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

• Published in: Quality of life research 2016-12, Vol.25 (12), p.3209-3219
• Main Authors: Maxwell, Aimee, Özmen, Mehmet, Iezzi, Angelo, Richardson, Jeff
• Format: Article
• Language: English
• Subjects: Adult; Age; Databases, Factual - utilization; Demography; Female; Happiness; Health care economics; Humans; INSTRUMENT DEVELOPMENT; Male; Medicine; Medicine & Public Health; Men; Middle Aged; Norms; Pain; Polls & surveys; Population; Population estimates; Public Health; Quality of life; Quality of Life - psychology; Quality of Life Research; Quantitative psychology; Reproducibility of Results; Social psychology; Sociology; Standard error; Surveys and Questionnaires; Wellbeing; Women; Young Adult
• ISSN: 0962-9343; 1573-2649
• DOI: 10.1007/s11136-016-1337-z

Objectives (i) to demonstrate a method which ameliorates the problem of self-selection in the estimation of population norms from web-based data and (ii) to use the method to calculate population norms for two multi-attribute utility (MAU) instruments, the AQoL-6D and AQoL-8D, and population norms for the sub-scales from which they are constructed. Methods A web-based survey administered the AQoL-8D MAU instrument (which subsumes the AQoL-6D questionnaire), to members of the public along with the AQoL-4D which has extant population norms. Age, gender and the AQoL-4D were used as post-stratification auxiliary variables to construct weights to ameliorate the potential effects of self-selection associated with web-based surveys. The weights were used to estimate unbiased population norms. Standard errors from the weighted samples were calculated using Jackknife estimation. Results For both AQoL-6D and AQoL-8D, physical health dimensions decline significantly with age. In contrast, for the majority of the psycho-social dimensions there is a significant U-shaped profile. The net effect is a shallow U-shaped relationship between age and both the AQoL-6D and AQoL-8D utilities. This contrasts with the almost monotonie decline in the utilities derived from the AQoL-4D and SF-6D MAU instruments. Conclusions Post-stratification weights were used to ameliorate potential bias in the derivation of norms from web-based data for the AQoL-6D and AQoL-8D. The methods may be used generally to obtain norms when suitable auxiliary variables are available. The inclusion of an enlarged psycho-social component in the two instruments significantly alters the demographic profile.