From my manual, those are 'count' and 'start' parms.
0 for count means 'change all', 1 for start means start at the beginning of the record.
CHG_OUT is described in the "External REXX Functions" in the "Functions" chapter.
Not sure which version of FileManager you have, but the books are on the IBM website.
What's really happening there is that a record is being read, then a rexx is being called for each record. the rexx is defined instream (PROC=*). The only line of the rexx is "chg_out..." That rexx function does the work of the replacing, the output record then gets copied to the output DD.
You can do a LOT of stuff with FileManager, and the more I use it, the more I learn. This concept of running a rexx on each record took me a while to understand though.