topic Re: PROC MIANALYZE to pool hazard ratios for continuous exposure in Statistical Procedures

PROC MIANALYZE to pool hazard ratios for continuous exposure

sseb — Mon, 04 Jan 2021 15:42:35 GMT

Hi,

I am looking to pool results after using a macro by @LinusS_ to output hazard ratios for each level of a continuous exposure. The dataset looks something like this:

_imputation_     Exposure     Estimate     StdErr
1                10           -0.3         0.4            
...              ...          ...          ...
1                60           -0.6         0.2

And so on for each imputation cycle for up to 10 imputed datasets.

I'm a novice when it comes to using PROC MIANALYZE, so I'm not sure how to use it when the dataset I'm trying to pool the data from wasn't created by a SAS PROC. Do I use parm= or data=? Any help would be greatly appreciated!

Re: PROC MIANALYZE to pool hazard ratios for continuous exposure

LinusS_ — Mon, 04 Jan 2021 19:03:52 GMT

It's probably easier to use data, I used this code recently:

proc sort data=Estimates;
	by Label _Imputation_;
run;
ods select none;
proc mianalyze data = Estimates;
	by Label;
	modeleffects Estimate;
	stderr StdErr;
	ods dataset ParameterEstimates=parms;
run;
ods select all;

Note that I use the unexponentiated estimates, since Rubin's rules requires the estimates to be approximately normally distributed, so just exponentiate the results in parms to get HR:s and CI:s.