, Hoaglin DC, Kaptchuk TJ, Burstein R. Sci Transl Med. 2014 Jan 8;6(218):218ra5. doi: 10.1126/scitranslmed.3006175
PMID: 24401940

Details of the clinical trial can be found on (identifier NCT00719134).

ResearchBlogging.orgKam-Hansen S, Jakubowski M, Kelley JM, Kirsch I, Hoaglin DC, Kaptchuk TJ, & Burstein R (2014). Altered placebo and drug labeling changes the outcome of episodic migraine attacks. Science translational medicine, 6 (218) PMID: 24401940


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s