To the Editor.
—In their critical and controversial appraisal of quality-of-life measurements, Drs Gill and Feinstein1 recommend the use of global ratings by patients, because this approach can "reflect the disparate values and preferences of individual patients." This argument is particularly important in controlled clinical trials with quality-of-life end points for two reasons. Values and preferences may not just be different among individuals but also change within individuals over time and across different situations (eg, whether they are on or off treatment),2 and values and preferences may also differ across the multiple cultural and social class groups often involved in clinical trials.3,4In a comparative clinical trial, of primary interest is the difference in quality of life between treatments. Quality of life is usually not comprehensively described; rather, treatments are compared with regard to their impact on patients' quality of life. A global rating may be sensitive
Hürny C, Bernhard J, Coates A, Peterson HF, Gelber RD. The Quality of Quality-of-Life Measurements. JAMA. 1995;273(11):843. doi:10.1001/jama.1995.03520350025013