LAUSR

Difficulties in Defining and Measuring Cancer-Related Fatigue To the Editor The recent article by Mustian et al1 shows a common lack of understanding of test theory that is pervasive in quality-of-life research. This meta-analysis was performed for something that does not objectively exist, cannot be measured, and even if both of those assertions are not accepted, cannot be compared across multiple studies. First, even simpler psychological concepts, such as cold, cannot be measured. Cold is always relative because there is no agreed-on definition of the specific objective measure (temperature) that defines “cold.” Fatigue has an even more tenuous existence. There is no physical measurement of fatigue. Each person defines for himself or herself what fatigue is and how severe it is. The same person might describe feeling fatigued after running a marathon, not sleeping well, taking an antihistamine, or having muscle weakness. And each would have a different severity scale. When a survey instrument purports to measure fatigue, it is not measuring fatigue, it is defining fatigue. And that definition exists only within that instrument and, by measurement theory, can only be validated within the specific population being tested. Also, by test measurement theory, a test cannot be “verified” by statistical correlations. It is a more complex process that requires in-depth proof of the veracity that what is purported to be measured is being measured.2(pp11-47) And space is too limited here to discuss the lack of statistical rigor involved in the method of quantification of the surveys, wherein, qualities such as good, fair, and bad are given numbers (1, 2, and 3, ...) and therefore suddenly become quantities rather than qualities. Because fatigue (or any symptom) does not concretely exist, it cannot be accurately measured. The literature on validation of health-related quality-of-life measures mistakes precision for accuracy. When one shoots arrows at clouds, all the arrows may land on the ground next to each other, representing great precision. And archers may use different bows and the arrows land in the same place, showing excellent correlation between the bows. But no arrow ever hit a cloud, and thus none had accuracy. However, by current health-related quality-of-life and symptommeasuring techniques, the arrows were validated for shooting at clouds. In conclusion, this highly scientific meta-analysis highlights the entire pseudoscience of health-related quality-oflife measurement and symptom measurement. Giving numbers to qualities that have no concrete meaning allows for measurement, but the results are empty.