QUALITY-OF-LIFE ASSESSMENT - CAN WE KEEP IT SIMPLE

The importance of general statistical principles of study design and analysis to quality-of-life assessment in clinical trials is emphasized. Basic methods are reviewed briefly, with reference to three examples. Careful use of standard tools supplemented with context-specific scales is recommended. Problems of weighting and aggregation are discussed; the use of simple weighting schemes supplemented by sensitivity analysis is suggested. Some technical issues are explored, including factorial question structure, components of variance to distinguish mean treatment and patient-specific treatment effects and informative loss to follow-up. Simplicity of design, analysis and presentation are stressed.