Incomplete information and irrelevant attributes in stated-preference values for health interventions

Area of Science:

• Health economics
• Behavioral economics
• Decision science

Background:

• Discrete-choice experiments (DCEs) are widely used to estimate preferences for health interventions.
• Violations of complete information (CI) and independence of irrelevant alternatives (IIA) assumptions can bias preference estimates.
• Limited evidence exists on how CI and IIA violations affect health preferences and if standard models suffice.

Purpose of the Study:

• To assess the appropriateness of CI and IIA assumptions in a DCE for long-term health problem prevention.
• To evaluate the sensitivity of preference estimates to changes in decision context and available alternatives.
• To investigate if random-parameters logit models adequately address these assumption violations.

Main Methods:

• A DCE survey was administered to a nationally representative consumer panel.
• Participants elicited preferences for interventions (surgery, medication, watchful waiting) to prevent health problems.
• Experimental variations in treatment options and decision contexts (severity, likelihood) were employed.

Main Results:

• Evidence of IIA violations was found when health prevention options changed.
• Additional substitutes decreased the monetized value of alternatives, as expected.
• Decision context moderated these effects, a novel finding.

Conclusions:

• IIA assumption violations are present in health preference DCEs, particularly when choice sets vary.
• The decision context plays a significant role in moderating preference estimates.
• Further research is needed to refine DCE methodologies for health preference elicitation.