Endoscopy 2019; 51(09): 813-815
DOI: 10.1055/a-0967-1523
© Georg Thieme Verlag KG Stuttgart · New York

Essential building blocks for colonoscopy quality improvement initiatives: a dedicated database, automation, and appropriate financial incentives

Referring to Bugajski M et al. p. 858–865
Subhas Banerjee
Stanford University School of Medicine, Stanford, California, USA
› Author Affiliations
Further Information

Publication History

Publication Date:
28 August 2019 (online)

Colorectal cancer (CRC) remains a leading cause of cancer-related mortality in the western world. Colonoscopy has assumed a predominant role in CRC screening, because of its high sensitivity, the ability to perform concurrent polypectomy, and its advantage, over other screening strategies, of allowing longer intervals between testing. Although colonoscopy is associated with a reduction in CRC mortality [1], interval cancers, which account for 3 % – 9 % of all colon cancers, remain a persistent and troubling issue. Although some interval cancers may represent de novo accelerated biology-related cancers, it is believed that the majority are attributable to missed, or to incompletely resected lesions. High quality colonoscopy is therefore a fundamental prerequisite for reduction in CRC mortality. Consequently, quality indicators/performance measures for colonoscopy have been developed by gastrointestinal societies in the United States [2], and more recently by the European Society of Gastrointestinal Endoscopy (ESGE) [3], with the goal of quantifying the quality of colonoscopy. Substantial variability has been demonstrated in the quality of colonoscopy between endoscopists, resulting in suboptimal CRC prevention [4]. This has provided impetus for the implementation of quality assurance programs, typically comprising continuous quality assessment, with intermittent feedback and quality improvement interventions. In this issue of Endoscopy, Bugajski and colleagues evaluate the performance of the Polish Colonoscopy Screening Program (PCSP) in adhering to the ESGE quality guidelines over the years 2014 – 2015, capturing the performance of 25 endoscopy units and 117 endoscopists over a total of 43 277 screening colonoscopies.

“Mid- and late-career colonoscopists, who may be challenging to retrain, will gradually be diluted and replaced by younger colonoscopists trained in a culture where quality is stressed.”

What lessons may others embarking on quality improvement initiatives draw from the Polish experience? Firstly, the Polish group should be applauded for having established not just a robust CRC screening program, but for having had the foresight to develop a dedicated database designed to facilitate quality monitoring. The system acquires an updated list of eligible individuals annually from the national population registry and invitation letters to participate in screening are sent out. Patient registration in the database is mandatory and financial incentives are built in, with reimbursement only provided for procedures and pathology registered in the database. Linking of pathology and colonoscopy reports within the database allows for automatic calculation of the adenoma detection rate (ADR), a performance measure that is typically problematic to determine. The database additionally calculates the rates of adequate bowel preparation, cecal intubation, and painful colonoscopy. Finally, it automatically generates recommendations for appropriate surveillance intervals based on ESGE guidelines and the nature and number of polyps; the surveillance interval may be altered at the discretion of the endoscopist.

Tracking of two measures was more complex. Assessing complication rates required accessing additional national registries. As this is not ideal for routine monitoring, the PCSP is testing the functionality of self-reporting of complications. Additionally, the database could not directly determine whether snare polypectomy was utilized for polyps larger than 3 mm, as it is currently set up to capture polyp sizes of 1 – 5 mm and higher ranges. Therefore, this performance measure had to be determined manually, and to simplify the task it was determined only in a subset of colonoscopies where a single polyp was excised. This minor impediment to an otherwise largely automated process should be easy to resolve with a software patch, allowing capture of polyp sizes and polypectomy methods consistent with quality assurance priorities, thereby obviating the need for and costs associated with additional human resources

