Authors Margo McCaffery and Paul Arnstein respond: The APS position statement opens the door to a more general definition of N-of-1 trials, one that does not require IRB approval and informed consent. If both the APS and the ASPMN had used the same definition of N-of-1 trials in the two articles cited, we might agree with Winslow and Hutchison. We do not disagree with the use of placebo in N-of-1 trials for which patient consent and IRB approval are required and that involve randomization of treatment, blinding, and the use of predetermined outcome measures. But the APS did not use that definition, nor did they address the need for IRB approval in the body of their statement. The articles in which that description appears were cited by both organizations but not for the purpose of defining N-of-1 trials.

The APS position statement speaks of a single patient in whom response "to a placebo injection or temporary procedure, for example, might suggest that the risks inherent in a permanent procedure are not warranted at present."1 Such a positive response from a placebo given one time in a "trial" without blinding, IRB approval, or randomization is not a legitimate reason for denying the patient active treatment. All three elements-random order of intervention with blinding, written patient consent, and IRB approval-are needed to conduct a legitimate N-of-1 trial. This is why the ASPMN disagrees with the APS.

In our article we describe placebo use in a hypothetical N-of-1 patient situation, and not a randomized, controlled N-of-1 trial. The difference is that the situation we described did not involve IRB approval, blinding of the professional to the treatment, or written consent. The APS position statement could be interpreted to allow this type of trial or other trials missing just one of these three elements.

REFERENCE