Caroline and others,

A trial that is commonly used to support the value of an invasive sham surgical control intervention is that of Moseley et al. in evaluating the purported efficacy of Arthroscopic Surgery for Osteoarthritis of the Knee (1).

The rationale for the use of a placebo in a clinical trial is to 1) enable blinding, and 2) to control for the placebo effect. With respect to 2), I mean a so-called psychologically -mediated expectancy effect.  This particular effect is claimed to be important for certain particular outcomes, such as pain.

In the Moseley trial, the outcomes were: "Outcomes were assessed at multiple points over a 24-month period with the use of five self-reported scores — three on scales for pain and two on scales for function — and one objective test of walking and stair climbing."

Unlike the case with acute pain, it is certainly debatable whether the expectancy effect could have a meaningful effect on these "subjective" outcomes at 24 months. I'm not aware of any research that attests to the "durability" of these expectancy effects.

Another trial published in the NEJM on the same topic was that of Kirkley et al. (2) that did not employ the sham surgical control intervention. If, counterfactually, it had been reported before the Moseley trial, it would, arguably, have established the inefficacy of the procedure without the need for a trial employing an invasive sham control intervention!

As the degree of invasiveness of a proposed surgical control is increased, a useful manner of assessing its merits (approvability) is by employing the following notions:

1. Assessing the appropriateness of an invasive sham intervention

* Identify the specified outcome measures.

* Assess whether they are indeed susceptible to a placebo effect and/or potential reporting biases, and also in relation to the time point(s) when these measures are made.

* Consider whether an alternative research design may answer the research question(s).  

2. An alternative approach to determining the need for a sham intervention control

* Conduct a pragmatic RCT comparing the intervention to standard therapy (without the sham intervention).

* If the intervention is superior to standard therapy AND
If a credible assessment asserts that a psychologically-mediated expectancy effect likely dominates the postulated mechanism(s) whereby the intervention produces its beneficial effects, then…

* An explanatory RCT utilizing a sham control intervention may be justified to assess this further.


From: London et al. Placebos that harm: sham surgery controls in clinical trials Stat Methods Med Res. 2002 Oct;11(5):413-27.

.............

The last reference should be "mandatory" reading for those proposing, and evaluating the merits of, a trial employing an invasive sham surgical control.

It is available here:

http://www.hss.cmu.edu/philosophy/london/London--ShamSurgeryReasonableRisk.pdf 


Regards,


Howard Mann
University of Utah


1. N Engl J Med 2002; 347:81-88

2. N Engl J Med 2008; 359:1097-1107