RE: [NMusers] Minimum patients number...

Hi

Having 2 DVs does make the process more complex - but it depends =
entirely on
the relationship of the DVs to each other and how you want to do your
analysis that determines the complexity of the problem.

If you had 2 DVs in the form of a nested single model multiple response
framework (as in PKPD):
1) If you intend to model the data simultaneously, then the PK model is
nested within the PKPD model and hence PD data will provide information
about both the PK and PKPD components of the model - whereas PK will =
only
yield information about the PK model. So the design needs to be aware =
about
the amount of information that is yielded from each of the DV types =
based on
each of the models.

2) If you intend to model the data sequentially, then in essence the PK =
and
(PK)PD models need to provide sufficient information to stand alone. =
This
setting is obviously not preferred (but sometimes necessary) and in =
essence
the designs can be considered separately as if they were independent. =
If
the 2 DVs are strongly linked (e.g. parent-metabolite) by common =
residual
variance parameters then their interaction cannot be quantified easily =
in
the naïve sequential analysis.

If you have 2 DVs that are not nested then the situation becomes more
complex.

The main point here - is that rules of thumb are probably not sensible =
as
they depend on too many non-quantified guesses that cannot be justified. =
 I
think you may as well just go and figure out how to quantify the
informativeness of your design using your preferred method.

Steve
--
Professor Stephen Duffull
Chair of Clinical Pharmacy
School of Pharmacy
University of Otago
PO Box 913 Dunedin
New Zealand
E: stephen.duffull_at_otago.ac.nz
P: +64 3 479 5044
F: +64 3 479 7034

Design software: www.winpopt.com
  

> -----Original Message-----
> From: owner-nmusers_at_globomaxnm.com
> [mailto:owner-nmusers_at_globomaxnm.com] On Behalf Of AJ Rossini
> Sent: Sunday, 28 January 2007 9:38 p.m.
> To: Mark Sale - Next Level Solutions
> Cc: nmusers_at_globomaxnm.com
> Subject: Re: [NMusers] Minimum patients number...
>
> Greetings -
>
> You have twice the data (total observations), but only the
> same amount of information (independent observations), and it
> is the latter which drives the precision/accuracy of the
> estimation procedure, adjusting downwards if the total number
> of observations goes up (adjusting upwards with the
> correlation of the additional observations, i.e. multiple
> measurements of "pretty much the same thing").
>
> best,
> -tony
>
> blindglobe_at_gmail.com
> Muttenz, Switzerland.
> "Commit early,commit often, and commit in a repository from
> which we can easily roll-back your mistakes" (AJR, 4Jan05).
>
> On Saturday 27 January 2007 20:49, Mark Sale - Next Level
> Solutions wrote:
> > Mats,
> > The "rules of thumb" are indeed empiric observations from one
> > variable. But, if you have two dependent variables, you have twice
> > the data (one time point, two DVs), and twice the
> parameters. So, I
> > think the rule may be consistent, but I don't know how
> valuable it is
> > in the first place.
> >
> >
> >
> > Mark Sale MD
> > Next Level Solutions, LLC
> > www.NextLevelSolns.com
> >
> > > -------- Original Message --------
> > > Subject: RE: [NMusers] Minimum patients number...
> > > From: "Mats Karlsson" <mats.karlsson_at_farmbio.uu.se>
> > > Date: Sat, January 27, 2007 11:41 am
> > > To: "'Ahmed Hawwa'" <ahawwa1_at_yahoo.com>, <nmusers_at_globomaxnm.com>
> > >
> > > Hi,
> > >
> > > I think that all the rules of thumb that have been mentioned are
> > > based on the assumption that you have a single observed
> variable. If
> > > you observe several, you certainly can expect to estimate
> more parameterers.
> > >
> > > Mvh,
> > > Mats
> > >
> > >
> > > Mats Karlsson, PhD
> > > Professor of Pharmacometrics
> > > Div. of Pharmacokinetics and Drug Therapy Dept. of Pharmaceutical
> > > Biosciences Faculty of Pharmacy Uppsala University Box 591
> > > SE-751 24 Uppsala
> > > Sweden
> > > phone +46 18 471 4105
> > > fax +46 18 471 4003
> > > mats.karlsson_at_farmbio.uu.se
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > -----Original Message-----
> > > From: owner-nmusers_at_globomaxnm.com
> > > [mailto:owner-nmusers_at_globomaxnm.com]
> > > On Behalf Of Ahmed Hawwa
> > > Sent: Saturday, January 27, 2007 02:10
> > > To: nmusers_at_globomaxnm.com
> > > Subject: RE: [NMusers] Minimum patients number...
> > >
> > > Hi all,
> > >
> > > Thank you all indeed for the valuable commenets. I have
> actually 19
> > > patients only and I need to look into the PK parameters of two
> > > metabolites.
> > >
> > > Minimum 10 patients per eta would mean 50-60 patients.
> > > I'm sure this number would give optimal results but as a
> matter of
> > > fact I don't think I would be able to reach it.
> > >
> > > > I would suggest that you assess each case on its merits and
> > > > determine the effectiveness of any proposed design using either
> > > > simulation linked with estimation (which is tedious, slow, not
> > > > optimal but often effective)
> > >
> > > Regarding using simultation do you mean to put certain values of
> > > thetas and then run a simulation and compare results with
> the actual
> > > concentrations or did I miss the point?
> > >
> > > > or an
> > > > information theoretic technique (such as optimal design).
> > > > The optimal design software WinPOPT (www.winpopt.com), which is
> > > > freely available, allows you to rapidly assess the
> effectiveness
> > > > of various designs as well as optimize a design within your
> > > > specific study constraints (e.g.
> > > > clinic visit times etc).
> > >
> > > I think one of my study constrains is clinic visit times since
> > > patients taking our drug are actually outpatients and we
> are able to
> > > collect one sample only at each clinic visit. would this
> software be
> > > of any help??
> > >
> > > Looking to hear from you soon.
> > >
> > > Regards,
> > > ahawwa
> > >
> > >
> > >
> > >
> > >
> >
> >_____________________________________________________________________
> > >____
> > >___ ________
> > > Cheap talk?
> > > Check out Yahoo! Messenger's low PC-to-Phone call rates.
> > > http://voice.yahoo.com
>
> e_at_gmail.com
>

Received on Sun Jan 28 2007 - 15:13:33 EST