Should additional domains be added to the EQ-5D health-related quality of life instrument for community-based studies? An analytical descriptive study

Background: There is increasing interest in monitoring the health-related quality of life (HRQoL) of populations as opposed to clinical populations. The EQ-5D identifies five domains as being most able to capture the HRQoL construct. The question arises as to whether these domains are adequate within a community-based population or whether additional domains would add to the explanatory power of the instrument. Methods: As part of a community-based survey, the responses of 310 informants who reported at least one problem in one domain filled in the EQ-5D three-level version and the WHOQOL-BREF (World Health Organization Quality of Life Scale - Abbreviated version). Using the EQ-5D visual analogue scale (VAS) of rating of health as a dependent variable, the five EQ-5D and four selected WHOQOL-BREF items were entered as dummy variables in multiple regression analysis. Results: The additional domains increased the explanatory power of the model from 52 % (EQ-5D only) to 57 % (all domains). The coefficients of Self-Care and Usual Activities were not significant in any model. The most parsimonious model included the EQ-5D domains of Mobility, Pain/Discomfort, Anxiety/Depression, Concentration, and Sleep (adjusted r(2) =.57). Conclusions: The EQ-5D-3L performed well, but the addition of domains such as Concentration and Sleep increased the explanatory power. The user needs to weigh the advantage of using the EQ-5D, which allows for the calculation of a single summary index, against the use of a set of domains that are likely to be more responsive to differences in HRQoL within community living respondents. The poor predictive power of the Self-Care and Usual Activities domains within this context needs to be further examined.