Context effects in health state utility assessment: Etiology, framing, and delay of health outcomes

This research examined the effects of two aspects of health state context, etiology and delay in onset, on preferences for health states. Research questions about the effect of etiology were derived from regret theory and prospect theory, theories of behavioral decision making: (1) Does etiology (iatrogenic versus natural) affect utility for a health state? (2) Does framing iatrogenic morbidity as a transaction cost reduce its disutility? Research questions about delayed health outcomes were based on the theory of intertemporal choice: (1) Does time preference for health differ for health gains and losses? (2) Does the time preference difference between gains and losses apply to both short and long-term changes in health? Each contextual factor was investigated in a separate mixed factorial experimental design. Subjects were randomly assigned to rate three hypothetical health states described in one of three frames: unexplained loss of health, iatrogenic loss of health, or iatrogenic cost of overall health gain, and to evaluate either health gain or loss profiles in which the duration and delay of onset of the health change were manipulated. One hundred and eight volunteers were obtained from a university staff population, and outpatient and volunteer populations of a Veterans Affairs hospital. Data for both studies were collected in one hour interviews. Utilities for health states and health profiles were assessed with the standard gamble and the category rating scale respectively. Inferential tests of hypotheses were based on mixed factorial analyses of variance. In the framing study hypotheses were tested with planned comparison t tests; in the time preference study they were tested with F tests of double and triple interactions. Conclusions were that greater disutility is assigned to iatrogenic morbidity than to natural morbidity, that time preference for health gains differs from time preference for losses, and that devaluation of a delayed health loss is dependent on its duration. The theory of intertemporal choice was supported in the health domain. Conclusions cannot be extended beyond preferences elicited with hypothetical health scenarios. Lack of support for a framing effect for iatrogenic health states was attributed to inadequate statistical power