There is no consensus on which methods to use to estimate an opportunity cost threshold for the efficient allocation of resources. Researchers have attempted to estimate an evidence-based threshold value, but only a few approaches have been considered and any estimate is currently used by policy makers. This study aims at exploring three assumptions normally applied in the threshold estimation: (i) approaches assume that there is always a displacement involving a loss of health; however, empirical studies suggest that one of the first responses of local health care purchasers is to squeeze greater efficiency out of providers; (ii) to be sure about the appropriate threshold it is necessary to know which health services purchasers are giving up to introduce a new treatment; current estimates bypass this lack of information by averaging the effects of changes in expenditure by clinical area; (iii) recent methodologies consider a single health outcome: mortality; however, health outcomes of many clinical areas may not be well reflected in mortality.

We propose data envelopment analysis (DEA) as a methodology that can help to address these issues by considering efficiency to measure opportunity cost per Primary Health Trust (PCT) in England and by including several outcomes in addition to mortality. This is the first time that DEA is tested in this context.

Results suggest that the majority of health locations have the possibility of decreasing their expenditures between 1 percent and 15 percent without affecting outcomes.

Estimation of the threshold should allow for observation of the actual level of inefficiencies as well as an ability to consider the previous capacity of health locations to respond to changes in expenditures. Moreover, it is crucial to select the appropriate set of health outcomes, such that they reflect health system priorities, otherwise, we would be estimating a threshold that does not reflect likely displacement.