Longitudinal investigation of the stable and dynamic components of the World Health Organization Quality of Life Measure (WHOQOL-BREF) using generalizability theory

The WHOQOL-BREF is a widely used 26-item self-report measure of Quality of Life (QOL), but there is lack of research investigating to what extent it measures dynamic (state) and stable (trait) aspects of QOL. Generalizability Theory is the most suitable method to examine the state-trait distinction and to evaluate the overall reliability and generalizability of the WHOQOL-BREF. Generalizability theory was applied to a longitudinal study dataset of 130 participants who had completed the 26-item WHOQOL-BREF at three time points spaced 6 months apart. The results indicated that the total 26-item WHOQOL-BREF has strong reliability in measuring enduring QOL with a G coefficient of 0.90, confirming that the scores are generalizable across persons and occasions. The results also indicate that the social domain subscale of the WHOQOL-BREF has marginally acceptable reliability in measuring trait, with a G-coefficient of 0.77. However, the remaining three domain subscales, physical, psychological and environmental, and the two short versions, WHOQOL-5 and EUROHIS- QOL 8-item index were found to be less reliable in measuring either state or trait QOL with G-coefficients below 0.65 and SCI below 0.15. The ability to distinguish to what extent items of each subscale of QOL are measuring state or trait will advance QOL research by identifying which QOL aspects are likely to change in one's subjective QOL if targeted by interventions. Using the WHOQOL-BREF total score permits evaluation of long-lasting QOL changes resulting from interventions and/or environmental impact.