On 19 Feb, Michel.guillaume wrote:
Hi Brian
> I paid a visit to Ronald F. Vetter <[log in to unmask]> 's WEB
> site " http://www.ridgecrest.ca.us/~rfvetter " and from there, using
> " Brian Collins " link, I got a very pleasant trip to your model
> description. You certainly are ahead of me using pharmacokinetics. I'm
> certain that we are working on the same wavelength.
> In the model that I'm using, the way doses are fading is described using
> the decreasing part of a gaussian distribution. i.e. the impulse response
> of a Modopar dose, is computed using the following Excel Formula :
> IF(h-t<0;0;IF(h-t<=m;a*(1-COS(PI()*(h-t)/m))/2;IF(h-t<m+p;a;a/EXP((h-t-m-p
> )^2/d^2))))
> Where :
> " h " is the current time
> " t " is the time at which the dose is taken
> " m " is the rise time (1.5 hour)
> " p " is the duration of the constant level (0.5 hour)
> " d " is the width at half level of the gaussian distribution
> whose center occurs at the time = t+m+p
> d = 1.5 hour for non CR doses an 4 hours for Modopar LP doses
> for memory, respectively a= 9.8 and 5.2
> This modelisation gives a very good representation of what the available
> Dopa is proportional to.
> It allows me to predict what is going to happen in the near future (with
> a +/- 5 minutes accuracy) knowing the state in which the previous dose
> has worked.
> MEG
>
Hello Michel, I must apologise for not responding to your most interesting
emails. I seem to have taken on a number of duties connected with rkinson's
Disease which do not allow me to spend much time at my computer.
I think I have understood the way that you make your program work- it is
interesting that we arrive at very similar coclusions, from quite different
directions.: considering the tail-off as the dose is falling away, the shape
of the curve in my model is derived in the same way as the rest of the drug
characteristics - by continually altering the shape of the curve describing
the drug (by dragging the points to different positions) until the effect of
those drug shapes, integrated across the whole day give a good match of
the observed to the predicted line.
I assume that you use the tailing off shape of the Normal distribution
because it "looks right" or is there a more technical reason? I must say
that the shapes of my decay lines do look very much like a normal distribution,
but I have a totally different theory about why it looks the way it does.
(We will have to come back to that another time.)
one point that is not clear to me is that even if I assume that you use a
similar curve to define the start-up ofthe tablet, you still have 2 more
variables to account : the height of the envelope and the width (is it
tall and thin, or short and fat ?) Is this where you use somethong like my
'condition' scale? I see that you write "The amplitude that is used is 9.8
with non-CR tablets, against 5.2 with CR tablet. Where did these numbers come
from?
I will writew again when I can - there are several points I would
like to discuss.
Regards,
--
Brian Collins <[log in to unmask]>
