Deriving population norms for the AQoL-6D and AQoL-8D multi-attribute utility instruments from web-based data

Area of Science:

• Health Economics
• Psychometrics
• Survey Methodology

Background:

• Estimating population norms from web-based surveys is challenging due to self-selection bias.
• Existing multi-attribute utility (MAU) instruments like the AQoL-4D and SF-6D have limitations in capturing psycho-social health dimensions.
• There is a need for robust methods to derive representative population norms for newer MAU instruments.

Purpose of the Study:

• To demonstrate a statistical method for mitigating self-selection bias in web-based survey data.
• To calculate population norms for the Assessment of Quality of Life (AQoL)-6D and AQoL-8D multi-attribute utility instruments.
• To establish population norms for the sub-scales of the AQoL-6D and AQoL-8D.

Main Methods:

• A web-based survey collected data on the AQoL-8D (which includes the AQoL-6D) and AQoL-4D.
• Post-stratification weighting, using age, gender, and AQoL-4D scores as auxiliary variables, was employed to correct for self-selection bias.
• Jackknife estimation was used to calculate standard errors for the weighted samples.

Main Results:

• Physical health dimensions in both AQoL-6D and AQoL-8D showed a significant decline with age.
• Psycho-social dimensions exhibited a significant U-shaped relationship with age for the majority of dimensions.
• The overall utility scores for AQoL-6D and AQoL-8D demonstrated a shallow U-shaped relationship with age, contrasting with monotonic declines in AQoL-4D and SF-6D.

Conclusions:

• Post-stratification weighting effectively reduced bias in deriving population norms for AQoL-6D and AQoL-8D from web-based data.
• The developed methodology is applicable for generating population norms when suitable auxiliary variables are available.
• The inclusion of expanded psycho-social components in AQoL-6D and AQoL-8D significantly influences their demographic profiles compared to older instruments.