Making judgements about treatment effectiveness based on health outcomes: theoretical and practical issues

This article considers the problem of deciding which health care outcomes are important and relevant for (1) developing management recommendations for clinical practice guidelines and (2) evaluating patients' responses to treatment.

The Heart Failure Guideline Panel sponsored by the Agency for Health Care Policy and Research (AHCPR) decided that for both purposes the relevant outcomes are those experienced directly by patients: mortality and health-related quality of life (HRQOL). Changes in intermediate outcomes, such as test results of various kinds, were deemed insufficient evidence of effectiveness.

In the context of heart failure, mortality risk (prognosis) can be measured using a variety of biochemical and physiological variables, but changes in these variables do not appear to correspond to changes in prognosis. For this reason, the Heart Failure Guideline Panel recommended that patients' responses to treatment be guided by signs and symptoms, rather than test results (for example, echocardiographic measurement of left-ventricular function or exercise-tolerance testing). HRQOL is best assessed by direct patient self-reports. Although patients may be influenced by a host of other variables (for example, mood, adaptation to chronic disease, placebo effect), self-reports will probably always represent the "gold standard" in assessing HRQOL. The reliability and validity of these reports can be enhanced by using standardized instruments or by incorporating questions from such instruments into the history-taking aspect of patient evaluation and monitoring. Finally, physical examination and submaximal exercise testing can provide additional information that can supplement patient reports. Information from these sources must be evaluated carefully in light of patients' self-reported HRQOL.