Non-physician performance of lower and upper endoscopy: a systematic review and meta-analysis

Background and study aims: Demand for endoscopic procedures worldwide has increased while the number of physicians trained to perform endoscopy has remained relatively constant. The objective of this study was to characterize nonphysician performance of lower and upper endoscopic procedures. Patients and methods: Bibliographical searches were conducted in Medline, EMBASE, and Cochrane Library databases. Studies were included where patients underwent flexible sigmoidoscopy, colonoscopy, or upper endoscopy done by a non-physician (nurse, nurse practitioner, physician assistant) and outcome measures were reported (detection of polyps, adenomas, cancer, and/or adverse events). Pooled rates were calculated for specific outcomes and rate ratios were determined for selected comparison groups. Results: Most studies involved nurses performing flexible sigmoidoscopies for colorectal cancer screening. Nurses and nurse-practitioners/physician assistants performing flexible sigmoidoscopies showed pooled polyp detection rates of 9.9% and 23.7%, adenoma detection rates of 2.9% and 7.2%, colorectal cancer detection rates of 1.3% and 1.2%, and adverse event rates of 0.3 and 0 per 1000 sigmoidoscopies, respectively. There was no significant difference between polyp and adenoma detection rates in sigmoidoscopy performance studies comparing nurses or nurse-practitioners/physician assistants with physicians. For the 3 studies of non-physician performance of colonoscopy, pooled adenoma detection rate was 26.4%, cecal intubation rate was 93.5%, and adverse event rate was 2.2/1000 colonoscopies. In the few studies examining upper endoscopies, 99.4% of upper endoscopy procedures performed by nurses were successful with no reported adverse events. Conclusion: Available studies suggest that when non-physicians perform endoscopic procedures, especially lower endoscopies, outcomes and adverse events are in line with those of physicians.