Estimating a preference-based index from the Japanese SF-36.

Area of Science:

• Health Economics
• Psychometrics
• Cross-cultural Health Outcome Measurement

Background:

• The Short Form (SF)-36 is a widely used health survey, but a preference-based index (SF-6D) is needed for economic evaluations.
• Existing SF-6D valuation studies are primarily based on UK data, necessitating region-specific indices for accurate health outcome measurement in other populations.
• Translating and valuing the SF-6D in Japan is crucial for its application in Japanese health economic assessments.

Purpose of the Study:

• To estimate a preference-based SF-6D index from the SF-36 for the Japanese population.
• To compare the performance of Japanese SF-6D models with existing UK models.
• To identify differences in health state valuations between Japan and the UK.

Main Methods:

• Translation of the SF-6D into Japanese.
• Valuation of 249 Japanese SF-6D health states by 600 general population members using standard gamble (SG).
• Modeling of health-state values using parametric (OLS) and nonparametric Bayesian methods.

Main Results:

• Parametric models based on Japanese data showed poorer fit, more inconsistencies, and greater prediction errors compared to UK models.
• Nonparametric Bayesian models demonstrated substantial improvements in out-of-sample predictions for the Japanese SF-6D.
• The physical, role, and social dimensions exhibited larger decrements in Japanese valuations relative to the UK, unlike pain and mental health.

Conclusions:

• Significant differences exist between Japanese and UK valuations of the SF-6D.
• The Japanese SF-6D valuation dataset, particularly when estimated using nonparametric Bayesian techniques, is essential for accurate health economic evaluations in Japan.
• Region-specific SF-6D valuation data are critical for valid cross-cultural comparisons and policy decisions.