Estimating costs of illness and injuries: a criticism.

Cost of illness estimates are important in health care policy making, as agreed upon, consistent measures for comparing the economic burden of various illnesses. As Hartunian, et al, say in their article' in the December 1980 issue of the Journal, "We should take economic costs into account in setting research priorities ... [and also, it is important to avoid] ... false impressions created by inexact, piecemeal measures of disease and injury effects." From this point of view, we are concerned that the Hartunian, et al, paper may itself do a disservice to the use of cost of illness measurements for policy making. We feel that it presents an unnecessary dichotomy of methodology, which may confuse policy makers. Furthermore, and perhaps most important, by employing a less than complete methodology, it creates a false impression that a change has occurred in the relative costs of illness generally accepted in the literature. Hartunian, et al, present the "incidence approach" and "prevalence approach" as two different methods of cost of illness measurement when, in fact, they are both variants of the "human capital" approach. (The principal other method in the literature is the so-called "willingness to pay" approach.) Hartunian, et al, say, "In considering preventive measures, the prevalence approach is ... basically inappropriate-just as the incidence approach is unsuitable when reduction of ongoing treatment costs or amelioration of work capacity in the prevalence population is at issue." Their statement implies there is one method for one set of problems and a different method for other problems. While not defending the "human capital" method-theoretically, philosophically, or practically-it can, when properly employed, provide reasonable comparative estimates of the potential gains from programs. The incidence approach, as described in the paper, is relevant for comparing two primary prevention programs, which would only impact the