Hi Ziad,
I think I had that problem a while ago. What I did was to just create another variable equal to studyid and then replace studyid with some other random studyid of your choice (not in the set of studyids you actually have) whenever visit is either -1 or 999. Then just use this temporal studyid like you do below, and get rid of the _N for those observations that have visit = -1 or 999.
I guess you could also drop all observations with visit = -1 or 999, collapse the data by studyid, and then merge your collapsed data with the original dataset.
Hope this helps.
Adrian
> Date: Thu, 28 Aug 2008 10:01:42 -0700
> From: [email protected]
> To: [email protected]
> Subject: st: help with bys
>
> Hi!
> I am dealing with cohort data for patients followed over different
> visits, before starting treatment (coded -1) then on different time
> points coded as 0, 2, 4, 8, 12, 20, 24, 36, 52, 78 and 104, also
> number 999 will be used if it was an unexpected visit to clinic.
>
> Not all patients came to all visits, so I am using -bys- command
> (pasted below) to calculate total number of visits, excluding codes -1
> and 999.
> The command below is still calculating all visits by living . value if
> visit=-1 or visit=999
>
> bys studyid: gen nvisits = _N if visit>=0 &visit
>
> Any tips?
> Thank you and best regards
> ziad
>
> *
> * 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/
_________________________________________________________________
Be the filmmaker you always wanted to be—learn how to burn a DVD with Windows®.
http://clk.atdmt.com/MRT/go/108588797/direct/01/
*
* 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/
