Just prior to the OMERACT conference, participants completed a questionnaire that solicited explicit opinions on the issues discussed at the conference. The response rate was 77%. To determine the minimum level of important difference in a clinical trial comparing 2 active drugs, participants were asked to think of each of 6 separate measures in turn as designated primary outcome measure. In this situation, to decide that an important difference between the 2 groups was present, participants required a median of 20% difference in painful joints, swollen joints, and in disability, 30% in pain and patient global assessment, and 40% in physician global assessment. On each measure, between 3 to 12% of participants felt they could not decide on an important difference in that situation. Similar questions were asked for the minimum important improvement in a patient; required levels of improvement were similar but slightly higher than the responses given for trials, and more participants felt they could not decide. Correspondents indicated that acute phase reactants are also very important for assessing minimum levels of important difference and improvement in trials and patients. A large majority was in favor of applying an index of aggregated outcome measures if sensible and valid: 72% in patients, and 93% in trials.