Programming the statistical procedures from SAS

proc genmod

Reply
User
Posts: 1

proc genmod

I used “proc genmod” to calculate mean difference of medical cost between two groups (BC1 and BC2), and this dataset was from survey data.

SAS code and results as follows:

proc genmod data =total;
class  BC_G
AGE42X_G  
RACEX_G
EDUCYR_G 
MARRY08X_G  
Com_mobility_G
model TOTEXP_PHCE = 
BC_G
AGE42X_G  
RACEX_G
EDUCYR_G 
MARRY08X_G  
Com_mobility_G
/dist=gamma link=log type3;
lsmeans BC_G /diff CL ilink;
Run;
Spoiler
Spoiler
Spoiler

BC_G Least Squares Means

BC_G

Estimate

Standard Error

z Value

Pr > |z|

Alpha

Lower

Upper

Mean

Standard Error of Mean

Lower Mean

Upper Mean

1

8.9551

0.07159

125.10

<.0001

0.05

8.8148

9.0954

7747.48

554.60

6733.28

8914.43

2

8.3905

0.01140

736.08

<.0001

0.05

8.3682

8.4128

4404.99

50.2118

4307.67

4504.51

 

Differences of BC_G Least Squares Means

BC_G

_BC_G

Estimate

Standard Error

z Value

Pr > |z|

Alpha

Lower

Upper

1

2

0.5646

0.07110

7.94

<.0001

0.05

0.4253

0.7040

My question are how to show CI of mean difference (anti-Log), and how to combine “proc surveymeans” into that model.

 

Proc surveymeans;
stratum varstr;
cluster VARPSU;
weight perwt08f;
var TOTEXP08_PHCE;
Run;
Ask a Question
Discussion stats
  • 0 replies
  • 214 views
  • 0 likes
  • 1 in conversation