It sounds like there is a possibility that others might be identifiable? Do you trust in your friend to fix that? Could you email the IRB anonymously? If edits are done that remove the possibility of identification, maybe I wouldn't say anything knowing that you could always bring it up later when the initial feelings subside.

If there was a chance the paper would be published despite the possibility of identifying people, I'd feel compelled to tell T. He could, however, be open about whatever it is and not be concerned - or - it could be very awkward. But how would you know.