TESTING FOR DIFFERENCES IN MULTIPLE QUALITY-OF-LIFE DIMENSIONS - GENERATING HYPOTHESES FROM THE EXPERIENCE OF HOSPITAL STAFF

In clinical trials with a quality of life (QoL) component, it is usual to monitor several QoL dimensions at several points in time. Multiple significance tests without formal hypotheses are problematic. It is n ot always feasible to specify a priori hypotheses for all variables. C an such studies be used to generate hypotheses for testing in later re search only? We developed a method which can allow for formal hypothes is testing on a data set collected without a priori hypotheses in the protocol. We surveyed experienced physicians and nurses treating patie nts, to obtain independent expectations about differences in QoL dimen sions. These 'staff expectations' will be used in the analysis of QoL data collected from breast cancer patients taking part in three random ized trials of adjuvant therapy. We propose frameworks for the informa l and formal use of the experience of the staff in testing for group d ifferences in patients' QoL scores. The method described here is antic ipated to be useful for QoL studies in general, even when a priori hyp otheses were specified before the studies were initiated. Qual. Life R es. 7:479-486 (C) 1998 Kluwer Academic Publishers.