Primary care physician (PCP)-performed colonoscopy screening: Does PCP procedure experience affect performance quality?

Background Screening colonoscopy to clear the colon of precancerous polyps could greatly reduce colorectal cancer incidence. There is 50% shortage of endoscopists relative to the annual need for 23 million screening colonoscopies. Scarce specialist resources cause de facto rationing, and consequently may foster disparities among underserved populations. Training PCPs in high quality colonoscopy performance may be one solution.

Methods To compare the quality of PCPs' colonoscopy performance with that of specialists, and among PCPs, to evaluate whether their training procedures are of lower quality than their post-training procedures. Quality is assessed by procedure time, adenoma detection rate (ADR), and mean number of adenomas per screened person (MNA). Since 2001, an innovative, polyp detection-maximizing colonoscopy protocol along with systematic data documentation has been implemented by an endoscopy center in South Carolina. The center trained 54 PCPs, and post training, provided them hands-on technician assistance and onsite specialist for rescue assistance. A previous study of a subsample of the PCP series (2001-2007 series) documented an adenoma detection rate (ADR) of 29.9% with a 0.06% complication rate, both rates being comparable to high performing specialists' series documented in the literature.

Results Retrospective colonoscopy data on 19,205 screening colonoscopies performed during 2001 to 2009 by PCPs (75%) and 5 specialists (25%), at an endoscopy center in South Carolina were analyzed. PCP rates on all indicators were not inferior to those of specialists, and within PCPs, training procedures were not inferior to post training procedures. PCPs had longer average procedure time than specialists (26.4 minutes vs. 24.6 minutes), marginally higher ADR than specialists (32% vs. 30.3%), and somewhat higher MNA (0.51 vs. 0.46). Grouping procedures by performing PCP's experience level (1st-20th, 21st-40th,…, 300+), showed highest procedure time up to the 40th procedure, gradually decreasing to 23.3 minutes (±10.75 minutes) at 300+ level. Because the gastroenterology credentialing standard for colonoscopy specifies 140 training procedures under specialist supervision, we stratified PCP procedures into 0-140th procedures and 141+ procedures. Analysis shows that PCPs' training procedures are not inferior to post-training procedures (ADRs 33.43% vs. 31.22% respectively, p=0.001) with similar findings for MNAs.

Conclusions PCP-performed colonoscopies, either during training or post training are not of inferior quality relative to those by specialists, likely because of uniform, polyp-detection maximizing protocol being implemented at this center, and post-training technical support provided to PCPs. This center's model of PCP utilization for screening colonoscopies could be replicated for addressing the colonoscopy capacity gap.