In Search of Quality Colonoscopy
Article Outline
See “Effect of institution-wide policy of colonoscopy withdrawal time >7 minutes on polyp detection,” by Sawhney MS, Cury MS, Neeman N, et al, on page 1892; and “Bleeding and perforation after outpatient colonoscopy and their risk factors in usual clinical practice,” by Rabeneck L, Paszat LF, Hilsden RJ, et al, on page 1899.
Recent guidelines representing multiple societies promote colonoscopy as the optimal test for colorectal cancer (CRC) screening.1 Although no direct evidence (ie, randomized, controlled trials) exist to prove the efficacy of colonoscopy to reduce mortality from CRC, substantial indirect evidence has been published supporting a general belief that this strategy will decrease cancer incidence and mortality more effectively than other available means.2, 3 The recent decline in CRC mortality has been attributed to the wide-scale implementation of CRC screening and, because all screening strategies involve colonoscopy as either a primary or evaluative instrument, this seems to further support colonoscopy as the best intervention.4
However, other data question the effectiveness of screening colonoscopy to reduce cancer incidence.5 How then can one resolve these divergent observations? Effectiveness reflects the ability of an intervention to achieve desired outcomes in clinical practice; contrast this with efficacy, which is observed in the controlled environment of clinical trials. Perhaps the gap between clinical trials and clinical practice can be explained through differences in endoscopist performance of colonoscopy with regard to detection and safe removal of neoplasia. Two articles published in this issue of Gastroenterology reveal potential factors relating to procedural performance that impact the effectiveness of colonoscopy and direct us to refine this technique so as to optimize strategies to reduce CRC mortality.
The study by Rabeneck et al6 confirms a low rate of serious complications associated with colonoscopy. Bleeding occurred in <2 per 1,000 colonoscopies and perforation was limited to <1 per 1,000 procedures. Mortality related to colonoscopy was <1 in 10,000 procedures. These data are congruent with estimates from population based studies in the United States, but higher than reports published from selected endoscopy centers.7, 8, 9 The discrepancy is most likely related to the difference between efficacy and effectiveness, represented here by the difference between patients and providers in academic medical centers versus those in clinical practice. For population-based CRC screening, it is certainly more relevant to use the population-based complication rates, because these reflect community practice and the results that could be expected outside of research settings. For this reason, rates of hemorrhage, perforation, and death associated with colonoscopy described to patients for the purposes of obtaining informed consent before performing colonoscopy should use these updated statistics. Additionally, health care policy should be based on these population-based estimates, which means that future decision and economic analyses focused on CRC screening should incorporate data from these sources.
The investigators were able to identify several factors that were associated with higher complication rates. Patient factors associated with increased complications included older age, male gender and greater comorbidity. Procedural factors predicting increased complications were limited to performance of a polypectomy, whereas provider factors associated with increased bleeding or perforation included lower volume endoscopists, especially those performing <300 colonoscopies per year. A secondary analysis examining the number of colonoscopies performed by an endoscopist in the previous 5 years revealed similar results, suggesting that experience rather than repetition could be the factor most linked to complications.
Most revealing was the subanalysis restricted to gastroenterologists that excluded surgeons and nongastroenterology internists or family physicians. This analysis failed to identify a factor predictive of complications other than performance of a polypectomy. The interpretation of the dichotomy between factors elicited in the overall cohort versus the gastroenterology subgroup could be that endoscopy training, specifically the requirements for gastroenterology fellowship training, can overcome the increased risk of procedural complications otherwise invoked by patient age, gender, and comorbidity, and mitigate the impact of endoscopic volume. There are stark differences in training requirements for gastroenterology fellows compared with surgical residents. The recommended threshold for assessing competence for endoscopic procedures among trainees in gastroenterology is 140 colonoscopies, 30 of which should involve snare polypectomies.10 Despite a recent increase in requirements for surgical residents, the total procedural requirement is 50 colonoscopies.11 It is conceivable that the lower procedural volume in surgical programs results in greater reliance on posttraining experience to reduce the risk of complications otherwise posed by patient factors. The implications for these findings might include a mandate for increased volume of procedures that are required during training before assessment of competency that is consistent across all providers who obtain credentialing for colonoscopy.
