JointPoint / test of secular trends

[JN Nielsen]

Dear all,
I've calcuted incidence rates (using Poisson with offset[logtime] and 1-year age bands and predicitive margins) at four different time points.
Now, I would like to test for trends over time. My plan was to use the JointPoint software, so I created a dataset with all the incidence rates, SE and 95%CI by year, sex, race and age groups. But I can't figure out how to format the dataset to be used in JointPoint - even after downloading two additional softwares (SEERSTAT SEERPREP).
Would anyone in here know how to export data from Stata such that I can run the test in JointPoint?

Kind regards
Jannie

[William Lisowski]

The Joinpoint User's Guide PDF I found at

https://surveillance.cancer.gov/join...lp_4.7.0.0.pdf

explains on page 6 what the format of the input should be. It appears that Joinpoint requires a delimited text file, so once you have a Stata dataset with the required variables sorted into the correct order, it seem that Stata's export delimited command will do what you need, and since all your data is numeric, preparing a csv file with comma as the delimiter should suffice. I will add that for this sort of thing, I generally use the menu-driven interface from Stata's File menu rather than try to work out the options on my own.

[JN Nielsen]

Thanks! Will check it out now!

[JN Nielsen]

Hi again.
It was easy to get the data into Joint Point in the CSV format! Thanks for helping out.