Selection of the physician-scientists for this meeting were based on a number of criteria. First, was to be sure that attendees represented those with active and recent research in the area of breath testing research. This involved a literature search to understand those in North America with active publications in the area. Many of these included physicians from major academic motility programs. The second was to also have representation from clinician scientists from high volume breath testing referral centers to ensure those with a vast experience in the use and interpretation of breath tests. The third was to ensure representation from various regions in the US as well as Mexico and Canada. An email was sent to those meeting these criteria. A total of 17 clinician scientists from North America were invited to participate in the survey and attend the consensus meeting, of whom 10 were able to attend the meeting in person. The half day meeting was held on 16 May 2015 in Washington, DC. At the consensus group meeting, results of the survey and comprehensive literature review were presented for discussion among the group members.
Distinct topics of discussion included indications, preparation, performance, interpretations of results, and future directions in breath testing. Based on the results of these discussions, a series of draft consensus statements was compiled and sent to committee members for online voting using SurveyMonkey (SurveyMonkey, Palo Alto, CA) with revision as necessary. Using a modified Delphi process (19 (link), 20 (link)), group members anonymously voted on their level of agreement with each statement on a scale of 1–3 (disagree, uncertain and agree, respectively). A statement was accepted if >70% of participants voted 3 (agree), and it was rejected if >50% of participants voted 1 (disagree). If neither of these criteria were reached, it was stated that “a firm position statement could not be reached due to lack of conclusive data”. The strength of recommendation for each statement was assigned by the consensus group as either strong (“we recommend...”) or weak (“we suggest...”). The strength of each statement was based on resource and cost benefit, patients' values, risk/benefit balance and quality of evidence. The quality of evidence for each consensus statement was classified as high (⊕⊕⊕⊕), moderate (⊕⊕⊕⊕), low (⊕⊕⊙̇), or very low (⊕⊙⊙⊙) based on Grading of Recommendations, Assessment, Development and Evaluations (GRADE) system (21 (link)). The manuscript was circulated to all group members for review, revisions and approval.