It is important to identify unknown and unquantified benefits and risks of therapeutic intervention that are reflected in the quality of life. This is especially important when the benefit : risk analysis is not clearly in favour of treatment. Quality of life is defined and target areas identified. The steps in the measurements of quality of life are discussed, including the validity, repeatability and sensitivity to change of existing methods; their acceptability, analytical problems; the presentation of the results; and the use of the results in economic evaluations. The Sickness Impact Profile, the Nottingham Health Profile and the Quality of Well-Being scale are compared for the assessment of patients with angina. The advantages and disadvantages of each is discussed together with the results from a double-blind trial of treatment in angina where the Sickness Impact Profile was employed. Generic and disease-specific instruments are compared. The advantages and disadvantages of economic evaluations employing quality-of-life outcomes are discussed. Methods should be employed that withstand rigorous scientific evaluation. Both a health profile and a summary statistic such as a Health Status Index should be measured. If different treatments are to be compared, a randomised controlled trial should be employed. If the more expensive treatment is likely to be superior then costs should be collected and a cost-utility analysis performed.