The use of cost-effectiveness by the National Institute for Health and Clinical Excellence (NICE): no(t yet an) exemplar of a deliberative process.

Democratic societies find it difficult to reach consensus concerning principles for healthcare distribution in the face of resource constraints. At the same time the need for legitimacy of allocation decisions has been recognised. Against this background, the National Institute for Health and Clinical Excellence (NICE) aspires to meet the principles of procedural justice, specifically the conditions of accountability for reasonableness as espoused by Daniels and Sabin, that is, publicity, relevance, revisions and appeal, and enforcement. Although NICE has adopted a highly standardised approach and continuously publishes key documents on its website, its technology appraisal programme does not fulfil the publicity condition of accountability for reasonableness. Economic models are not made sufficiently transparent to enable public scrutiny, and decision criteria other than cost-effectiveness remain enigmatic. NICE's reliance on cost-utility analysis and "plausible" cost-per-quality-adjusted life year (QALY) benchmarks further raises serious issues with regard to the relevance condition of accountability for reasonableness. This is illustrated by counterintuitive cost-per-QALY rankings that are difficult to justify using reflective equilibrium methods, and by the current debate surrounding expensive therapies for rare diseases ("orphan" treatments). In addition, an excessive focus on QALYs may stand in the way of exploiting the best available effectiveness evidence. The NICE mechanism for revision and appeals is also more restrictive than provided in accountability for reasonableness. As to the enforcement condition, no effective quality assurance processes are in place for technology assessments, and implementation of guidance remains imperfect. NICE, despite impressive efforts, appears to have a long way to go before meeting the conditions of accountability for reasonableness.