The purpose of this paper is to argue that the emphasis in the pharmac
oeconomic literature on cost-outcomes ratios as key decision variables
in determining drug choice is misplaced. Unless strict assumptions ar
e made (and believed) as to the relationship between numbers of patien
ts treated and the marginal costs and outcomes of therapy, the only ba
sis on which an evaluation of alternative therapy options can be made
is an equilibrium-to-equilibrium modeling framework. This specifically
refers to the assumption that costs and outcomes functions exhibit co
nstant returns to scale. In this framework, given the expected changes
of distribution of patients between therapy options, estimates can be
made of net changes in overall treatment costs and outcomes.