The use of placebo in cancer research

Written by Vinjar Fønnebø and Edzard Ernst and the CAM-Cancer Consortium.
Updated March 11, 2011

Use of placebo in cancer research

By Vinjar Fønnebø, NAFKAM, University of Tromsø, Norway and Edzard Ernst, Complementary Medicine, Peninsula Medical School, Universities of Exeter & Plymouth, UK

When confronted with a cancer patient, the health professional faces several treatment challenges, some of them involve the use of CAM in one form or another.

A substantial proportion of cancer patients will seek CAM treatments at some point in their disease trajectory, and some of them will ask their GP, surgeon, oncologist or cancer nurse about it. When giving advice about these treatments it is important to have thought through in which situations one should look for placebo-controlled trials and when other trial designs are more relevant.

A placebo intervention is most commonly used in a clinical trial where the research question is to determine whether (most commonly) a drug exhibits a biological influence on the indicated endpoints in the study apart from important non-specific interventions like the clinician’s or patient’s expectation, enthusiasm, empathy, time and dedication and other therapeutic interventions occurring simultaneously. The placebo drug should be truly inert, meaning that it has no content that even theoretically can influence the disease endpoints measured in the study. It is also important in these trials that the patient, doctor and investigator are unable to tell the difference between the experimental drug and the placebo.

CAM “drug” treatments

If a CAM treatment is claimed to have an independent effect similar to a pharmaceutical drug, research should document the efficacy of the CAM treatment using the same methodology. Examples of this type of treatment would be Black Cohosh, Boswellia and Ukrain. In these trials all other aspects related to taking the drug (expectation, interaction etc.) are understood to be the same in those patients taking the real drug as in those taking the placebo, given that they are blinded with regard to what they are taking. Interpretation of these trials should be done using the same precautions as would be used in any double-blind, placebo-controlled randomized clinical trial.

Complex interventions

Most experts believe that things like a clinician’s expectation, enthusiasm, empathy, time and dedication can affect a patient’s outcome. There is also good evidence to demonstrate that this is, in fact, the case. Good clinicians use these and other factors wisely in order to help their patients. This means that, in many therapeutic situations, there is a more or less deliberate attempt to maximize the effects of these non-specific (or context) interventions, which often are called placebo effects. Most, if not all, medical interventions can therefore correctly be labelled complex interventions. If a treatment is prescribed or given as a complex system of several treatment components simultaneously or as a comprehensive package with sequential treatment delivery, the documentation of efficacy thus becomes more challenging. In these situations the treatment package often includes some of the non-specific treatments that in drug therapy research should be “controlled for”. Theoretically it might be possible to develop a truly inert complex treatment system that would appear identical to the intervention treatment, but this has not been demonstrated in practice.

In this situation we can only provide evidence of effectiveness, that is documentation of whether a package works as well, worse, or better than another package or different system of care. This ”control” system of care can either be usual care, best currently available evidence care, a specifically designed control system of care, or no care. It is important to remember that all complex interventions, whether they are conventional (e.g. consultation and conventional drug treatment) or alternative (e.g. consultation and homeopathic treatment) should be assessed in the same way.

Interpretation of these two uses of ”placebo”

It is important to make clear that when studying treatment packages, the research cannot determine efficacy of any single component of the complex intervention. A study showing that patients treated by a homeopath fare better or worse than patients treated by their GP for a certain condition cannot be used to claim that the prescribed homeopathic remedy works better or worse than a placebo remedy. Likewise efficacy research done on any component alone cannot be used to conclude on a system of care. Research showing efficacy or not of a homeopathic remedy cannot be used to claim that the system of homeopathic care is effective or ineffective.

Relevance for cancer

CAM treatments used by cancer patients are based on drug-like and complex interventions research. The reader of these summaries therefore needs to consider closely where placebo-controlled studies are appropriate and where other study types should be preferred. We will provide you with evidence at both levels, sometimes even within the same therapy (for example acupuncture/acupressure for nausea and vomiting).

Vinjar Fønnebø and Edzard Ernst, CAM-Cancer Consortium. The use of placebo in cancer research [online document]. March 11, 2011.