Following up on Steve's thought in light of the (very
interesting) piece by Sackett, I was struck by the latter's
observations that ' front-line clinicians do not want to
"generalize" an RCT's results to all patients, but only to
"particularize" its results to their individual patient, and
already routinely adapt the trial result (expressed, say, as
a "number-needed-to-treat" or NNT, which is the inverse of
the absolute risk reduction) to fit the unique risk and
responsiveness of their individual patient, ... the
patient's preferences and expectations, and the like'.
This seems eminently sensible, but I wondered whether such a
strategy is sustainable in all areas: in the light of
possible developments in new drug systems based on
pharmacogenetics, for example, the genotypic
particularisation of patients becomes a _requirement_ if the
benefical and harmful effects of a new compuond are to be
effectively controlled. Will patients' 'preferences',
clinical discretion, etc be relevant? Or will clinicians
avoid the finely grained segmentation of their patient
population in this way? More generally, pharmacogenetics
might encourage some to argue that trials for new compounds
need not go through the conventional RCT process since the
efficacy of the drug can be determined biogenetically -
trials become validated as a sort of lab experiment rather
than a trial and error across a large sample population that
looks for the narrowest confidence intervals. Any thoughts?
Andrew Webster
SATSU
Univ of York.
> "Simon, Steve, PhD" wrote:
>
> David Sackett wrote a fascinating article for CMAJ, and I
> was curious what people thought of his comments that
> generalizability is not all that we make it out to be
> (particularly on page 1232). I guess he has made this
> argument before, but this is the first I had heard of it.
>
> This seems to go against the recommendations of the users
> guides, for example (how similar is my patient?).
>
> Any thoughts from this group?
>
> Steve Simon
>
> Why randomized controlled trials fail but needn't: 2.
> Failure to employ physiological statistics, or the only
> formula a clinician-trialist is ever likely to need (or
> understand!) David L. Sackett. CMAJ 2001;165(9):1226-37.
>
> http://www.cma.ca/cmaj/vol-165/issue-9/1226.asp
|