The accuracy of an endoscopy nurse in interpreting capsule endoscopy

Objective: The average physician time required to view a wireless capsule endoscopy study at our institution is 50 min. It is unknown whether a nurse could preview the capsule endoscopy video and accurately detect all significant lesions. If so, it may allow the physician to review only the predetected abnormalities and thereby greatly reduce physician reading time. Our aim was to evaluate whether a nurse can accurately detect lesions on capsule endoscopy. Methods: An endoscopy nurse who was trained to read capsule endoscopy reviewed 20 consecutive capsule endoscopy studies and recorded all findings. The same studies were viewed independently by a gastroenterologist. The two sets of recorded findings were reviewed, and a comparison of the accuracy in detecting landmarks and clinically significant lesions was made. Results: The nurse missed two of 27 significant lesions seen by the gastroenterologist (93% sensitivity, 95% CI=74-99%), and the gastroenterologist missed three seen by the nurse. The nurse accurately recorded gastric emptying time and time of passage through the ileocecal valve to within I min of the times recorded by the gastroenterologist in 18 of 20 patients (90% agreement, 95% CI=67-98%). Conclusions: In this study, the endoscopy nurse detected 93% of the clinically significant lesions seen by the gastroenterologist. The clinical implication of this is that a physician extender could preread capsule endoscopies, allowing the gastroenterologist to view only the demarcated abnormalities. This could improve the cost effectiveness of capsule endoscopy and lead to wider physician acceptance of the test.