Propensity score matching

GS2 — Wed, 16 Aug 2023 13:28:52 GMT

Using SAS 9.4

I have a data set with 299 obs. I want to propensity match the data set between 2 groups. Group 1 has 98 obs and group 2 has 201 obs. I would like to propensity match based on procedure_type (exact), age, sex, BMI, size. What would be the best way to get propensity match and get data set output? Below is my attempt but I am not sure this is correct. Thank you

proc logistic data=lw.analysis;
class procedure_type;
model group = procedure_type age bmi size / link=glogit rsquare;
output out = ps_los pred = ps xbeta=logit_ps;
run;

Re: Propensity score matching

ballardw — Wed, 16 Aug 2023 14:47:34 GMT

I suggest looking at the examples for Proc PSMATCH and the documentation as a good starting place. https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.4/statug/statug_psmatch_syntax.htm

The example in the Getting Started tab shows code to create a data set that uses Age, Gender and BMI to create matches for a Drug (the role of your procedure_type I think).

Steps would be to create a score, similar that in the example that incorporates "size" as well.

topic Propensity score matching in SAS Programming

Propensity score matching

Re: Propensity score matching