A potential factor separating effectiveness from efficacy is the quality of colonoscopy performed in a screening strategy. One may postulate that higher quality colonoscopy performed in academic medical centers in the context of a clinical trial compared with screening colonoscopy performed in a community setting could explain the deficit in outcome realized between research and community estimates of the benefit of screening colonoscopy. One of the factors associated with the performance of quality colonoscopy is withdrawal time, defined as the time elapsed between cecal intubation and completion of the colonoscopy, during which the majority of mucosal inspection is performed. A well-publicized study by Barclay et al12 demonstrated a linear association between mean withdrawal time for a screening colonoscopy during which no polyps were found and the overall adenoma detection rate associated with an endoscopist. It should be noted that no asymptote was demonstrated in this study, because even the practitioner who had a 17-minute average withdrawal time during a “negative” colonoscopy exhibited a continued linear increase in adenoma detection beyond his counterparts. Therefore, the withdrawal time that maximizes adenoma detection rates is unknown, and slower withdrawal times could yet be associated with even greater adenoma yield.
The study by Sawhney et al13 published in this issue of Gastroenterology seems to refute the hypothesis that institution of a quality measure based on colonoscopy withdrawal time can increase rates of polyp detection during screening colonoscopy. In this prospective study, an intervention of provider education supplemented by profiling (individualized feedback regarding results of a quality measure) was used to influence endoscopists to increase their withdrawal times to >7 minutes. The program seemed to successfully increase the proportion of endoscopists adherent to the 7-minute threshold; however, the investigators were unable to demonstrate a significant increase in the number of polyps or the proportion of patients in whom polyps were detected. This study calls into serious question the usefulness of a withdrawal time threshold as a quality measure or a metric for endoscopist competence.
An apparent limitation of this study was that an endpoint of polyp detection was used because histopathologic confirmation was not available and adenoma detection could not be assessed. This omission would have likely biased the study to observe an increase in polyp detection because “imaginary” lesions could be documented to artificially increase the desired polyp detection rate; however, this was not observed. A more plausible explanation for the findings of this study is that withdrawal time is a marker identifying “good” endoscopists and not necessarily the cause of improved adenoma detection. Endoscopists who have a longer withdrawal time may differ from their counterparts with respect to technique, meticulousness, or powers of observation. Thus, the use of withdrawal time as a quality standard may not adequately reflect the true characteristics of a good endoscopist. Certainly, this supports a view that not all 7 minutes are equal. From one extreme this can be illustrated by comparing one endoscopist who divides a 7-minute withdrawal time equally among all segments of the colon versus another endoscopist who rapidly withdraws the scope, thereby resulting in poor visualization of proximal segments but spends 6 minutes in the rectum to fulfill the “quality standard.”
Another potential reason that this intervention failed to increase polyp detection could have been that the endoscopists had already reached the limit of detection. At the onset of the study, when only 63% of endoscopists recorded a withdrawal time of ≥7 minutes, the baseline polyp detection rate (number of polyps identified divided by the number of colonoscopies performed) was 48% and the proportion of patients in whom polyps were identified was 35%. Considering that the standards for quality colonoscopy include threshold adenoma detection rates of 15% in women and 25% in men, the baseline detection rate before initiation of this intervention may have left little to be gained in this cohort of endoscopists.14
Because the primary goal of screening and surveillance colonoscopy is to remove lesions possessing malignant potential, it is much more thoughtful and reasonable to assume that if adenoma detection rates for a particular endoscopist were greater than an established threshold appropriate for her patient population, further efforts to measure effectiveness in quality intrinsically are less important. Perhaps the 7-minute rule should not be a universal quality measure, but employed only among those endoscopists who fail to meet the primary goals of adenoma detection. Certainly, proposed pay-for-performance measures should take into account these observations and be developed with a priori hypotheses that are most likely to improve the true quality of colonoscopy.
