# Re: st: absolute excess risks - test for trend

 From "Austin Nichols" To statalist@hsphsun2.harvard.edu Subject Re: st: absolute excess risks - test for trend Date Wed, 2 May 2007 08:35:50 -0400

```Raoul--
I am not familiar with this use of the acronym AER, but I think the
right approach here depends on whether E in the formula is estimated
or assumed fixed.  If E is the estimated mean of O, then you can
regress O/personyrs on age--the estimated constant will be E/personyrs
and the coef on age will be your estimate of trend in age. Like so:
. poisson o_pyrs age, r
. test age

On 5/2/07, raoul reulen <r.c.reulen@gmail.com> wrote:
```
```Hi,

I have calculated absolute excess risks of death in a cohort of people
with cancer by calculating the observed (O) and expected (E) deaths
,and using the formulae:

AER = ((O-E)/person-years) * 10,000

This gives me the excess risk per year per 10,000 persons. I have done
this for various age groups and would like to use a test for trend to
see whether there is a significant increase in the risk. How do I do
this in STATA? I probably should use an additive model, but where can
```*