Performance Issue in Data Module from 2 joined Data Sets

[dlblake32]

I'm working with a client who has implemented row level security in a Data Module by using 2 Data Sets with a join, some calculated fields using macros for current user groups, etc. that returns a 1 or 0 used to filter the appropriate records. 
Performance w/o the join is reasonable but after joining the 2nd Data Set, performance has taken a huge hit, sometimes timing out or causing other errors.  When it does execute, the row level security is working as they expect but actually getting results is sporadic.
They are on Cognos 11.0.6 in the Dev environment and 11.0.5 in Production, both showing similar results.
My question, is there a way to index a Data Set to improve the performance on the join?