Determining the indications for diagnostic upper endoscopy using the A/S/G/E guidelines: Can non physicians do just as well as physicians?

Purpose: To determine whether non physicians can assess the appropriateness of indications for diagnostic upper endoscopy as well as physicians. Methods: All written Gastroenterology consultation requests were reviewed by a trained research assistant (RA), a nurse practitioner (NP) experienced in Gastroenterology, and attending Gastroenterologists (MD). The RA received formal training on the A/S/G/E approved indications for diagnostic endoscopy and used a check list of indications to review the consultation requests. The NP and MD were familiarized with the indications but did not use the check list. All reviewers were independent and blinded from other reviewers. The 'true' appropriateness was determined by consensus during a panel discussion by all of the reviewers using the A/S/G/E guidelines. Agreement was calculated using the K statistic. Results: 124 of 348 consecutive consultation requests reviewed during a 10 week period were for upper gastrointestinal problems. The final panel discussion identified 55 consultation requests as having one or more explicit A/S/G/E indications for diagnostic upper endoscopy. The RA correctly identified 43 (sensitivity=78%, accuracy= 88%,κ =0.75) cases, the NP correctly identified 39 (sensitivity=71%, accuracy=85% κ=0.69) cases, and the MD correctly identified 43 "sensitivity=78%, accuracy=90%, κ=0.79) cases. The reviewers concurred in only 31 (56%) cases. The MD correctly identified 8 appropriate cases that the RA or the NP did not; 4 omissions resulted from medical ignorance. The RA and the NP correctly identified 5 appropriate cases that the MD did not - all were due to oversight. The RA and the NP incorrectly identified 3 cases as having appropriate indications; there were no false positive identification by the MD. The RA and the NP failed to correctly identify 2 cases each, that other reviewers identified as being appropriate for endoscopy. 95% of cases in which all the reviewers agreed as being appropriate had pathology found including cancer (25%), ulcer (15%), esophagitis (25%), gastritis (20%). Conclusion: 1) RA, NP, and MD correctly identified <80% of consultation requests with appropriate indications for diagnostic upper endoscopy. 2) RA and NP were as sensitive and accurate as the MD in identifying A/S/G/E approved indications for endoscopy. 3) MD failed to correctly identify appropriate indications due to oversight, whereas RA and NP failed to correctly identify indications because of oversight or lack of understanding. 4) These data suggest that trained non physician assistants can accurately identify appropriate indications for diagnostic upper endoscopy.