PROC GENMOD/PROC GEE for repeated County-level data

[johnrizk]

Hi everyone,

I am currently running county-level data (3130 counties) to study the association between prescription rates (continuous, independent variable) and a county being labeled as "high-risk" (dependent variable, can be 1=high-risk or 0=non-high-risk).

The prescription rates  are at the county level, and so is the outcome variable "high-risk". While I'm not looking at state-level data, it's important to bear in mind that counties are clustered within states, so we need to account for clustering of counties within states. 

Based on this idea, should the "repeated" statement be for the county or the state? I am sharing 2 SAS codes below and would appreciate advise on what be more appropriate. Thank you so much!

proc genmod data=dataset descending;

CLASS State  /param=ref;

model highrisk= Opioid_Rx_Rate  /dist=binomial link=logit;

repeated subject = State/type=exch;

estimate 'Opioid_Rx_Rate' Opioid_Rx_Rate 1 / exp;

run;

quit;

OR

proc genmod data=dataset descending;

CLASS County /param=ref;

model highrisk= Opioid_Rx_Rate  /dist=binomial link=logit;

repeated subject = County /type=exch;

estimate 'Opioid_Rx_Rate' Opioid_Rx_Rate 1 / exp;

run;

quit;

[StatDave · Accepted solution]

The purpose of the SUBJECT= option in the REPEATED statement of PROC GENMOD is simply to distinguish those observations that are correlated from those that aren't. That is, it defines the clusters of correlated observations. Observations with the same value of the SUBJECT= effect belong to the same cluster and are assumed to be correlated. So, if you feel that there is correlation among all of the observations in the same state, then you should specify SUBJECT=STATE.