gms | German Medical Science

Background: The Institute of Medicine standards for developing practice guidelines requires an explanation of the part played by values, opinion, theory, and clinical experience in deriving recommendation. Such considerations usually require obtaining consensus.

Context: Unstructured discussions between experts can rapidly attain consensus. However informal consensus processes often introduce biases from dominant personalities and groupthink. Formal consensus processes can minimize these sources of bias. However, the need for anonymous voting, facilitated discussion and statistical feedback increases resource demands on guideline developers.

Description of best practice: A professional organization developed an electronic meeting-free method for developing consensus using a modified Delphi process. Expert panel participants receive an electronic survey to obtain their opinions on multiple recommendation-related issues. Votes and comments are automatically anonymized and entered into an electronic spreadsheet. The spreadsheet is programmed to determine if consensus is attained by pre-specified rules and to generate vote distribution graphics. The program also collates comments solicited from the experts during the time of voting. If necessary, panel members re-vote on items not attaining consensus. During the re-vote, panel members receive vote distribution feedback and the anonymous comments provided by other panel members. This process is repeated until consensus is attained or a maximum of three rounds of voting. These techniques have been successfully used to develop guideline recommendations on sports concussion, and the muscular dystrophies.

Lessons for guideline developers: The advantages of a structured consensus process can be realized with minimal investment in infrastructure while obviating the need for face-to-face meetings.