Validation of the WHOQOL-BREF and Shorter Versions Using Rasch Analysis in Traumatic Brain Injury and Orthopedic Populations.

OBJECTIVE

To use Rasch analysis to validate the World Health Organization Quality of Life-BREF (WHOQOL-BREF) and existing short versions in individuals with traumatic brain injury and orthopedic injuries, with comparisons to a general population group.

DESIGN

The Partial Credit Rasch model was applied to evaluate the WHOQOL-BREF as well as shortened versions using a cross-sectional study design.

SETTING

Regional hospital, and national electoral sample in New Zealand.

PARTICIPANTS

Individuals with traumatic brain injury (n=74), individuals with orthopedic injuries (n=114), general population (n=140).

INTERVENTIONS

None.

MAIN OUTCOME MEASURE

WHOQOL-BREF.

RESULTS

The WHOQOL-BREF met expectations of the unidimensional Rasch model and demonstrated good reliability (person separation index [PSI] =0.82) when domain items were combined into physical-psychological, social, and environmental superitems. Analysis of shorter versions, the EUROHIS-QOL-8 and World Health Organization Quality of Life-5 (WHOQOL-5), indicated overall acceptable fit to the Rasch model and evidence of unidimensionality. The EUROHIS-QOL-8 showed good reliability (PSI=0.81); however, reliability of the WHOQOL-5 (PSI=0.68) was below acceptable standards for group comparisons, in addition to demonstrating poor person-item targeting.

CONCLUSIONS

The WHOQOL-BREF and the 8-item EUROHIS-QOL-8 version are both reliable and valid in the assessment of quality of life in both injury and general populations. Ordinal-interval conversion tables published for these validated scales as well as for the WHOQOL-5 can be used to improve precision of assessment. The transformation of ordinal scale scores into an interval measure of health-related quality of life also permits the calculation of a single summary score for the WHOQOL-BREF, which will be useful in a wide range of clinical and research contexts. Further validation work of the WHOQOL-5 is needed to ascertain its psychometric properties.