Recognising diversity in public preferences: the use of preference sub-groups in cost-effectiveness analysis. Authors' reply