A general thread on the PACE trial!

The "placebo response" in patients is highly likely to be the same as it is in any other group of people. The problem is a variety of reporting biases are being bundled up as a "placebo effect" and then claiming that the placebo effect* is magically high in some cases and low in others. When the real reason is the reporting biases are better controlled in some contexts, compared to others.

*placebo effect defined as: a beneficial effect produced by a placebo drug or treatment, which cannot be attributed to the properties of the placebo itself, and must therefore be due to the patient's belief in that treatment.

 

The “placebo effect” can also be nothing to do with expectations, but can show up because of the chance that some individuals out of both groups will make a spontaneous improvement. Ie Regression to the mean.

 

This is to a large extent what my grand round presentation was about.

I think there is a simple sense in which the placebo effect for healing of a broken bone is small and the placebo effect for anything where the ascertainment of outcome is subjective, as for pain, is likely to be larger. ME is bound to fall more in the second category at least in the wider context of a spurious positive result that gets bundled under placebo effect.

 

Regression to the mean is not a placebo effect, it is a bias that is often falsely attributed to the placebo effect.

 

Fair point. Unfortunately improvements in the placebo group are often all attributed to “the placebo effect” without any other acknowledgement of what can cause a placebo group to improve (or appear to improve)!

 

Is there a "real" placebo effect or is it just entirely due to hidden biases?

 

Yes, the real placebo effect is a transient reduction in pain due to conditioning of the bodies endorphin system. There is some biological evidence of this. This system has obvious evolutionary advantage, namely being able to temporarily suppress pain for the purpose of escaping danger.

 

But @Keela Too did put it in quotes - "placebo effect" so she was absolutely correct. Just as one might say, and Noam Chomsky did, that the "mind-body problem" has in fact been a body-body problem since Newton's theory of gravitation.

"Placebo effect" is all too often used to include regression to the mean by my colleagues.

 

What one might call semantic finesse, or finesse semantique?

 

The endorphin release associated with placebos could be part of the body's behaviour control system. It could be seen as the reward for finding a solution a for health problem.

It makes sense that it hurts until you find a solution to the pain. Then it no longer need to hurt as much and your efforts should be rewarded so that you continue to take care of yourself.

Seeing a doctor and leaving with the feeling that you've done something useful seems to count as "finding a solution". Even if the treatment is actually ineffective.

 

Another thought on PACE, and probably other trials. Given it is often ethically unacceptable to include a control arm having no treatment, each arm therefore has one of the interventions plus SMC (standard medical care). But what about the case where SMC actually conflicts with the intervention, as would seem to have been the case with PACE. How does that work out?

For example, SMC for pwME pre-GET, would have been much more inclined towards real pacing, and not pushing yourself beyond sensible limits, and to listen to your body; this is the sort of advice my wife received in 2006'ish. So how could you be having that treatment, whilst also having treatment encouraging you to ramp up your activity levels and to ignore your symptoms. That alone sounds bonkers.

ETA: Just discussing this with my wife, and she reminded me that the ME specialist we saw in 2007, recommended pacing, and enquired at great length how my wife managed her days, and how she paced herself. He told her that although they could offer CBT they were not doing so for my wife because he felt it would provide no additional benefit to her. So I'm just clarifying the advice my wife was given at the time.

 

This is from the SMC manual:

But it also says:

So I guess the SMC doctor would need to tailor the advice they gave depending on which group the participant was in.

 

Now this is getting weird. The current NICE guidelines (CG53) recommending CBT & GET were put in place in 2007. I’m not sure of timescales, but is it possible that some PACE trial participants were getting CBT/GET as part of SMC?

Or had the trial period ended before CG53 came into effect?

 

My memory is that recruitment to PACE was 2005-2007? The SMC patients clearly did not get CBT or GET but it does all seem very muddled and I think Barry's point is basically valid. Adding CBT or GET to SMC is not necessarily a reality.

 

And as patient experiences of conventional care differ widely, it is likely there was nothing standard about SMC!

On top of that many GPs use the less formal version of GET - it goes along the lines of “I’ve heard that exercise helps with ME, you should make sure to fit some exercise into your day if you want to improve.”

If SMC had to be defined by PACE (eg don’t offer CBT/GET to any trial patients as part of the SMC you offer them) then it really wasn’t “standard” medical care, but yet another form of care.

All very muddy indeed.

 

In which case the SMC-only arm was useless as a control.

 

Pretty much a textbook apples to citruses control.