Does the table(s) in question have a surrogate key (i.e. identity column or something)? If so use your date query but add a statement to get anything that is greater than the highest value in your history table.

If you have to worry about updates to existing data maybe use a two step approach where you update the current records based on the PK and verify that something has changed (using binary_checksum() or something like it). Then do your inserts based on new data.

Would something like that work? or did I miss something in your post?