Usefulness of consensus conferences: the case of albumin.

The usefulness of consensus conferences has been questioned. We have assessed the impact of a 1989 consensus conference about the treatment of hypovolaemia on prescribing practice at our hospital. Data available from the blood bank enabled us to follow albumin use and costs by department between 1987 and 1991. Medical journals, direct mail, and meetings were used to disseminate the recommendations of the consensus conference, which were that the only indications for albumin are massive haemorrhage, plasmapheresis, massive hypoalbuminaemia, and volume expansion in pregnancy; for hypovolaemia caused by septic shock or vasoplegia, fluid gelatins and crystalloids should be used. In the year after the conference, and in subsequent years, albumin use and costs were 40% lower than before the conference, even though total medical expenditure continued to rise and numbers of patients admitted did not change. We believe that consensus conferences can be a useful means of improving medical practice. The features that ensured success in this programme were the small number of prescribers, the homogeneous setting, and the frequent restatement of the recommendations.