DISCUSSION
Guidelines are intended to review systematically the available evidence, rate its validity, and distill the findings into easy to understand recommendations that practitioners can follow. Many different methodologies have been developed for this purpose21,174,175 The GRADE method (Grading of Recommendations Assessment, Development and Evaluation)175, has been recently adopted by a number of professional organizations for the development of guidelines, but we did not choose it because it relies heavily on RCTs and systematic observational studies that are not available in the literature on EA-TEF.
We developed our recommendations on the basis of the Rand Appropriateness Method24–26, offered certain important advantages for the type of evidence we had available such as: a) the responders do not have to accept positions they do not agree with, in order to reach a consensus; b) disagreements are not suppressed, thus reflecting more accurately the uncertainties and variability that exist in the actual clinical practice; c) each item is rated by all responders, and their responses carry the same weight (thus preventing a single issue (or a single person) from dominating the process; d) the anonymity of the responses allows the participants to express their opinions more freely; and e) it provides timely feedback by communicating the results of each round to the responders, allowing them to compare their responses to others and potentially adjust them.
Our recommendations were based on the strength of the agreement among the RCWG members. Thus, Level A recommendations refer to unanimous adoption of specific interventions and care practices. Level B & C recommendations reflect both the lack of evidence and actual controversies on a specific topic, as well as the different areas of expertise, skill mix, resource availability and hospital culture among the RCWG’s representative institutions. This differs from other consensus based methods that assign different weight on the evidence depending on the design of the study, its size, the type of analysis etc.)176.
As diagnostic methods improve and therapeutic interventions become more precise, it is likely that many of the current recommendations will have to be revised. In the meantime, the document provides a common approach for the evaluation and management of these complex patients at levels that can be provided by most practitioners and institutions. It may also serve as the basis for multicenter collaborations that will eventually provide a higher level of evidence.
ACKNOWLEDGMENTS
The authors would like to acknowledge the contribution of Dr. Michael Rutter and Dr. Marlene Soma for their input in the development of the original statements. We would also like to thank Drs. Frederic Gottrand, Christophe Faure, Luigi dall’ Oglio, and Usha Krishnan for their critical review of the manuscript.