The effect of periodic monitoring and feedback on screening colonoscopy withdrawal times, polyp detection rates, and patient satisfaction scores

Abstract

Previous studies showed a correlation between mean withdrawal times during screening colonoscopy and polyp/neoplasia detection rates. To assess the effect of a monitoring and feedback program on withdrawal times, polyp/neoplasia detection rates, and patient satisfaction. Comparison of retrospective and prospective data. Teaching hospital. Asymptomatic adults undergoing screening colonoscopy. Monitoring and feedback program. Withdrawal times, polyp and neoplasia detection rates, and patient satisfaction scores. We retrospectively reviewed 850 screening colonoscopies, recording withdrawal times, polyp findings, and patient satisfaction scores. All procedures were performed by 10 experienced gastroenterologists who were then informed that periodic confidential monitoring and feedback of withdrawal times, polyp detection rates, and satisfaction scores would be started. We then prospectively collected data on another 541 screening colonoscopies. We compared pre- and postmonitoring outcome measures. Overall, after monitoring had begun, there was an increase in mean withdrawal times (from 6.57 to 8.07 minutes; P < .0001), and polyp detection rates (from 33.1% to 38.1%; P = .04, significance removed by Bonferroni correction). Nine of the 10 endoscopists increased their withdrawal times significantly. There was a small, nonsignificant increase in the neoplasia detection rate (from 19.6% to 22.7%; P = .17), but no significant change in mean satisfaction scores. Across endoscopists, there was a moderate correlation (r = 0.63; P = .04, significance removed by Bonferroni correction) between withdrawal times and polyp detection rates, but not between withdrawal times and satisfaction scores. No randomization, possible response bias, confounding of intervention effects, and sample size limitations. Monitoring and feedback are associated with increases in mean withdrawal times and polyp detection rates, but not patient satisfaction scores. Neoplasia detection rates showed a statistically nonsignificant trend toward an increase.