Crosswalk algorithms for the conversion of mean EQ-5D indices calculated with different value sets

Area of Science:

• Health Economics
• Psychometrics
• Quality of Life Research

Background:

• The EQ-5D questionnaire is vital for health research and quality registers.
• Sweden previously lacked a country-specific EQ-5D value set, often using the UK TTO values.
• This necessitated the development of a method for converting between value sets.

Purpose of the Study:

• To create a user-friendly tool for bidirectional crosswalking of EQ-5D mean values.
• To establish conversion algorithms between the UK time trade-off (UK TTO) and Swedish TTO value sets.
• To enable more accurate comparisons of EQ-5D health states across different value sets.

Main Methods:

• An artificial dataset covering all EQ-5D outcomes was generated.
• EQ-5D indices were calculated using both UK TTO and Swedish TTO value sets.
• Least-squares and major axis regression models were employed to establish crosswalk algorithms, followed by simulations.

Main Results:

• Major axis regression proved more effective than ordinary least-squares regression for crosswalking.
• Conversion from the UK TTO scale to the Swedish TTO scale was more accurate.
• Crosswalking EQ-5D values near the index's floor and ceiling presented greater challenges.

Conclusions:

• Algorithms based on major axis regression were successfully developed for crosswalking EQ-5D values between UK TTO and Swedish TTO sets.
• The developed algorithm aids in comparing EQ-5D values when only mean values are accessible.
• This crosswalk algorithm simplifies the transition from using the UK TTO to the Swedish TTO value set.