Severity of comorbidity was a better predictor of patients' current health rating, rating for angina-free health, and anticipated gain from relief of angina than was severity of angina. Since patient perceptions of a symptom may be distinct from self-reported symptom severity, treatment-outcome studies should assess patient preferences in addition to symptom severity. Comorbidity should also be measured in such studies. Having patients rate current health and symptom-free health may be a useful measure of treatment effectiveness for specific symptoms in clinical trials and patient care, and may help patients and clinicians prioritize multiple health problems.