Most of women with a breast cancer in sub Saharan African countries are diagnosed after the advanced stage and the survival rates reported are low.
Despite the women diagnosed after experiencing longer time interval between first symptom recognition and diagnosis, most of the literatures I come through calculated the survival rate from date of diagnosis to date of event. I feel this might affect the survival estimates.
In my data I do have both the date of first symptom recognition and date of diagnosis but I am not convinced on which time points to calculate the survival rate.
Any advice you have for me?
Regards,
The reason why survival analysis from time of first symptom is not generally done (although it can be) is that ascertainment of time of first symptom is usually less reliable than date of certain diagnosis.
I suggest you do it both ways. We did this with mesothelioma in Australia.
