Propensity scores more than one treatment

[Martin Imelda Borg]

Hi I have a column with procedure type. Which could be:
procedure type - 0 (stent)
procedure type - 1 (cardiac bypass surgery)
procedure type - 2 (venous dilation)

Now I am only interested in the main treatment (stent) compared to the control (cardiac bypass surgery). So i created a column that has values 0 - 2

To create propensity scores, the question is to evaluate if STENT procedure have a HIGHER RISK of death.
It will be using the information from procedure column.

For psmatch2 to work, this needs to be binary
0 (stent) or 1 (cardiac bypass surgery)

How do I address those patients who have had procedure 2 - I can not give them the value of 0 or 1.... as this would be the wrong treatment and thus wrong results

* Example generated by -dataex-. For more info, type help dataex
clear
input str2 hesid float admidate double RIP float(deathdate proceduretype death_30 sort_id)
"1a" 22665 1 23012 1 0 .05222337
""       . .     . . 0 .08942585
"2a" 22706 1 22719 0 1 .18564783
"3a" 22740 0     . 2 0  .8825376
"3a" 23011 1 23012 2 1  .9440776
"2a" 22705 0     . 1 0  .9457484
"1a" 22646 0     . 1 0  .9472316
"3a" 22736 0     . 1 0  .9487334
"2a" 22677 0     . 0 0  .9743183
end
format %td admidate
format %td deathdate

[George Ford]

https://www.rand.org/content/dam/rand/pubs/tools/TL100/TL170z1/RAND_TL170z1.pdf
https://www.stata.com/statalist/archive/2014-02/msg01187.html
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3710547/
https://www.hsrd.research.va.gov/for_researchers/cyber_seminars/archives/3970-notes.pdf