Last month, I wrote an article that identifies indexes that are not used in DB2 UDB. Well, IBM has made it easier in DB2 V9.7. They now have a monitor routine called MON_GET_INDEX that can help identify the indexes that are not used. Alternatively you can still use db2pd and tcbstats to find the unused indexes as well. Here is a script that will work on your DB2 V9.7 . You need to create a table (the ddl is on the script). The script will load the output to a table, so you can query the table all you want. The script can also be modified to identify the most used index and soforth.
As always do due deligence before you execute the script. Please provide your feedback and findings.