I am trying to do an analysis using matching on propensity score weights. In my data, treatement is a 3-class variable: no treatment, treatment 1, treatment 2. Outcome is a 2-class variable: 0 or 1
As the number of patients in the no treatment category is very high (8000) compared to treatment 1 (900) and treatment 2 (700), I would like to use 1 to many matching
I used in R the twang package with the mnps command to generate propensity scores. I followed the twang tutorial but the analysis with svyglm seems to only use inverse probability weighting...is there a command that enables matching on the propensity score rather than IPTW?
if matching is possible, how we do carry out the 1 to many matching? (I tried MatchIt in R but this only enables matching with treatment as a 2-class variable)
thank you
