Measurement of health outcomes in treatment effectiveness evaluations: conceptual and methodological challenges.

Major challenges in the evaluation of the "end results" of health services include ensuring that concepts are correctly defined and measured, that the validity of measures used in different applications and populations is well documented, and that observed effects can be clearly interpreted. Health status is the most widely interpretable concept to apply in the context of health services. Quality of life connotes inclusion of the environment outside the context of the person and of health care and may or may not be health related, depending on the evaluation context and the impact of disease and treatment. All concepts and constructs must be defined in reference to their theoretical origin and to a model of relationships among different concepts. Modern test theory offers the potential for individualized, comparable assessments and for the careful examination and application of different measurement models. Selection and critique of measures should be based on the intended application and accumulated evidence for that application. Thus, there are no valid instruments per se. Validity in use, including responsiveness, interpretation of effects, and generalizability to diverse populations, is the most important measurement characteristic for treatment effectiveness. An evaluation of the validity of preference-based measures is particularly important for the interpretation and comparability of outcomes in cost-effectiveness evaluations. The successful translation of research into policy and practice is limited by the extent to which these critical issues are addressed in actual treatment evaluations.