| Quality of life (QOL) is rapidly becoming an important outcome measure in randomized clinical trials of cancer treatments. The analysis of quality of life data is, however, complicated by the correlation between repeated measurements and large amount of missing assessments. Many statistical methods have been proposed to solve these problems but there is still no agreement as to what is the optimal approach of analysis.; The Quality of Life Committee of the National Cancer Institute of Canada Clinical Trials Group (NCIC-CTG) recently proposed an approach that combines cross-sectional analysis with a global test calculated from the proportion of patients whose quality of life improved or deteriorated significantly based on some clinical criteria during the course of the study. This approach seems easier for clinicians to understand but its statistical properties are not investigated.; The aim of this project is to implement CTG standard protocol and to exemplify the use of other methods, and directly compare their results by performing a comprehensive analysis on a clinical trail QOL data set. Approaches included are NCIC-CTG standard protocol, summary measures and summary statistic-based methods and model-based approaches. (Abstract shortened by UMI.)... |
