Reporting endoscopy quality and adverse events: Dutch step forward

Quality assurance protocols in endoscopy require the reporting of crucial quality measures as well as adverse event rates [1] [2]. Defining endoscopy quality by performance measures, including adenoma detection rate, cecal intubation rate, and adequate bowel preparation rate, has become standard procedure in many centers and/or screening programs. However, the reliability of methods to measure adverse event rates is still uncertain. Although it is a good clinical practice to discuss local complications at mortality and morbidity conferences, such meetings cover only selected cases and do not present a broader perspective on the quality of service or provide sufficient detail for research purposes. Much effort has been made in the past to propose a unified methodology for measurement of the adverse event rate. Self-reporting by endoscopy units is hampered by an unwillingness to report their own complications. Efforts to obtain feedback from patients require usage of limited human resources. Electronic tools to increase completeness of feedback within 30 days post-procedure are promising, but failed to achieve spectacular success in one randomized controlled trial [3]. Finally, cross-checking against health registries is considered to be the most objective method; however, this approach only allows data on hospitalizations and deaths to be collected and does not always allow for the assessment of causality. In addition, registry analysis is not always a feasible method as it requires a lot of time and human resources. Furthermore, the true estimates are not always clear as the method used to collect the data can have a significant impact on results. Adler et al. revealed a 3.1-fold difference in complication frequency depending on whether registry data, patient feedback, or hospitalization data were used [4]. Another problem lies in the varying definitions of adverse events used in studies and the differing methods of assessing causality (i. e. whether the complication was related or unrelated to the procedure in question), resulting in inconsistent findings. From a public perspective, 30-day adverse event rates should always be put into context with data from a comparison group that did not undergo the procedure in question, which is usually lacking. “Subtracting” natural history events during a 30-day period should be included in the equation. Such an example has been reported previously [5]. Finally, the indication for a procedure is often not taken into account in large databases of adverse events.

“The study has a unique value as it presents data obtained by linkage of appropriate registries in a large nationwide cohort of colonoscopies. These data may be used for benchmarking according to indications for endoscopic procedures.”

Some of the aforementioned problems were resolved by Nass et al. in a Dutch study published in this issue of Endoscopy [6]. The authors should be congratulated for attempts to simplify the whole process. The main achievement lies in the creation of two national registries facilitating up-to-date feedback on endoscopy performance and rate of adverse events. The Dutch Gastrointestinal Endoscopy Audit (DGEA) allows for automated extraction of quality data from standardized endoscopy reports without engaging any human resources. This is also the solution used in the Polish Colonoscopy Screening Program [7] and is considered to be the target standard when creating an endoscopy database. The other registry, the Dutch Registration of Complications in Endoscopy (DRCE) unfortunately requires manual recording of 30-day adverse events by endoscopists themselves. All types of adverse events are registered, including cardiovascular and pulmonary. Severity is noted, as well as likelihood of causality (e. g. “related,” “likely related,” “possibly related,” and “unlikely related”). Finally, the process leads to external linkage of data by a certified third party institution and anonymization for research purposes. Noticeably, the DGEA covers 80.5 % of all endoscopy services in the country; the coverage for DRCE is 96.1 %. The authors intend to obtain nearly full coverage in the near future.

A total of 277 913 colonoscopies were recorded in DGEA and 1784 adverse events were recorded in DRCE. Importantly, authors were able to provide results in relation to indication for colonoscopy: fecal immunochemical test (FIT)-positive screening, surveillance, diagnostic, and therapeutic colonoscopies. On a national level, the cecal intubation rate was 92.4 %, the highest being among FIT-positive screening colonoscopies (97.1 %) and the lowest (understandably) for therapeutic colonoscopies (83.1 %). Adequate bowel preparation rate overall was 95.1 %, with, again, the highest for FIT-positive examinations (97.1 %) and the lowest for therapeutic colonoscopies (90.8 %). For all procedures, the colonoscopy-related adverse event rate was 0.58 %, including bleeding (0.35 %) and perforation (0.06 %); for procedures with polypectomy in FIT-positive screening colonoscopies, these values were 1.17 %, 0.93 %, and 0.06 %, respectively.

The study by Naas et al. has a unique value as it presents data obtained by linkage of appropriate registries in a large nationwide cohort of colonoscopies. These data may be used for benchmarking according to indications for endoscopic procedures. However, there are still problems to be resolved globally, the most prominent being the reliance on self-reporting of adverse events by endoscopists. Another issue is the very serious legal barrier, which blocks access to individual data in line with the European Union General Data Protection Regulation (GDPR). The scientific community should strive to influence policy makers in order to overcome this GDPR barrier for research purposes. Nass et al. discussed this issue in their article, as regulations precluded them from obtaining more detailed data, especially on adverse events. In order to be able to present the most objective data and full picture of endoscopy safety, an important aspect of health services for the public, this issue needs to be worked out as soon as possible.

Let us hope that the future reporting of quality and adverse events in endoscopy services will improve further. The Dutch approach presented by Naas et al. should be regarded as a really important step forward – one that should be followed by other countries.

Publication History

Article published online:
25 March 2022

• References

• 1 Kaminski MF, Thomas-Gibson S, Bugajski M. et al. Performance measures for lower gastrointestinal endoscopy: a European Society of Gastrointestinal Endoscopy (ESGE) quality improvement initiative. Endoscopy 2017; 49: 378-397
• 2 Valori R, Cortas G, de Lange T. et al. Performance measures for endoscopy services: a European Society of Gastrointestinal Endoscopy (ESGE) quality improvement initiative. Endoscopy 2018; 50: 1186-1204
• 3 Bugajski M, Wieszczy P, Pisera M. et al. Effectiveness of digital feedback on patient experience and 30-day complications after screening colonoscopy: a randomized health services study. Endosc Int Open 2019; 7: E537-E544
• 4 Adler A, Lieberman D, Aminalai A. et al. Data quality of the German screening colonoscopy registry. Endoscopy 2013; 45: 813-818
• 5 Kobiela J, Spychalski P, Wieszczy P. et al. Mortality and rate of hospitalization in a colonoscopy screening program from a randomized health services study. Clin Gastroenterol Hepatol 2020; 18: 1501-1508
• 6 Nass KJ, van der Schaar PJ, van der Vlugt M. et al. Continuous monitoring of colonoscopy performance in the Netherlands: first results of a nationwide registry. Endoscopy 2022; 54: 488-495
• 7 Bugajski M, Rupinski M, Wieszczy P. et al. Key performance measures for colonoscopy in the Polish Colonoscopy Screening Program. Endoscopy 2019; 51: 858-865