I was very pleased to see that the update of CONSORT asks for description rather than use of what is inherently unclear terminology. Moreover, I was even more pleased to see that assessing the success of blinding has been removed. In my experience of trials it has never been possible to know what was being assessed here. I guess it is more meaningful if the treatment is not effective, but even so it is , to my mind, over-rated.
Strictly speaking in many trials true blinding is very difficult to achieve, but it is important to aim for this and important to describe how it has been done. For the issue of blinding clarity will usually require description.
Anna Hart
>>> Keith Rennolls <[log in to unmask]> 4/19/2010 11:07 pm >>>
Dear Doug, and others,
When grant applications are written, important clinical trials are
designed, and when the results are published, it is usually regarded as
essential to use the words "double blind".
But, how many such important clinical trials actually evaluate the
efficacy of the blinding process, as part of the trial?
After all, is it realistic to expect that every single patient and every
clinician (to include all assessors) will be completely in the dark
about the treatment a particular subject receives. Treatments often have
effects (fortunately) which must, at least a partially, break blindness.
We might define a "truly double blind" clinical trial to be one in which:
(i) the subjects' "best guesses" of their treatment regime turns out to
be "random". (this would need to be elaborated)
(i) the clinicians' "best guesses" of the treatment regimes of patients,
(or a sample of patients), turns out to be "random".
In the case of the clinicians, their best-guesses might "random" when
the probability they assign to any given treatment is the same as the
proportion of the treatment in the trail. One must assume that the
clinician is honest in giving a best guess.
How many truly double blind clinical trials are there?
I suspect very few.
Some years ago, when working in the MRC on one of the elderly mild
hypertension trials, a small test of blindness was conducted. The
results were that blindness was not as blind as expected. There was
significant "sight" of both subjects and clinicians in reputedly double
blind trials. I cannot recall the figures, since this study was a small
aside 20 years ago.
If most "double-blind" clinical trials are not "truly double blind", but
are only "partially double blind", then clinical trials really should
assess degree of blindness of participants, and use this covariate
information in the analysis of the study. Of course, doing this would
not be simple or quick. But it should be possible in principle.
Is it worth the bother? There are strong reasons not to bother. But we
will never know if we never try.
Keith Rennolls
Emeritus Professor,
University of Greenwich.
Doug Altman wrote:
> Labels are convenient but are untrustworthy for scientific
> communication. Many terms are widely misused even when in principle
> they have a clear definition (although many don't even have that).
>
> Several terms used in clinical trials fall into this category - as
> well as "single blind" and indeed "double blind" and "triple blind",
> other examples include "randomised" and "intention to treat". When it
> really matters, eg in a publication or grant application, all such
> terms are inadequate and the authors need to say exactly what they did
> or plan to do.
>
> Doug
>
>
>
>
>
> At 19:16 16/04/2010, Dr Philip Sedgwick wrote:
>
>> Dear Allstaters
>>
>> With respect to clinical trials, I was taught the definition of
>> single-blind as the assessor was blind to allcoation but tthe patient
>> was not. The concept I suppose was based on the fact it was always
>> possible to blind the assessor to allcoation, but not necessarily the
>> patient. Needless to say I have no reference. I was further taught,
>> like most things with time, this definition has relaxed and it is
>> possible for the assessor to be aware of allcoation, but the patient
>> not - still resulting in single-blind trial.
>>
>> Does anyone have thoughts as to an original definition?
>>
>> Best wishes
>>
>> Philip Sedgwick
>>
>> St. George's, University of London
>> London SW17 0RE
>>
>> You may leave the list at any time by sending the command
>>
>> SIGNOFF allstat
>>
>> to [log in to unmask], leaving the subject line blank.
>
>
> _____________________________________________________
>
> Doug Altman
> Professor of Statistics in Medicine
> Centre for Statistics in Medicine
> University of Oxford
> Wolfson College Annexe
> Linton Road
> Oxford OX2 6UD
>
> email: [log in to unmask]
> Tel: 01865 284400 (direct line 01865 284401)
> Fax: 01865 284424
> www: http://www.csm ( http://www.csm/ )-oxford.org.uk/
>
>
> CONSORT Statement updated March 2010
> www.consort-statement.org
>
> EQUATOR Network - resources for reporting research
> <http://www.equator ( http://www.equator/ )-network.org/>www.equator-network.org/
>
>
>
>
> You may leave the list at any time by sending the command
>
> SIGNOFF allstat
>
> to [log in to unmask], leaving the subject line blank.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|
