Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Area under the curve


From   Peter DeGaris <pdegaris@bigpond.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: Area under the curve
Date   Wed, 01 Oct 2008 20:46:25 +1000

Hi All,

I am currently using Stata 9.2 July 2007.

I have a rather unusual pharmacokinetic type data set that I need to run AUC's on. The time axis is both positive and negative and hence the two sections of the data need to be analysed separately (day -35 to day 0 and day 0 to day +35). The follow up times are not identical for patients either before or after day 0 (ie different time points for the data for different patients). Some patients have a missing value at time = 0.

I have split the data set to allow calculation of the AUC for the negative time period by making these negative times positive.

I want to fit a cubic spline to the data and then calculate the AUC for varying time periods (ie day - 10 to day 0 or day 0 to day +10). I have been trying to do this using the pkcollapse function.
Due to the irregular follow up times for different patients and some missing data at time=0, I need Stata to fit the spline and calculate the AUC under the spline for each patient for a consistent period say (day 0 to day +10) even if that particular patient has only data points at say day 0, 1, 4, 8, 12, 16 . ie day 10 data value would need to be predicted from the cubic spline to have the AUC calculated over day 0 to 10 and not the AUC from day 0 to 8. In addition I don't want Stata to insert a value of 0 at time=0 if that measurement is missing.
I hope that's a clear enough explanation???

Cheers

Peter DeGaris
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index