Selection and evaluation of empirical research in technology assessment.

Technology assessment involves application of the scientific method to the practice of medicine. Finding all of the assessment reports in a given field is not an easy task. Proper evaluation of those assessments requires the conduct of a prospective experiment in which the sources and results are blinded when the choice is made of papers to exclude and to include, and the process should be carried out in duplicate. There are several available data bases for carrying out the search, but because of indexing problems they should be supplemented by reference to the bibliographies of pertinent published articles. Clinical trials included in meta-analyses should be graded by quality and thus facilitate sensitivity analyses. Attention must be paid to the possibility of publication bias. Finally, the advent of meta-analysis makes it desirable to begin randomized controlled trials in areas of uncertainty, even when there is no possibility that individual investigators will encounter enough patients to draw valid conclusions.