S Brown1, H Thorpe, V Napp, J Brown. 1. Clinical Trials Research Unit, University of Leeds, 17 Springfield Mount, Leeds LS2 9NG, UK. s.brown@leeds.ac.uk
Abstract
AIMS: It has been recommended that closeness to death is considered when designing and reporting quality-of-life (QoL) studies in patients with advanced cancer. The aim of this work was to assess whether this finding holds when adjusting for previous QoL. MATERIALS AND METHODS: The study sample comprised 107 advanced non-small cell lung cancer patients randomised to the Big Lung Trial-QoL Study. Baseline and week 12 QoL data and patient characteristics were considered. Univariate and multivariate analysis of week 12 QoL considered baseline QoL scores, closeness to death, patient characteristics and treatment. RESULTS: Univariate analyses showed significantly different week 12 QoL according to closeness to death for nine of 15 QoL domains. Multivariate analyses showed closeness to death to be independently predictive of only three QoL domains, with performance status and baseline QoL being more predominant. CONCLUSIONS: In the present study, closeness to death was not found to be as important a factor in multivariate analysis as previously found. We recommend that previous QoL and performance status are considered in the design and analysis of QoL studies in which QoL is assessed at multiple time points in patients with advanced cancer, and that closeness to death may also be considered.
RCT Entities:
AIMS: It has been recommended that closeness to death is considered when designing and reporting quality-of-life (QoL) studies in patients with advanced cancer. The aim of this work was to assess whether this finding holds when adjusting for previous QoL. MATERIALS AND METHODS: The study sample comprised 107 advanced non-small cell lung cancerpatients randomised to the Big Lung Trial-QoL Study. Baseline and week 12 QoL data and patient characteristics were considered. Univariate and multivariate analysis of week 12 QoL considered baseline QoL scores, closeness to death, patient characteristics and treatment. RESULTS: Univariate analyses showed significantly different week 12 QoL according to closeness to death for nine of 15 QoL domains. Multivariate analyses showed closeness to death to be independently predictive of only three QoL domains, with performance status and baseline QoL being more predominant. CONCLUSIONS: In the present study, closeness to death was not found to be as important a factor in multivariate analysis as previously found. We recommend that previous QoL and performance status are considered in the design and analysis of QoL studies in which QoL is assessed at multiple time points in patients with advanced cancer, and that closeness to death may also be considered.