Feb 06 2007

Take Two Placebos and Call Me in the Morning

I often get questions about the placebo effect – what exactly is it, how can we exploit it, and what does it all mean. In my experience the placebo effect, briefly defined as a measurable response to an inert treatment, is almost completely misunderstood by the public – a fact that is exploited by purveyors of dubious treatments.

In order to demystify the placebo effect, I will try to first describe exactly what it is. The operational definition of a placebo effect is any health effect measured after an intervention that is something other than a physiological response to a biologically active treatment. In practical terms this is measured by comparing two groups in a clinical study, one group receives only a placebo (an inactive treatment) and the other received an active treatment (such as a pharmacologically active substance – a drug), and the outcome of the two groups is compared. If the placebo group has a 30% response (in whatever is being measured) and the treatment group has a 60% response, then it is concluded that the placebo was responsible for the 30% response and the active treatment for an additional 30% (it is assumed that the active treatment group also gets the placebo effect plus the treatment effect).

At face value this would make it seem as if the placebo has just as much effect as the added effect from the treatment. This leads to frequent questions, the two most common of which are: does this mean that placebos are legitimate and effective treatments in their own right; does this prove a “mind over matter” effect? The answer is no and no, and here’s why:

We need to look deeper by what exactly is being measured in that hypothetical 30% measured response. What is being measured and how. Subjective outcomes like pain, fatigue, and an overall sense of wellbeing, are subject to a host of psychological factors. For example, subjects in clinical studies want to get better, they want to believe they are on the active experimental treatment and that it works, they want to feel that the time and effort they have invested is worthwhile, and they want to make the researchers happy. In turn, the researchers want their treatment to work and want to see their patients get better. So there is often a large reporting bias. In other words, subjects are likely to convince themselves they feel better, and to report that they feel better, even if they don’t.

It has also been clearly demonstrated that subjects who are being studied in a clinical trial objectively do better. This is because they are in a clinical trial – they are paying closer attention to their overall health, they are likely taking better care of themselves due to the constant reminder of their health and habits provided by the study visits and attention they are getting, they are being examined on a regular basis by a physician, and their overall compliance with treatment is likely to be higher. So basically, subjects in a trial take better care of themselves and get more medical attention than people not in trials.

But what about the mind-over-matter effect? This effect varies greatly depending on the outcome being studied. Pain, for example, is entirely a subjective experience (although certainly a neurological phenomenon). The perception of pain can be modified by mood, and there are many non-specific factors that can even biochemically suppress pain. For example, increased physical activity can release endorphins that naturally inhibit pain. So the placebo effect for pain is typically high, around 30%.

But the more concrete and physiological the outcome, the smaller the placebo effect. Survival from serious forms of cancer has a very small placebo effect, limited to the objective factors listed above. There is no compelling evidence that mood or thought alone (independent of better care and compliance) can help fight off cancer or any similar disease.

Other conditions are more objective than pain, but are special because they have a strong influence from the neuro-endocrine system. This system translates psychological stress into physical stress, by releases stress hormones and increasing activity in the sympathetic nervous system. So, for example, for heart disease mood matters quite a bit. Someone who has an A-type personality and is always angry and upset is at higher risk of a heart attack than someone who is mellow and unstressed. But here there is a known physiological connection between mood and a specific organ – the heart. This cannot be extrapolated to other diseases; it doesn’t mean you can smile your cancer away.

Many people talk about the neuroendocrine system’s effect on the immune system. Again, here there is a physiological connection. Stress hormones do suppress the immune system, and it is probably true that extreme stress leaves us physically susceptible to disease for this reason. But the effects of moderate levels of stress are not established. Also, we cannot extrapolate from the risk of getting a cold to the ability to fight off cancer. You have to look at the evidence for each disease unto itself. So while this is a potential contributor, it is overall probably a small effect except in extreme situations.

So the bottom line is that the placebo effect is fairly complex and is largely an artifact of observation. Any real benefits that contribute to the placebo effect can be gained by more straightforward methods – like healthy habits, compliance with treatment, and good health care. It is not evidence for any mysterious mind-over-matter effect, but since the mind is matter (the brain) and is connected to the rest of the body, there are some known physiological effects that do play a role (although often greatly exaggerated).

In light of all this, I do not feel that knowingly prescribing a placebo treatment is effective or ethical medicine. Modern scientific medicine should strive for interventions that physiologically are scientifically plausible and have sufficient evidence for safety and effectiveness. But I do think there are lessons to be learned from the placebo effect – there are aspects of therapy that do go beyond the physiological intervention. Medicine is not only an applied science; it is the art of humans treating other humans. As part of effective treatment it is helpful to try to maximize all those human intangibles that contribute to a good outcome. But we can do this in the context of scientifically valid treatment, and without crossing the ethically dubious line of deception.

As I said above, real treatments have a placebo effect too.

One response so far