Methodologies involving the use of quality-of-life patient outcomes in observational and interventional studies of health are drawn from a large and diverse field of research methods. The multidimensional way in which quality of life is conceptualized will affect the way it is measured and the complexity of the measurement. At the earliest stages of research, one must rely on methods common to the fields of tests and measurement, survey research, psychometrics and sociometrics to measure constructs that are not directly observable. Indices measuring performance can either focus on the scale's ability to perform in noninterventional, cross-sectional studies or interventional, longitudinal studies. Indices of stability, internal consistency, responsiveness with respect to true changes in quality of life, and sensitivity to treatment effects can be used to assess the scale's adequacy as a dependent variable of interest. Respondent variability can occur due to factors such as different reporters (patient, spouse, physician), the manner and form of administration (long form vs short form; self-administration vs interview) and the assessment environment (clinic, home). Finally, since quality-of-life research often involves inferential statistics and hypothesis testing, the statistical and epidemiologic principles of good study design should be followed. In addition, one should account for the reliability, responsiveness, and the sensitivity of the scale when designing the scientific hypotheses, and should specifically address the meaning of quality-of-life effect sizes by interventional-based validation. Design considerations must address the statistical issues of power, the determination of effect sizes through validation by external criteria, longitudinal data, effects of withdrawal and early termination, ceiling and floor effects, and heterogeneity of responsiveness and sensitivity among individuals. The problem of estimating quality-of-life summary parameters for use in pharmacoeconomic models is receiving increasing attention in this era of health-care reform and fiscal restraint. While medical decision theory has used cost-effectiveness models and quality-adjusted life years since the early 1970s, estimation of population parameters to differentiate among different medical interventions is relatively new. The assessment of the patient outcomes associated with medical interventions in terms of the risks, benefits and costs will clearly be a major focus of health-care reform. Development of new methodologies in quality-of-life research should build upon the strong foundation already established in the areas of clinical research, epidemiology, biostatistics, economics and behavioral science.(ABSTRACT TRUNCATED AT 400 WORDS)