Struggling to contain health care costs, the United States has devoted considerable attention to health technology assessment, producing clinical guidelines and measures of appropriateness, organising consensus conferences, and carrying out studies of practice variation and “outcomes research.” Serious reservations have been expressed about the scientific validity of some of these approaches. Controversy focuses particularly on the use of routine health records to assess the effectiveness of treatments. Such studies conflict with the British tradition of using randomised controlled trials. Last year the New York Academy of Sciences brought together the exponents of different approaches to evaluating health care. The results have now been published in a book, which also provides edited accounts of the often heated discussion.1

When Congress created the Agency for Health Care Policy and Research in 1989 it was anxious for quick results, assuming that analysis of databases would provide technical solutions to the United States' health care crisis.2 Fifteen patient outcomes research teams (PORTs) with a planned budget of …