Back to the Top
Dear all,
How can I calculte the Tmax (Nonparametrics analysis) in
bioequivalnce studies with SAS?
Best regards,
Carlos
Back to the Top
Hi Carlos,
One way is to use proc twosample (for 2X2 crossover) to test for
period, sequence and treatment effects. Though this is a SAS
procedure, it is not built in and you have to purchase it from Cytel
software. It's part of StatXact procs - (statistical software for
exact nonparametric inference).
Best regards,
Nav
--
Nav Coelho, M.Sc.
Biostatistician
Biovail Contract Research
Back to the Top
The following message was posted to: PharmPK
I will describe to you probably the most inefficient method to obtain
individual tmaxs:
Say data 'pk' has 3 variables:id, time and conc.
Proc sort data=pk;
By id time;
Run;
Proc means data=pk;
Var conc;
By id;
Output out=cmaxout max=cmax;
Run;
Data cmaxtmax (drop=time conc);
Merge pk cmaxout;
By id;
If conc=cmax;
Tmax=time;
Run;
Tmaxout will now have id, cmax and tmax.
I think this should work.
Joga
Back to the Top
The following message was posted to: PharmPK
Dear Carlos,
I think SQL in SAS (Proc SQL) can be used to calculate TMax.
proc sql;
create table tmax as( select sub,time,max(conc) as CMax
from concdata
group by sub,period
having conc=max(conc));
Here in dataset-Conc data, we have variables like subject, period,
time and conc(concentration).
The above query create a table having subject, tmax and Cmax values.
Hope this helps..........
Rgds,
Vikesh S
Back to the Top
Hi SAS-users!
Carlos Gonzales wrote:
>How can I calculte the Tmax (Nonparametrics analysis) in
>bioequivalnce studies with SAS?
>
and Navdeep Coelho answered:
>One way is to use proc twosample (for 2X2 crossover) to test for
>period, sequence and treatment effects. Though this is a SAS
>procedure, it is not built in and you have to purchase it from Cytel
>software. It's part of StatXact procs - (statistical software for
>exact nonparametric inference).
>
ok, but are you *really* sure that results are correct (I mean, have you
validated it)?
At least it does not work that way in my versions of the standalone
product (StatXact v5.0.3, v.6, v6.1)
If using the ln-transformed data given in Table 3 of the 'classical'
paper by Hauschke et al. (Int.J.Clin.Pharmacol.Ther.28/2-1990 [72-78])
one gets in StatXact [Non-Parametrics > Two Independent Samples (with or
without Strata) > Hodges-Lehmann Estimates
]:
PE: +0.013521 (exp > 1.014)
90% CI: -0.150521, +0.169747 (exp > 0.860, 1.185)
whereas Hauschke et al. reported:
1.006 [0.899, 1.109]
which I could confirm by EquivTest (v2, Statistical Solutions), BIOEQV80
(v1.2, H.P. Wijnand), BIOEVALUAT (Biokinet in-house), and a manual
calculation (yes, that's possible ;-)
If I remember it correctly, Cytel's support suggested to save the
distribution, calculate the CDF, and 'manually' pick out the
corresponding values, which closest match the condition P>=0.9.
Maybe the new version 5 has improved, or the SAS-plugin behaves
differently from the standalone product?
regards
Helmut
--
Helmut Schuetz
BEBAC
Consultancy Services for Bioequivalence and Bioavailability Studies
Neubaugasse 36/11
1070 Vienna/Austria
tel/fax +43 1 2311746
http://BEBAC.at
Bioequivalence/Bioavailability Forum at http://forum.bebac.at
http://www.goldmark.org/netrants/no-word/attach.html
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)