Assessing the Quality and Value of Quality-of-Life Measurement in Breast Cancer Clinical Trials

Abstract

It was more than 30 years ago that the first efforts to incorporate quality-of-life (QOL) measurements into cancer treatment trials began in studies of women with advanced breast cancer (1). In that study, the investigators used an ad hoc instrument with linear analog scales to focus on the common side effects of chemotherapy that was being given to women with advanced metastatic breast cancer, and they proposed that self-report by the patient herself would provide valuable information about toxicity and tolerability. This proposed assessment strategy paralleled the first use of multimodal chemotherapy in the setting of metastatic breast cancer, in which endocrine therapies had dominated treatment for nearly a century. At that time, women routinely received radical mastectomy as primary surgery and no one would have dreamed of the dramatic changes that ensued in early detection, surgical management, and multimodal adjuvant therapies that are now standard for women with early-stage breast cancer (2). Nor would those early QOL investigators have anticipated the explosion in the development and psychometric evaluation of various tools to evaluate QOL and patient-reported outcomes (3), as well as the routine incorporation of these assessments in a wide range of randomized clinical trials of treatment and supportive care for women with breast cancer (4). In this issue of the Journal, Lemieux et al. (5) extend a previously published systematic review of randomized clinical trials in breast cancer that included studies published through 2000 (4). In this update of studies from 2001 to 2009, which used a systematic search and reporting strategy similar to that in the previous review, these authors further our understanding of how measurement of QOL outcomes adds value to randomized clinical trials in breast cancer. During the past decade, an impressive number of trials were published (n = 190) that added to those previously reported on (n = 66), reflecting the increasing interest in the inclusion of QOL and patient-reported outcomes in clinical trials, as well as the high volume of treatment and behavioral intervention research focused on patients with breast cancer. However, despite this expansion in the number of trials, the authors identified persistent limitations in the quality of statistical methods used in these studies—with new emphasis on description of statistical power and/or sample size calculation for the QOL outcomes, documentation of missing data, and reporting of the clinical significance of the QOL findings. Inclusion of these elements of trial design and analysis has become a standardized part of QOL study design in recent years, especially with respect to labeling claims for devices