• Sonja Boehmer and Aleksandra Luszczynska.
• Department of Health Psychology, Free Universität of Berlin, Berlin, Germany. boehmers@zedat.fu-berlin.de
• Qual Life Res. 2006 Feb 1; 15 (1): 131-41.

AbstractTwo kinds of variables may be differentiated within instruments measuring health-related quality of life (HRQOL). Physical symptoms may be called causal variables because their occurrence can cause a change in HRQOL, but poor HRQOL need not necessarily imply that a patient suffers from these symptoms. Other items can be regarded as indicator variables. They reflect the level of HRQOL and a patient with a poor HRQOL is likely to have low scores on all indicator variables. The aim of this study was to test the relationship of the latent HRQOL construct to items and domains on the EORTC QLQ-C30. Confirmatory factor analysis was performed on EORTC QLQ-C30 data from 205 gastrointestinal cancer patients assessed one month after surgery. Functioning scales were established as variables indicating the level of HRQOL whereas all other subscales and items were established as variables causing a change in HRQOL. The model-data fit was satisfactory. Results support the hypothesis that two kinds of variables may be differentiated within HRQOL instruments. Further research should investigate whether both kinds of variables could be combined into a higher order, aggregate score for HRQOL.

24 keywords...

hide…