Comparing the EQ-5D and the SF-6D Descriptive Systems to Assess Their Ceiling Effects in the US General Population


Murtuza Bharmal, Quintiles Transnational Corp., 3130 Fairview Park Drive, Suite ♯ 800, Falls Church, VA 22042, USA. E-mail:


Objectives:  The EuroQol (EQ-5D) and SF-6D (derived from the SF-12) were compared to assess any ceiling effect in the EQ-5D and the SF-6D descriptive systems. In addition, the Physical Component Summary (PCS-12), the Mental Component Summary (MCS-12) and the EuroQol Visual Analog Scale (EQ-VAS) were compared on their discriminative ability to detect differences among individuals with different morbidities and sociodemographic characteristics.

Methods:  Data from the 2000 Medical Expenditure Panel Survey were used for the analysis. A total of 11,248 individuals that were 18 years or older and had data on all the study variables were included in the analysis.

Results:  A total of 5104 individuals (47%) reported no limitations on all of the EQ-5D dimensions and only 683 (5.8%) were classified in full health based on the SF-6D descriptive system. Approximately 49% of the respondents that reported no limitations on the EQ-5D reported feeling “tense or downhearted and low,”“a little,” (level 2) or “some” (level 3) of the time on SF-6D. PCS-12 scores and EQ-VAS scores among individuals reporting no limitations on the EQ-5D descriptive system were significantly lower for respondents reporting coronary heart disease, angina, diabetes, myocardial infarction, high blood pressure or joint pain compared with respondents that reported no medical condition. Effect sizes for medical conditions using the PCS-12 were larger than the effect sizes using the EQ-VAS.

Conclusions:  Unlike the EQ-5D descriptive system, the SF-6D descriptive system derived from the SF-12 does not seem to have a ceiling effect. Nevertheless, the SF-6D does not discriminate between individuals with different morbidities who report full health on the EQ-5D, as does the PCS-12 and the EQ-VAS.