In a 2011 article published in this journal, Baker
et al.set out to resolve a nasty dilemma for NICE by reconciling two approaches for determining whether adopting a new intervention would increase total health gains produced from available resources and hence increase system efficiency. In this response we show how the proposed reconciliation, as well as the two approaches on which it is based, fail to inform decision makers about the efficiency of a new intervention. We show how this arises from the misuse of incremental costs and effects of between-intervention comparisons as measures of changes in costs and effects associated with marginal adjustments to the scale of an intervention. Ironically, incremental data represent the choices faced by decision makers and we illustrate a method for determining unambiguously whether a new intervention represents an improvement in efficiency.