Optimal sequential sampling rules for the economic evaluation of health technologies

Referring to the literature on optimal stopping under sequential sampling developed by Chernoff and collaborators, we solve a dynamic model of the economic evaluation of a new health technology, deriving optimal rules for technology adoption, research abandonment and continuation as functions of sample size. The model extends the existing literature to the case where an adoption decision can be deferred and involves a degree of irreversibility. We explore the model's applicability in a case study of the economic evaluation of Drug Eluting Stents (DES), deriving dynamic adoption and abandonment thresholds which are a function of the model's economic parameters. A key result is that referring to a single cost-effectiveness threshold may be sub-optimal.