The impact of consensus conferences on physicians was studied in 1985 for the first four conferences Sweden held in 1982–1985. For each conference a main target group was defined as hospital-based physicians in supervisory positions within relevant clinics. Data were collected through a postal survey of all 1,668 members of the four different target groups. The total response rate was 86%. Reported awareness of a particular consensus conference as well as professed knowledge of a single conference statement was high—roughly 90%—for all four conferences. The consensus statement evoked changes in clinical practice according to 7%–10% of the respondents and changes were proportionally higher for physicians in smaller hospitals. Most of the respondents (61%–83%) reported no change, claiming that the consensus statement reflected clinical practices established prior to the conference. The validity of self-reported data is discussed.