Health preference instruments in a breast cancer population: Is there a ceiling effect when compared to the quality-of-life instruments?

Abstract

6597 Background: Determination of health preference is a common outcome measure used in economic evaluations and represents the societal valuation of the patient's health state. It can be measured with generic multi-attribute instruments. Multi-attribute health preference (HP) instruments have similar domains to oncology quality of life (QOL) instruments. It is possible that HP instruments may over or under estimate a patient's QOL. Methods: Consecutive breast cancer patients irrespective of stage and treatment attending an outpatient clinic were approached to enrol in a work productivity and QOL study. The Functional Assessment of Cancer Therapy- General (FACT-G) / -Breast (FACT-B), and the EQ-5D were used to assess QOL and HP, respectively. Total scores are 108 (FACT-G) and 148 (FACT-B). EQ-5D is a generic, multi-attribute instrument measuring health preference from 0 to 1 for 243 possible health states. The analysis was carried out post-hoc to investigate the ceiling effect with the EQ-5D and FACT. Results: A total of 151 patients with breast cancer participated in the study. Mean±SD age was 55.2±12.5 years and time since diagnosis ranged from 0.1 to 25.0 years. Mean±SD FACT-G and FACT-B scores were 79.3±16.6 and 105.2±21.6, respectively. A perfect score was observed in only one patient for the FACT-G, none for the FACT-B, and in 35 (23.2%) patients for the EQ-5D. Four EQ-5D health states accounted for half the sample, with perfect health being the most common. Mean±SD utility score was 0.80±0.16. A correlation test between the EQ-5D and both FACT scores demonstrated convergent validity (p<0.01). Conclusions: Ceiling effect was observed with the EQ-5D which has fewer questions and a smaller scale compared to the FACT. This may indicate that the EQ-5D is not as sensitive in a general breast cancer population. However, the mean utility score was comparable to other studies in women with breast cancer. Future studies should investigate the relationship between long term QOL and HP in a larger cohort of women with breast cancer taking into account disease stage and treatment strategy. No significant financial relationships to disclose.