Should the consumption of survivors be included as a cost in cost-utility analysis?

Survivor costs are those costs associated with a treatment because it extends the patient's life. A controversy exists regarding whether survivor consumption costs should be included in cost-utility analyses. The present paper uses this controversy to motivate a general reexamination of what costs to include in cost-utility analyses. Rather than the ad hoc inclusion rules currently used--a causal relationship between the intervention and the costs, and a proscription on double counting--this paper suggests three inclusion principles based on standard welfare economics. Thus, costs should be (1) included if they represent resources that directly produce the utility that is being measured in the denominator of the cost-utility ratio, (2) excluded if they represent resources that produce utility that is not being measured in the denominator, even though the costs are causally associated with the intervention, and (3) included if they represent resources consumed that are causally related to the intervention, but that have no counterveiling utility gains. These principles suggest important changes in how we account for recuperation time and unrelated medical care. They also suggest that survival consumption costs and earnings be excluded from existing cost-utility analyses.