Background: Analyzing small-bowel capsule endoscopy (CE) images is time consuming.
Objective: To determine the effect of reducing the number of images on reading time and interpretation of CE procedures.
Design: Two techniques aimed at reducing the number of images to be viewed were studied. The number of images was reduced by removing every second image (study A) or by the Quickview mode (study B). In both studies, one endoscopist viewed the images in the conventional way, whereas another endoscopist viewed the reduced number of images.
Setting: A single-center prospective study.
Patients: Two hundred CE procedures, 100 consecutive procedures for each study.
Main outcome measurements: Reading times for small-bowel images were recorded. Kappa (kappa) values were used to calculate interobserver agreement between viewing techniques. Diagnostic miss rates were calculated.
Results: Median procedure reading times were significantly reduced by viewing half the number of images (10.2 minutes) or using the Quickview technique (4.4 minutes) compared with conventional viewing (17.0 minutes). Interobserver agreement was excellent (kappa = 0.91) in study A and good (kappa = 0.74) in study B. The diagnostic miss rate was 2% when half the number of images were viewed and 8% when the Quickview technique was used. Agreement between both techniques and conventional viewing was best when the indication for the procedure was suspected inflammatory bowel disease.
Limitation: Lack of a criterion standard.
Conclusions: Techniques that reduce the number of images examined in CE are time saving but are associated with considerable diagnostic miss rates. Such techniques may only be considered if inflammatory bowel disease is suspected.