We developed a set of preliminary response criteria for use in future clinical trials in fibromyalgia syndrome. We determined outcome measures from a previously reported clinical trial which best distinguished patients treated with effective medication from those treated with placebo or ineffective medication, using stepwise logistic regression analysis. Several combinations of outcome measures were identified and plotted in the form of receiver operating characteristic (ROC) curves. The combination of variables possessing the greatest area under the ROC curve included (1) physician global assessment score less than or equal to 4 (0 = extremely well, 10 = extremely poorly), (2) patient sleep score less than or equal to 6 (0 = sleeping extremely well, 10 = sleeping extremely poorly), and (3) tender point score less than or equal to 14 (maximum possible tender point score equalled 20). These criteria accurately distinguished those treated with effective drug from those treated with placebo when tested in an unreported therapeutic trial of cyclobenzaprine. The criteria identified 11 of 14 patients in the amitriptyline trial and 4 of 6 patients in the cyclobenzaprine trial who attained improvement measured independently. The methodology used to define these preliminary criteria may be applied to refine the criteria as additional sensitive and clinically relevant outcomes are developed.