Updated: Ethical concerns about a placebo run-in (created 2006-10-04, revised 2010-02-04)

Dear Professor Mean, Some of the trials that our Institutional Review Board looks at have a placebo run-in period. All patients are given a placebo before the start of treatment and anyone who responds well to the placebo is dropped from the trial. What are the ethical ramifications of such a study. You can't disclose the placebo run-in period to the research volunteers because it would defeat the purpose.

I've often wondered if there is such a thing as a placebo research study--a study that looks just like a real research study, but which imparts no useful knowledge to the reader.

Sorry about that, I was just daydreaming. You actually ask an interesting question and there is a closely related question--what about a research study that assesses patient compliance during a placebo run-in period and excludes poor compliers from the main study.

You should start by reading

Stephen Senn, who is supportive of placebo controls in general, is still sharply critical of the placebo run-in periods. There is a critical distinction in that the use of a placebo requires withholding information from a patient, but the use of a run-in period requires deception.

You need to think carefully about the words "deception" versus "withholding information". It may seem like a difference in semantics, but it's not. In general, witholding information is much more common in research studies, and it causes fewer ethical concerns.

There's a lot of guidance on the use of deception in research, and the major principles are:

  1. Only deceive if there is no other practical way to do the research.
  2. Never deceive (or withhold information) about the risks of the research.
  3. Debrief the subjects afterwards.

There's a scientific validity question here as well. Do you have the option as a clinician to say to a patient "You look like a placebo responder to me, so I won't treat you"? Of course not. So when researchers exclude placebo responders, they are making it more difficult to generalize the results. Even worse, research excluding placebo responders is almost certainly going to overstate the value of a new intervention.

Researchers want to exclude placebo responders because placebo responders add noise to a study and when you exclude them, you strengthen the signal and allow the research to be done with a smaller number of patients. This is actually an admirable goal, and needs to be balanced against the loss of generalizability.

As a general rule, if a study is intended to establish a proof of concept, I might be inclined to allow exclusion of placebo responders. But if it is intended to provide pragmatic advice to practicing clinicians, the loss of generalizability outweighs the other considerations.

Here's some additional reading on this issue:

I've often wondered if there is such a thing as a placebo research study--a study that looks just like a real research study, but which imparts no useful knowledge to the reader.

Sorry about that, I was just daydreaming. You actually ask an interesting question and there is a closely related question--what about a research study that assesses patient compliance during a placebo run-in period and excludes poor compliers from the main study.

You should start by reading

Stephen Senn, who is supportive of placebo controls in general, is still sharply critical of the placebo run-in periods. There is a critical distinction in that the use of a placebo requires withholding information from a patient, but the use of a run-in period requires deception.

There's a lot of guidance on the use of deception in research, and the major principles are:

  1. Only deceive if there is no other practical way to do the research.
  2. Never deceive (or withhold information) about the risks of the research.
  3. Debrief the subjects afterwards.

There's a scientific validity question here as well. Do you have the option as a clinician to say to a patient "You look like a placebo responder to me, so I won't treat you"? Of course not. So when researchers exclude placebo responders, they are making it more difficult to generalize the results. Even worse, research excluding placebo responders is almost certainly going to overstate the value of a new intervention.

Researchers want to exclude placebo responders because placebo responders add noise to a study and when you exclude them, you strengthen the signal and allow the research to be done with a smaller number of patients. This is actually an admirable goal, and needs to be balanced against the loss of generalizability.

As a general rule, if a study is intended to establish a proof of concept, I might be inclined to allow exclusion of placebo responders. But if it is intended to provide pragmatic advice to practicing clinicians, the loss of generalizability outweighs the other considerations.

Here's some additional reading on this issue:

Related pages on this website:

This page was written by Steve Simon while working at Children's Mercy Hospital. Although I do not hold the copyright for this material, I am reproducing it here as a service, as it is no longer available on the Children's Mercy Hospital website. Need more information? I have a page with general help resources. You can also browse for pages similar to this one at Category: Placebos in research.