Small bowel capsule endoscopy (CE) is a non-invasive and validated technique allowing the visualization of the small intestine mucosa. It generated more than 40,000 images per examination and induced a long median time of medical reading that may exceed 1 hour. In a transfer of skills’ program, the CE reading might be transferred to nurses. We herein evaluate and present a feasibility study of the CE to nurses.

Nurses experienced in endoscopy were trained for small bowel CE reading, selection and interpretation of pathological images. Two strategies were experienced: a partial delegation to nurses, restricted to the selection of picture and an interpretation of the selected pictures by experts in CE (strategy A) or a total delegation including the selection of pictures, their interpretation, and the realization of a report (strategy B). Nurses were volunteers to participate to the study. Strategies were compared to the reference procedure (control) and a blinded analysis of the CE by experts was performed. Selection of pictures and their interpretation were timely evaluated in all strategies. The stomach and small intestine transit times and the reading time were recorded. The accuracy of the whole CE interpretation was compared. Experts reviewed all discordant interpretations.

The extent of transit time in the stomach and the small intestine, and the identification of a normal endoscopic examination were not statistically different between strategy A and control (P=0.71). The accuracy of reports from strategy A and control was 95% (P=0.9). In one case, a nurse has not selected the pathological image, corresponding to a duodenal ulcer. The median medical time for reading was significantly shorter in strategy A than in control (9.2min [range: 4–20] vs 34.0min [range: 10–60]; P<0.01). The accuracy of reports performed by nurses in strategy B was 80% (P=0.11). In 4 cases, the medical interpretation in strategy B was discordant. The discrepancy was due to a lack of pathological images’ selection of the selected pictures for the final report (n=1) and a lack of interpretation (n=3). The median time for analysis was not significantly different between strategy B and control (34.7 [range: 10–75] versus 34.0min [range: 15–60]; P=0.53).

In small bowel CE, a pre-selection of pathological images by a trained nurse could be integrated in a transfer of skills’ program and appeared feasible. The gain of time for the gastroenterologists is significant and reduced the medical time for reading from 34 to 9minutes per examination. Meantime, an overall delegation of small bowel CE to nurses suffered for a lack of efficacy. Our study paved the way for a nurse’ training program in the pre-selection and identification of CE images.