Willingness-to-pay (WTP) measures of the effects of changes in health on a person's welfare are more comprehensive than traditional cost-of-illness (COI) measures, but they are sometimes difficult to obtain. The authors investigated two approaches for measuring heart patients' WTP for changes in their angina symptoms. First, actual expenditures and perceived angina episodes avoided were used to infer an averting-behavior measurement of WTP. Second, a contingent-valuation approach was used to ask direct WTP questions regarding a hypothetical medical treatment that could be purchased to avoid additional angina episodes. The results indicated that although negligible COI changes were expected with small changes in angina frequency, the subjects had significant WTP to avoid increases in angina. The average WTP to avoid additional angina episodes revealed by the averting-behavior questions was comparable to the directly-elicited WTP, providing a test of the validity of the contingent-valuation approach.