When quality-of-life outcomes are used to evaluate treatment effectiveness, the importance of the treatment effect relative to other clinical factors is often difficult to assess. A major methodological issue addressed in this review is the interpretation of quality-of-life treatment effects. The problem is challenging for a number of reasons, including the subjective nature of the quality-of-life construct, the indirect way which it is assessed, the multiple sources of measurement error, the heterogeneity of the stochastic properties of longitudinal changes over the full range of the scale, the complex associations among multiple outcomes, and the lack of clearly directed therapeutic goals defined in terms of quality-of-life changes. The interpretation question can be addressed at 2 levels: measurement and inference. At the first level of measurement, it is necessary to establish the relevance of the quality-of-life metric across the distribution of changes by establishing meaningful category intervals that are important to the individual patient. The second level of inference involves an evaluation of the relative benefit of a quality-of-life improvement or the risk of a quality-of-life worsening for alternative treatments in populations in whom other issues, such as overall cost and available health resources, must also be considered. This report focuses on the quantitative issues that must be addressed in an interpretation of the treatment-related changes in quality-of-life outcomes. The conceptual framework of the problem is outlined, and problems that contribute to the interpretation dilemma are discussed.