Dr. Beaird asserts that health can't be valued. However, values for differing qualities of life can be and are compared, although values can vary greatly between persons or groups. Measuring quality of life is difficult, but this difficulty does not imply that meaningful, individual differences in preferences for health states do not exist, nor does it "prove wrong the assumption of cost-effectiveness analysts." Instead, it shows that policy informed by cost-effectiveness analyses must account for a wide range of utilities and the uncertainties in their measurement. As we show in our article, sildenafil is reasonably cost-effective compared to accepted medical interventions when disutility from erectile dysfunction is varied widely through the values obtained from published studies. The utility of erectile dysfunction for wives alone in Volk and colleagues' article (1) is tempered in the same article by the value given by husbands and wives jointly, 0.84. We won't speculate on whether Adam Smith would value erectile function as diamonds or as water.

Groeneveld and Duncan advocate using willingness-to-pay and cost–benefit techniques to explore sildenafil coverage decisions, citing weaknesses in utility assessment done by using time-tradeoff techniques. We argue that erectile dysfunction and other transient and recurrent illnesses impart significant chronic disutilities that allow reasonable use of time-tradeoff assessment. In addition, many health care professionals are uncomfortable with placing an explicit value on human life, as is required in cost–benefit analysis, and with the tendency of analyses using willingness-to-pay techniques to favor wealthier populations (2). Cost-effectiveness analyses do not seek to place an absolute value on health or life. Instead they compare interventions, their costs, and their resulting quality of life. We agree that insurers use many criteria other than cost to make coverage decisions. However, it is unclear whether insurers would consider cost–benefit analyses differently than they do cost-effectiveness analyses. Outside of the United States, cost-effectiveness analyses are often used to inform pharmaceutical coverage decisions (3).

Use of decision analytic techniques allows examination of worst-case scenarios for various measures to understand their impact on results (4). As stated in our paper, the values for morbidity and mortality related to sildenafil therapy used in the analysis are much higher than have been reported, illustrating that even unrealistically high values have little impact on the cost-effectiveness calculation. We agree with Siegel and Glasser that there is little evidence from clinical trials linking sildenafil to increased morbidity and mortality.