Related Experiment Videos
Inconsistent responses in three preference-elicitation methods for health states.
1Institut Universitari de Salut Pública de Catalunya, Universitat de Barcelona, Spain. xbadia@bell.ub.es
Social Science & Medicine (1982)
|September 1, 1999
Summary
This study found that while the time trade-off (TTO) method generated more inconsistent health state values, overall inconsistencies were low and did not impact the final rankings. Older and less educated individuals produced more inconsistencies, highlighting a need for targeted improvements in preference elicitation methods.
Area of Science:
- Health Economics
- Psychometrics
- Decision Science
Background:
- Cost-utility analysis, using Quality Adjusted Life Years (QALYs), relies on general population health state values for resource allocation.
- Preference-elicitation methods like the visual analogue scale (VAS), time trade-off (TTO), and standard gamble are used but can yield inconsistent responses.
- Understanding and minimizing response inconsistencies is crucial for reliable health state valuation.
Purpose of the Study:
- To assess the degree of inconsistent responses from ranking, VAS, and TTO methods for health state valuation.
- To investigate the influence of sociodemographic and health variables on response inconsistencies.
- To evaluate the impact of inconsistencies on the ordering of health states in EQ-5D tariffs.
Main Methods:
- Examined inconsistencies (internal and criterion) in health state valuations using ranking, VAS, and TTO methods.
- Assessed a subset of 43 EuroQol-5D (EQ-5D) health states for a Spanish tariff.
- Analyzed the number and 'size' of inconsistencies, and their relationship with respondent characteristics and interviewers.
Main Results:
- The TTO method produced more inconsistent responses than VAS and ranking, but overall inconsistencies were low across all methods.
- Inconsistencies did not significantly affect the final ordering of health states in the EQ-5D VAS and TTO tariffs.
- Older age and lower educational attainment were associated with higher numbers of inconsistencies; health characteristics were not significant.
- Inconsistencies doubled for ranking and VAS methods between interviewers.
Conclusions:
- While TTO yields more inconsistencies, current levels are generally low and do not alter health state rankings in EQ-5D tariffs.
- Targeted efforts to reduce inconsistencies are needed, especially for older and less educated populations.
- Future research should explore how inconsistencies influence the actual values assigned to health states, beyond just their order.