An evaluation and comparison of utility measures for cost utility analysis
Date
1992
Authors
Nield, James Andrew
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Economic evaluation is gaining prominence in the health care sector as escalating health care costs are forcing policy makers to search for more efficient resource allocation. Cost utility analysis (CUA) is being hailed as a method to compare the marginal costs of a program or treatment with the marginal health improvement attributable to the program or treatment. This health improvement is measured using quality adjusted-life years (QALY's) gained as the outcome measure in the analysis. Health Status indexes (HSIs) are used to quantify the quality of life component of QALY's. This study compares two methods of generating HSIs, the standard gamble technique (SG) and the time trade-off technique (TTO), to see if they give equivalent results for hypothetical health states in the context of lung cancer treatment.
The equivalence of the SG and ITO is examined using the standard direct comparison as well as a more novel indirect comparison. The direct comparison directly compares the SG and ITO values using regression analysis to test whether they are equivalent. This is the method followed in other studies but it is potentially biased because the dependent variable is limited between O and 1. The indirect comparison compares two models relating each of the HSIs to the underlying lung cancer symptoms. The SG and ITO are equivalent if the models relating each HSI to the symptoms are equivalent. This method overcomes the problem of a limited dependent variable because if the coefficients are biased by a limited dependent variable, both models will be biased in the same manner.
Both the direct and indirect comparisons show that the SG and ITO are not equivalent. The indirect comparison also permits some insight into how the two HSIs are different. Specifically, the difference appears to be in only two of the thirty changes in symptom severity levels, when looking at changes between lung cancer health states. Therefore, in most cases, changes in lung cancer health states are valued equivalently. The other important discovery is that the SG varies over demographic groups with different risk attitudes. Males give lower SG values than females, and the females in the student sample give lower values than the females in the sample of clinic workers.
The results of this study have two important policy implications for CUA. First, SG and ITO results are not equivalent. Hence, extreme care should be taken when comparing the results of CU A studies to ensure that the same HSI is being used. Second, because values for the SG technique vary over demographic groups with different risk attitudes, demographic or cultural differences in the risk attitudes of samples used to generate SG values may bias the results of the study. Therefore, to conduct intra program comparisons using SG values, the demographic composition of the samples must be equivalent.
Description
Keywords
UN SDG 8: Decent Work and Economic Growth