OBJECTIVES: NICE evaluates whether a new technology is cost-effective versus an appropriate comparator to make NHS funding decisions. They define an appropriate comparator as “established practice,” which may include branded, generic and biosimilar medicines or best supportive care (BSC), representing non-pharmacological therapy. The research presented builds on a previously described paradox, that it can be harder to make a cost-effectiveness case for innovative treatments in a disease with high unmet need and limited treatment advancements, given low-cost comparators.

METHODS: Hypothetical examples are explored, based on the mathematical relationship between key inputs in the ICER equation, to understand the additional costs that can be justified for a new intervention versus a branded, generic/biosimilar medicine or BSC. Indicative costs per annum are considered: £10,000 – £50,000, £2,000 – £5,000 and £0 respectively (mid-point base-case), and the impact of varying QALY gains (0–2, 0.2 increments) and ICER thresholds.

RESULTS: For a new intervention offering 1 QALY gain, the justifiable additional costs comparing versus a branded medicine are 1.8/2.0-fold higher than when comparing versus a generic/BSC (£30,000/QALY threshold). Perversely, this differential further increases with reduced QALY gains, which may represent diseases with limited innovation. For example, a 0.4 QALY gain (which could represent a doubling in survival where baseline survival is low), there is a 3.5-fold differential for a branded vs BSC comparator. The differentials between comparisons reduce as the threshold increases to £50,000/QALY.

CONCLUSIONS: Comparators have a marked impact on additional costs justified within cost-effectiveness analysis. This paradox may have perverse impacts on investment decisions and access to novel treatments for diseases with limited innovation, especially where there are low baseline QALYs. To reduce inequity NICE should consider additional analysis comparing interventions with a common baseline comparator to highlight potential biases. Increasing the QALY threshold could also help to mitigate this paradox.