R help - Mar 2011 - why the survival function estimate using package 'mstate' & package 'cmprsk' vary from sas and LTA (From WHO).

hello,everyone:
  I am now confused in multistate survival , when I want to poccess a 
multistate survival analysis, I turn to R and the package 'mstate'  and
packege 'cmprsk'. When I come to publishing the article.
Follow  requirement  of the magzine, the statistic is carry out in LTA package
(which was said  to be a authority in intauterine from WHO) . There we found
difference in the results(incidence ratio or survival function) computed by the
two software. I also come to the result from SAS, carry out survival analysis
for each cause seperately using prod-limit method,. the result is closed to LTA
rather than R.
why ? is the result from R better than LTA or carry out survival analysis for
each cause seperately using prod-limit method?
thanks a lot!
sincerecely
chai jian
from population and family-planing institute of Henan Province, china
 		 	   		  
	[[alternative HTML version deleted]]