Delphi Surveys are a method for finding out where a true consensus lies in expert opinion. They are done by getting a group of about 20 experts to complete repetitive rounds of market research type structured questionnaires about specific aspects of the question of interest. The questionnaires are completed in private so the experts are not able to influence each other in the way they often do in discussion groups or advisory boards. The results of each round of questionnaires are summarised and fed back in an anonymised form to the participants before the next round of questionnaires is made. The results of one round are used to reframe and refine the questions asked in the next round. This process is continued until statements emerge to which most of the experts agree. By convention a consensus is declared when either 75% or 80% of the experts agree with a particular statement.
The method is particularly useful for building an expert consensus in a circumstance where no definitive scientific or experimental results are available to give a firm answer to an important question. It is therefore a useful way of obtaining credible Grade IV evidence for use in the practice of Evidence Based Medicine and in Clinical Guideline Development.