From a more global view, the efforts to improve colonoscopic screening should be taken in proper perspective. Despite our enthusiasm for colonoscopic screening, one must realize that the lifetime risk of CRC of 6% means that even if colonoscopy were completely effective in eliminating CRC, 94% of screened patients will not benefit from screening programs. Therefore, the task at hand is to develop reliable, noninvasive, and risk stratification strategies that accurately identify which patients harbor neoplasia. Moreover, it is unlikely that all adenomatous polyps are destined to undergo malignant transformation; therefore, it may not be necessary to remove all adenomas, but instead identify and remove only adenomas destined to undergo malignant transformation, a pursuit that would involve biomarkers and superior optical imaging. Colonoscopy is an excellent tool and may be the most efficacious to reduce cancer incidence at this point in time. However, the future lies in the development of noninvasive means to limit intervention to patients who are at risk for development of CRC. We should not be satisfied that we identify adenomas but should strive to identify the specific markers that herald future transformation to CRC. Ultimately, a noninvasive intervention that reduces cancer risk in high-risk groups may overtake colonoscopy as the optimal means of reducing cancer risk. Until these goals are achieved, however, we must continue to refine the practice of colonoscopy to ensure that it is a safe and effective means of reducing the incidence and mortality associated with CRC.
References
- Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. Gastroenterology. 2008;134:1570–1595
- Prevention of colorectal cancer by colonoscopic polypectomy (The National Polyp Study Workgroup). N Engl J Med. 1993;329:1977–1981
- Use of colonoscopy to screen asymptomatic adults for colorectal cancer (Veterans Affairs Cooperative Study Group 380). N Engl J Med. 2000;343:162–168
- Annual report to the nation on the status of cancer, 1975–2004, featuring cancer in American Indians and Alaska Natives. Cancer. 2007;110:2119–2152
- Colorectal cancer in patients under close colonoscopic surveillance. Gastroenterology. 2005;129:34–41
- Bleeding and perforation after outpatient colonoscopy and their risk factors in usual clinical practice. Gastroenterology. 2008;135:1899–1906
- Complications of colonoscopy in an integrated health care delivery system. Ann Intern Med. 2006;145:880–886
- . Endoscopic perforation of the colon: lessons from a 10-year study. Am J Gastroenterol. 2000;95:3418–3422
- . Colonoscopy completion and complication rates in a community gastroenterology practice. Gastrointest Endosc. 2006;64:556–562
- Methods of granting hospital privileges to perform gastrointestinal endoscopy. Gastrointest Endosc. 2002;55:780–783
- . Surgery policy information. http://www.acgme.org/acWebsite/RRC_440/440_policyArchive.asp2006;Accessed September 22, 2008
- Colonoscopic withdrawal times and adenoma detection during screening colonoscopy. N Engl J Med. 2006;355:2533–2541
- Effect of institution-wide policy of colonoscopy withdrawal time >7 minutes on polyp detection. Gastroenterology. 2008;135:1892–1898
- Quality indicators for colonoscopy. Gastrointest Endosc. 2006;63:S16–S28
The author discloses no conflicts.
PII: S0016-5085(08)01983-5
doi:10.1053/j.gastro.2008.11.003
© 2008 AGA Institute. Published by Elsevier Inc. All rights reserved.
Refers to article:
-
Effect of Institution-Wide Policy of Colonoscopy Withdrawal Time ≥7 Minutes on Polyp Detection
, 28 August 2008
-
Bleeding and Perforation After Outpatient Colonoscopy and Their Risk Factors in Usual Clinical Practice
, 16 September 2008
