Over the years, I have heard and read many claims of dramatic placebo effects. The claims of strong placebo effects have not been replicable, or if replicated, have been based on studies with poor control and subjective outcome scales without blinding/masking, that is, measurement bias almost certainly occurred. The belief in a strong, generalizable placebo effects – a belief based on writings from the 1950s, as far as I know -- have been effectively disconfirmed. “Placebo effects” need to be much better and more clearly defined. More limited psycho-somatic effects still occur in certain situations, in my opinion, but not a strong generalizable real effect.

Sorry If I sound dogmatic, but over the years I have so many times -- hundreds and hundreds of times -- heard claims of “placebo effect” used as a cover for provision of “treatment” is without evidence base.

For this audience, I suspect that my message is just the need to clearly define what is meant by a “placebo effect” and how it can be distinctly measured.

Mark V. Johnston, Ph.D.

Professor, College of Health Sciences,

University of Wisconsin – Milwaukee,

(414) 229-3616

From: Evidence based health (EBH) [mailto:[log in to unmask]] On Behalf Of Klim McPherson
Sent: Thursday, June 09, 2016 6:39 AM
To: [log in to unmask]
Subject: Re: placebos and side effects

But also the placebo effect itself is sometimes surprisingly dramatic - we do not know that they will not help.

Klim

From: "[log in to unmask]" <[log in to unmask]> on behalf of Jeremy Howick <[log in to unmask]>
Reply-To: Jeremy Howick <[log in to unmask]>
Date: Thursday, 9 June 2016 12:03
To: "[log in to unmask]" <[log in to unmask]>
Subject: Re: placebos and side effects

The reason this can happen is that, as Tom pointed out, placebo ingredients are not reported.

Jeremy

T: +44 (0)1865 289 258 E: [log in to unmask]

http://www.phc.ox.ac.uk/team/jeremy-howick

Nuffield Department of Primary Care Health Sciences, University of Oxford
Radcliffe Primary Care Building, Radcliffe Observatory Quarter, Woodstock Road, Oxford, OX2 6GG

From: "[log in to unmask]" <[log in to unmask]>
Date: Wednesday, 8 June 2016 23:43
To: Tom Jefferson <[log in to unmask]>, Jeremy Howick <[log in to unmask]>
Cc: "[log in to unmask]" <[log in to unmask]>
Subject: Re: placebos and side effects

I cannot see how equipoise can be obtained under these circumstances: we know that placebo will not help but it may harm.

Ben

Sent from my iPad

( please excuse typos & brevity)

On Jun 8, 2016, at 9:49 AM, Tom Jefferson <[log in to unmask]> wrote:

Jeremy, it does not quite work that way. The active placebo trick is not played to even out all potential harms across intervention and control arms, just one harm or a group of harms.

So for example for Relenza which is a dry powder for inhalation the placebo was the lactose based eccipient which could trigger asthmatic attacks in susceptibles just like the "live" arm. Explanation: lactose powder was the eccipient in the active arm so that's fair (Gardasil and alumiunium is similar although no explanation for the choice of placebo is reported in the CSRs). FDA did not agree.

Different for Tamiflu (dehydrocholic acid in low doses to mimick the bitter taste of oseltamivir powder). Except that d acid also causes diarrhoea and gut upset which is a known harm of oseltamivir. FDA seems to have missed this one.

Real? Mistake? Who knows?

Hope this helps,

Tom..

Dr Tom Jefferson

Honorary Research Fellow

Centre for Evidence Based Medicine

Oxford OX2 6GG

On 8 June 2016 at 13:57, Jeremy Howick <[log in to unmask]> wrote:

Question:

‘Active’ placebos (that imitate some side effect of the experimental treatment) are sometimes used in clinical trials (http://www.ncbi.nlm.nih.gov/pubmed/14974002); this is good for improving success of blinding.
However given that placebo ingredients are rarely reported (http://www.ncbi.nlm.nih.gov/pubmed/?term=howick+golomb+placebo) this creates a potential problem for side effect estimates in trials:
The side effects of the experimental treatment are compared with the side-effects of the ‘placebo’. But if the placebo contains an ingredient that induces a side effect, there will (by definition) be no difference in side effects between the treatment and placebo groups.

Is anyone aware of a real example illustrating point (3): where the experimental treatment was reported to have ‘no side effects’ because there was no drug/placebo difference in side effects, yet where the placebo contained an ingredient that imitated a side effect?

Thanks in advance,

Jeremy