STATISTICAL METHODS FOR QUALITY-OF-LIFE OUTCOME RESEARCH

生活质量对患者预后的影响越来越大

Quality-of-life patient outcomes measures are playing an increasingly important role in determining the effectiveness and net benefit of new and existing pharmaceutical therapies. The interpretation of quality-of- life data for therapeutic decision making and policy planning requires a) valid assumptions concerning measurement and statistical properties of the quality-of-life scales used in assessment, b) representative analytical models relating patient-specific quality-of-life outcomes to effectiveness, utility, costs and overall net benefits and 3) appropriate inferential statistical methodology to test therapeutic effects and differences. While a large body of literature exists debating the meaning, usefulness and validity of various models, instruments and definitions of quality of life, little research has focused upon those issues impacting on the analysis and interpretation of quality-of-life outcome data used in determining the relative risks and benefits of alternative pharmaceutical regimens. The major objectives of this proposal are a) to evaluate and analyze current and potential measurement and statistical techniques used in quality-of-life related patient outcomes through an exhaustive review of the therapeutic and statistical literature, b) to refine existing analytical and statistical methodology so that it is appropriate for therapeutic science and the evaluation of pharmacologic therapies and c) to develop data analysis, demonstration projects within three major quality-of-life analytical areas by applying the selected methods to existing quality-of-life clinical trials databases of hypertension (n = 2,500) diabetes (n = 300) and HIV (n = 1000). All these databases contain extensive prospective data on quality of life during randomized multicenter clinical trials of alternative drug therapy. The three-step methodology of model review, development and demonstration will be employed in the following major areas: 1) reliability, responsiveness, validity and sensitivity of health-related quality-of-life measures; 2) multivariate, longitudinal models (both parametric and nonparametric) to assess quality-of-life treatment effects in randomized clinical trials and 3) global models for summary estimates for use in quality-adjusted survival, cost-effectiveness, cost-utility and health state transition Markov models . The final product of this research will be a comprehensive compendium of quality-of-life analytical models, appropriate statistical methods suitable for quality-of-life related therapeutic outcomes research, and practical data analysis demonstration projects useful to clinical investigators, health policy planners and regulatory agencies for designing, evaluating and interpreting quality-of-life studies of pharmacologic therapies.