For the one-line strings, I would use the following wild card searches:
at_case_id [0-9]{1,}
at_user_id [0-9]{1,}
ca_case_ref */*/*/[0-9]{1,}/[0-9]{1,}
at_created_dt [0-9]{4}\-[0-9]{2}\-[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2}

For the one-line "at_audit" strings, I would use the following:
at_audit_type [A-Z]{1,}
at_audit_note [!^13]{1,}

For the "at_audit" strings that carry over to the next line, it would help to know which character ends the first line. Is it a paragraph end mark (a "pilcrow"), or a line break (Shift+Enter)?