On a simple quickest detection rule for health-care technology assessment

In this paper we propose a solution to the Bayesian problem of a decision maker who chooses, while observing trial evidence, an optimal stopping time at which either to invest in a newly developed health care technology or abandon research. We show how optimal stopping boundaries can be computed as a function of the observed cumulative net benefit derived from the new health care technology. At the optimal stopping time, the decision taken is optimal and the decision maker either invest or abandon the technology with consequent health benefits to patients. The model takes into account the cost of decision errors and explicitly models these in the payoff to the heath care system. The implications in terms of opportunity costs of decisions taken at sub-optimal time is discussed and put in the value of information framework. In a case study it is shown that the proposed method, when compared with traditional ones, gives substantial economic gains both in terms of QALYs and reduced trial costs.