topic Re: calculating total time at risk for denominator for incidence rate for males n females separately in SAS Procedures

calculating total time at risk for denominator for incidence rate for males n females separately

KH1 — Mon, 02 Nov 2015 18:17:50 GMT

hi,

I need to calculate total time at risk for males and females separately for incidence rate denominator.

I have a variables 'time_major' in days which denotes time in days from T0 to outcome/death/loss to follow up.

What I dont know is how to get the total time_major for each sex and use it to get incidence rate per 100 person years.

The reason is that thhe sum of the total time for males and females cannot exceed the overall time_major.

thanks

Re: calculating total time at risk for denominator for incidence rate for males n females separately

ballardw — Mon, 02 Nov 2015 18:25:09 GMT

I would help to provide some example data, fake is fine, that illustrates your starting point and what the desired outcome would be. The example data would be best as a data step we can execute and should provide examples of all the behaviors you expect in the data.

Re: calculating total time at risk for denominator for incidence rate for males n females separately

KH1 — Tue, 03 Nov 2015 15:57:16 GMT

thanks!

data b; set a;

input sex $ eventtype time;

datalines;

m 1 3000

f 1 3996

f 1 285

m 2 4000

m 3 893

m 2 1000

m 2 3952

f 1 123

f 1 964

m 1 478

m 2 2789

;

run;

note: eventtype: 1=contralateral major amputation (MAIN outcome of interest); 2=death; 3=migration from country

time is time to event (contralaterl major amputation/death/migration) from T0 (30 days after surviving a unilateral major amputation) in days

I want to calculate incidence of contralateral major amputation per 100 person years among males and females separately.

thank you

Re: calculating total time at risk for denominator for incidence rate for males n females separately

KH1 — Tue, 03 Nov 2015 16:16:25 GMT

What I would like to know is if the total time in the incidence rate denominator is the sum of all times or the longest follow up time.

just need to know how to calculate the denominator for incidence rate.

I did this:

proc sort; by sex;

proc print;

var time;

sum time;

by sex;

run;

and it added up time ind ays for all females and males separately. is this correct?

thanks