Dear Stata-listers
I am modelling the change in blood creatinine over a short time period (7
days) in a small number of patients (n=36), half of whom have been given
placebo and half a particular drug (variable "rx") after an "insult" (on the
first day of observation). The model is simple, in that there is adjustment
for baseline creatinine. I am using gllamm with random intercepts and slopes
over time. There is (statistical) reason not to use the "nocor" option. So,
in brief:
gen const=1
eq intercept: cons
eq slope: day
xi:gllamm creat basecreat i.rx*day, i( patient) eqs(intercept slope) ///
nip(8) nrf(2)
matrix a=e(b)
xi:gllamm creat basecreat i.rx*day, i( patient) eqs(intercept slope) ///
nip(20) nrf(2) from(a) adapt
The latter model seems to offer advantage and the graphics show reasonable
correspondence with a longitudinal plot of the "raw" data.
My question:
In the absence of the ability to model the "day" correlation structure (for
example AR1) with the current version of gllamm (correct me if I am wrong
here), is it possible to model the (random) slopes using, say, a spline and
if so, what would be the appropriate code (to enable each patient to have
their own (random) "slope" ("spline")).
Thanks in advance for any suggestions on this.
John moran
John Moran
Department of Intensive Care Medicine
The Queen Elizabeth Hospital
28 Woodville Road
Woodville SA 5011
Australia
Tel: 61 08 8222 6463
Fax 61 08 8222 6045
E-mail: [email protected]
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
