METHODS: A cross-sectional study was conducted of consecutive diagnostic EGDs performed at a university-affiliated, teaching hospital, which has an open-access endoscopy system for doctors who work in the hospital. The main indication(s) for EGD was recorded and assessed as appropriate or inappropriate by using American Society for Gastrointestinal Endoscopy criteria. EGD findings were recorded and classified as positive or negative. Referrals were categorized as being from endoscopists, primary care physicians, and others.

RESULTS: Of 1076 referrals for EGD, 88.3% were deemed appropriate. The group with the highest rate of appropriate referral was endoscopists (90.2%), followed by primary care physicians (89.6%) and "others" (81.9%). The rate of appropriate referrals was significantly higher for endoscopists and primary care physicians compared with "others" (respectively, p=0.001 and p=0.022). The most common appropriate indication was "upper abdominal distress that persists despite an appropriate trial of therapy" (35.4%). The most common inappropriate indication was "dyspepsia in patients aged 45 years or below without adequate empirical medical treatment" (48.4%); 42.2% with an appropriate indication had positive findings compared with only 25.6% of those with inappropriate indications (p=0.006). On multivariate analysis, the following were identified as independent predictive factors for positive findings at EGD: male gender (p=0.005), age over 45 years (p=0.011), smoking (p=0.005), none/primary education (p<0.001), and secondary education (p=0.026).