Instead of doing a de novo model in which you're replicating an earlier proceeding, effectively, with leave, why wouldn't you simply focus your efforts on making sure that the original hearing is robust and that it protects fully the ability of claimants to make their case?
