topic How to generate estimated probability of surviving beyond age t given survival to age 65 in Statistical Procedures

How to generate estimated probability of surviving beyond age t given survival to age 65

JWH1 — Mon, 25 Nov 2019 15:00:59 GMT

Hi,

I have a dataset of pensioners that is left-truncated and right-censored. I would like to compare (for males vs females) the estimated probability of surviving to e.g. age 70 conditioned on surviving to age 65. I also have a variable indicating last known annual pension amount that I'd like to use to adjust for socioeconomic status. The dataset includes pensioners alive as of 1/1/2010 and mortality was assessed during the period 2010-2016.

data p;

input BirthYr PensionYr DeathYr Pension_amt Gender Age2010 Dead;

datalines;

1921 1948 2014 84400 0 89 1
1927 1950 2014 80000 1 83 1
1962 1984 2016 54700 1 48 0
1951 1986 2016 76500 1 59 0
1923 1957 2016 14700 1 87 1
1940 1960 2016 25200 1 70 1
1939 1965 2011 51500 0 71 1
1947 1974 2011 10000 1 63 1
1944 1979 2010 79000 0 66 1
1918 1947 2010 36100 1 92 1

Would appreciate any suggestions..

Re: How to generate estimated probability of surviving beyond age t given survival to age 65

mkeintz — Mon, 25 Nov 2019 15:15:15 GMT

Look up SAS examples for PROC LIFETEST. I believe it will nicely provide the hazards you are asking for.

Re: How to generate estimated probability of surviving beyond age t given survival to age 65

JWH1 — Mon, 25 Nov 2019 15:32:47 GMT

thank you...I don't believe I can use PROC LIFETEST for left-truncated data.

Re: How to generate estimated probability of surviving beyond age t given survival to age 65

mkeintz — Mon, 25 Nov 2019 16:23:49 GMT

Am I correct in assuming that "conditioned on surviving to age 65" is what you mean by left-censored?

If so then "the estimated probability of surviving to e.g. age 70 conditioned on surviving to age 65" is, iirc, is the very definition of the 5-year hazard rate (or actually 1 minus the 5-year hazard rate) for 65 to 70 period, which I'm guessing can be produced by LIFETEST. And even if LIFETEST doesn't produce something literally called hazard-rate, isn't what you want equivalent to the probability of surviving to 70 divided by the probability of surviving to 65? Those values would certainly be generated by LIFETEST, no?

It's been a long time since I was exposed to survival analysis, so I admit to being a bit insecure about these statements.