The study evaluated performance across all tiers of delivery of endoscopic care, from the national program level, to service and individual endoscopist levels. Reassuringly, at the program level, the PCSP was noted to meet minimum standards for adequate bowel preparation rate, cecal intubation rate, ADR, and 7-day readmission rate. The largest variability in quality was noted in appropriate polypectomy technique. This is an important performance measure as it is estimated that 19 % – 25 % of interval cancers may be related to incomplete polyp resection [5]. For this performance measure, the minimum standard for polyps 4 – 5 mm in size was met by only 7 of 25 PCSP centers. These findings offer an opportunity for providing feedback and education to endoscopists. A software patch allowing the capture of desired polyp sizes and whether or not size-appropriate polypectomy techniques were applied, will simplify future data extraction and may also improve compliance because of the Hawthorne effect. The authors raise the possibility that cost-containment issues may have played a role in the low rate of adherence for this performance measure. Unlike the United States where one may code for both a cold biopsy and a snare polypectomy in the same procedure, the PCSP offers no additional financial incentive for endoscopists to perform polypectomy, and indeed a cost burden is imposed if both biopsy forceps and polypectomy snare are utilized in the same procedure. This inconsistency may need to be revisited by the PCSP. Appropriate financial incentives with “pay for performance” may be a stronger motivator of change in endoscopist practice, than standard quality assurance interventions [6].

None of the 25 centers met the target for measuring patient experience. This is not entirely surprising, as the process requires patients to fill in a questionnaire the day following the procedure, then mail it back. Such surveys inevitably have low response rates and are prone to bias. Immediacy and brevity are key to achieving high response rates, as evidenced by the runaway success of the Finnish “HappyOrNot” smiley face feedback terminals, which allow for effortless consumer feedback. A short, real-time, digital survey completed by patients just prior to discharge following colonoscopy may therefore be a superior instrument for capturing the immediate patient experience. As some patients may feel vulnerable while still in the endoscopy unit, anonymous surveys should be considered. Although this approach will not capture delayed complications, it appears that the PCSP may already be moving towards patient self-reporting of complications. Additionally, it is conceivable that patients “primed” by participation in immediate online feedback may be more likely to subsequently self-report delayed complications.

At the service and individual levels, inevitably, significant variability was evident across all performance measures. In particular, 7 of 25 endoscopy units did not meet the target ADR of 25 % despite this being an arguably low bar, given that a progressive reduction of interval CRC has been demonstrated with higher ADR rates [7]. As the most desired outcome of screening is a decrease in CRC mortality, ADR is arguably the most important performance measure, given the significant association of endoscopist ADR with the risk of interval CRC [8]. As such, quality assurance interventions should particularly emphasize improving ADRs.

The effectiveness of quality assurance interventions has been questioned, with several early studies indicating a depressing lack of improvement despite interventions [9]. However, recent quality assurance intervention studies have been more encouraging [10] [11]. These positive studies have emphasized formal performance improvement programs with continuous feedback at short intervals, intensive education with retraining where necessary, and training of leaders at endoscopy centers as a means to disseminate teaching of high quality colonoscopy. Similarly, the setting of a relatively high entry bar on quality for endoscopists wishing to participate in the UK Bowel Cancer Screening Programme, including an accreditation exam and ongoing performance audits, have resulted in an admirably high mean ADR of 46.5 % within the program [12].

Overall, there is reason for optimism that quality will continue to improve across all screening programs, with increasing implementation and acceptance of quality assurance programs, together with innovations in quality assurance interventions. Mid- and late-career colonoscopists, who may be challenging to retrain, will gradually be diluted and replaced by younger colonoscopists trained in a culture where quality is stressed. The transformative impact of financial incentives with greater linkage of payment to performance should not be underestimated, at least in healthcare systems where these can be implemented. Finally, underlying trends in colonoscopy including improvements in technology, techniques, and training will result in continual advances in quality, independent of quality assurance programs.

In summary, developers of the PCSP have created an admirable and robust national screening and quality assurance program by attention to two critical elements: automation, with careful design of a dedicated database, and leveraging of financial incentives, thereby compelling use of the database by all stakeholders. Further refinements to the database together with structured quality assurance interventions should result in improved quality across all tiers of the PCSP.