Re: st: stgtcalc and stset in version 5

[Enzo Coviello <enzo.coviello@alice.it>]

At 16.05 03/11/2006, you wrote:
> Dear all,
> I have a problem using stgtcalc:
> In order to test Proportional Hazards assumption in a Cox model, I would
> like to follow Patrick Royston and Peter Sasieni's suggestion
> (http://www.stata.com/meeting/4uk/nonprop.html) and fit fractional
> polynomials to the local estimates of the regression coefficients of the Cox
> model by using fracpoly after stgtcalc.
> I am using Stata9 and when I try to use stgtcalc as follows:
> ***
> stset time trtt
> stcox sex  agecateg2 agecateg3
> stgtcalc

To use -stgtcalc you should use the old syntax of 
-stset but it should not be necessary if you have 
Stata 6 or a more recent version.
Actually -stgtcalc computes scaled schoenfeld 
residuals and then -stgtplot creates a graph by 
running. So you can obtain the same results 
saving scaled Schoenfeld by using 
-scaled(newvars) in -stcox and then smoothing 
them with fracpoly or lowess (no confidence 
intervals) or running and plotting versus analysis time.

Best wishes.

Enzo





> ***
>
> I get the following error message after typing stgtcalc:
> ***
> Data are st/Stata 6 and not st/Stata 5.
> You are running the old Stata 5 version of st.
> We provide it so that researchers may reproduce old analyses.
> If you do not mean to be running the old system, type
>
>        . version 6
>
> if you do mean to be running the old system, type
>
>        . version 5
>        stset ...                       fill in the dots with old syntax)
>
> When you are done, type
>
>        . version 6
> r(119);
> ***
>
> I have tried
> ***
> version 5:stset time trtt
> version 5:stcox sex  agecateg2 agecateg3
> version 5:stgtcalc
> ***
>
> But I get exactly the same error message.
> I don't know how to stset in stata5 but in his example, P. Royston and P
> Sasieni use the same stset I have used. Is it the data I must change?
> Otherwise, is there an adaptation of stgtcalc to stata9?
> Thank you so much for your help
> Valérie
>
>
>         Valérie Jooste
>         Registre des tumeurs digestives
>         INSERM EMI 0106
>         Recherche épidémiologique et clinique en cancérologie digestive
>         Faculté de Médecine
>         BP 87900
>         21079 Dijon cedex
>         FRANCE
>         Tel : +33- 380 39 33 25
>         Fax : +33- 380 66 82 51
>
>
>
> *
> *   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/
^^^^^^^^^^^^^^^^
Enzo Coviello
Dipartimento Prevenzione ASL BAT/1
Ufficio Igiene Pubblica
via Barbera 27 - 70055 MINERVINO MURGE (BA)
Italy
tel. fax +39 0883 691053
tel (home) +39 0883 695055
mobile 347 5016016 

*
*   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/