ABSTRACT
In most clinical trials, investigators assess HRQoL longitudinally over the period of treatment and, in some trials, subsequent to treatment. Each assessment involves multiple scales that measure the general and disease-specific domains of HRQoL. For example, in the NSCLC study, there are three treatment arms, four assessments, and five subscales of the FACT-L. Thus, most studies entail multiple HRQoL measures that are assessed periodically. As a result, addressing the problem of multiple comparisons is one of the analytic challenges in these trials [79]. Not only are there concerns about Type I errors, but large numbers of statistical tests generally result in a confusing picture of HRQoL that is difficult to interpret [31].
