topic Re: Proc LIFETEST in SAS Procedures

Proc LIFETEST

rcleven2 — Fri, 28 Feb 2020 16:43:24 GMT

I am replicating a lifetime risk analysis that the CDC put out a few years ago regarding HIV infection among race and risk category. Currently the process uses a program called DEVCAN which was built for calculating lifetime cancer mortality. PROC LIFETEST is the only SAS procedure that is close to what I need.

Its looks as though PROC LIFETEST, for survival analysis starts with a cohort and follows that group over time. Well, I am conducting this for an entire state's population and don't have each citizen as an observation. Is there a way to calculate with LIFETEST using this data format:

AGE POPULATION RACE SEX RISK HIV_POS

1 8,000 White F None 5

2 9,000 White F None 0

3 10,000 White F None 0

85+ 13,000 White F None 7

1 8,000 Black M None 8

...Instead of having each line be an observation like i have seen in each paper and example so far?

Re: Proc LIFETEST

Reeza — Fri, 28 Feb 2020 20:55:03 GMT

Do you have dates attached to those? Part of PROC LIFETEST does require a indication of the time of when events happen and it doesn't look like your data has those components.

Re: Proc LIFETEST

rcleven2 — Fri, 28 Feb 2020 21:05:00 GMT

The date component would be met using Age as the increasing time value.

Re: Proc LIFETEST

Reeza — Fri, 28 Feb 2020 21:43:38 GMT

Then look at the FREQ statement in LIFETEST.