Quality Indicators for Colonoscopy

ACG: American College of Gastroenterology; ADR: adenoma detection rate; APC: adenoma per colonoscopy; ASGE: American Society for Gastrointestinal Endoscopy; CRC: colorectal cancer; PDR: polyp detection rate Colonoscopy is widely used for the diagnosis and treatment of colon disorders. Properly performed, colonoscopy is generally safe, accurate, and well-tolerated. Visualization of the mucosa of the entire large intestine and distal terminal ileum usually is possible during colonoscopy. Polyps can be removed during colonoscopy, thereby reducing the risk of colon cancer. Colonoscopy is the preferred method to evaluate the colon in most adult patients with large-bowel symptoms, iron deficiency anemia, abnormal results on radiographic studies of the colon, positive results on colorectal cancer (CRC) screening tests, post-polypectomy and post-cancer resection surveillance, and diagnosis and surveillance in inflammatory bowel disease. In addition, colonoscopy is the most commonly used CRC screening test in the United States (1). Based on 2010 data, over 3.3 million outpatient colonoscopies are performed annually in the United States, with screening and polyp surveillance accounting for half of indications (2). Optimal effectiveness of colonoscopy depends on patient acceptance of the procedure, which depends mostly on acceptance of the bowel preparation (3). Preparation quality affects the completeness of examination, procedure duration, and the need to cancel or repeat procedures at earlier dates than would otherwise be needed (4, 5). Ineffective preparation is a major contributor to costs (6). Meticulous inspection (7, 8) and longer withdrawal times (9, 10, 11, 12, 13, 14) are associated with higher adenoma detection rates (ADR). A high ADR is essential to rendering recommended intervals (15) between screening and surveillance examinations safe (16, 17). Optimal technique is needed to ensure a high probability of detecting dysplasia when present in inflammatory bowel disease (17, 18, 19, 20, 21). Finally, technical expertise and experience will help prevent adverse events that might offset the benefits of removing neoplastic lesions (22). Recent studies report that colonoscopy is less effective in preventing proximal colon cancer and cancer deaths (ie, colon cancer proximal to the splenic flexure) compared with distal cancer (ie, colon cancer at or distal to the splenic flexure) (23, 24, 25, 26, 27, 28). Decreased protection against right-sided CRC is likely due to multiple factors. These include missed adenomas or incompletely resected adenomas; suboptimal bowel preparation; precancerous lesions that are endoscopically subtle or difficult to remove, such as sessile serrated polyps and flat and/or depressed adenomas, and differences in tumorigenesis between right-sided and left-sided cancers. Improving prevention of right-sided colon cancer is a major goal of colonoscopy quality programs. Five studies have established that gastroenterologists are more effective than surgeons or primary care physicians at preventing CRC by colonoscopy (27, 29, 30, 31, 32). This most likely reflects higher rates of complete examinations (ie, cecal intubation) (30) and higher rates of adenoma detection among gastroenterologists (33, 34). All endoscopists performing colonoscopy should measure the quality of their colonoscopy. Institutions where endoscopists from multiple specialties are practicing should reasonably expect all endoscopists to participate in the program and achieve recommended quality benchmarks. The quality of health care can be measured by comparing the performance of an individual or a group of individuals with an ideal or benchmark (35). The particular parameter that is being used for comparison is termed a quality indicator. A quality indicator often is reported as a ratio between the incidence of correct performance and the opportunity for correct performance (4) or as the proportion of interventions that achieve a predefined goal (35). Quality indicators can be divided into 3 categories: (1) structural measures—these assess characteristics of the entire health care environment (eg, participation by a physician or other clinician in systematic clinical database registry that includes consensus endorsed quality measures), (2) process measures—these assess performance during the delivery of care (eg, ADR and adequate biopsy sampling during colonoscopy for chronic ulcerative colitis), (3) outcome measures—these assess the results of the care that was provided (eg, the prevention of cancer by colonoscopy and reduction in the incidence of colonoscopic perforation). METHODOLOGY In 2006, the American Society for Gastrointestinal Endoscopy (ASGE)/American College of Gastroenterology (ACG) Task Force on Quality in Endoscopy published their first version of quality indicators for colonoscopy (36). The present update integrates new data pertaining to previously proposed quality indicators and new quality indicators for performing colonoscopy (36). Indicators that had wide-ranging clinical application, were associated with variation in practice and outcomes, and were validated in clinical studies were prioritized. Clinical studies were identified through a computerized search of Medline followed by review of the bibliographies of all relevant articles. When such studies were absent, indicators were chosen by expert consensus. Although feasibility of measurement was a consideration, it is hoped that inclusion of highly relevant, but not yet easily measurable indicators, would promote their eventual adoption. Although a comprehensive list of quality indicators is proposed, it is recognized that, ultimately, only a small subset might be widely used for continuous quality improvement, benchmarking, or quality reporting. As in 2006, the current task force concentrated its attention on parameters related to endoscopic procedures; whereas the quality of care delivered to patients is clearly influenced by many factors related to the facilities in which endoscopy is performed, characterization of unit-related quality indicators was not included in the scope of this effort. The resultant quality indicators were graded on the strength of the supporting evidence (Table 1). Each quality indicator was classified as an outcome or a process measure. Although outcome quality indicators are preferred, some can be difficult to measure in routine clinical practice, because they need analysis of large amounts of data and long-term follow-up and may be confounded by other factors. In such cases, the task force deemed it reasonable to use process indicators as surrogate measures of high-quality endoscopy. The relative value of a process indicator hinges on the evidence that supports its association with a clinically relevant outcome, and such process measures were emphasized.Table 1: Grades of recommendationaThe quality indicators for this update were written in a manner that lends them to be developed as measures. Although they remain quality indicators and not measures, this document also contains a list of performance targets for each quality indicator. The task force selected performance targets from benchmarking data in the literature when available. When no data were available to support establishing a performance target level, “N/A” (not available) was listed. However, when expert consensus considered failure to perform a given quality indicator a “never event” such as monitoring vital signs during sedation, then the performance target was listed as >98%. It is important to emphasize that the performance targets listed do not necessarily reflect the standard of care but rather serve as specific goals to direct quality improvement efforts. Quality indicators were divided into 3 time periods: pre-procedure, intraprocedure, and postprocedure. For each category, key relevant research questions were identified. In order to guide continuous quality improvement efforts, the task force also recommended a high-priority subset of the indicators described, based on their clinical relevance and importance, evidence that performance varies significantly in clinical practice, and feasibility of measurement (a function of the number of procedures needed to obtain an accurate measurement with narrow confidence intervals and the ease of measurement). A useful approach for an individual endoscopist is to first measure their performances with regard to these priority indicators. Quality improvement efforts would move to different quality indicators if the endoscopists are performing above recommended thresholds, or the employer and/ or teaching center could institute corrective measures and remeasure performance of low-level performers. Recognizing that certain quality indicators are common to all GI endoscopic procedures, such items are presented in detail in a separate document, similar to the process in 20 06 (37, 38). The preprocedure, intraprocedure, and postprocedure indicators common to all endoscopy are listed in Table 2. Those common factors will be discussed in this document only insofar as the discussion needs to be modified specifically to relate to colonoscopy.Table 2: Summary of proposed quality indicators common to all endoscopic procedures (38)aPreprocedure quality indicators The preprocedure period includes all contacts between members of the endoscopy team and the patient before the administration of sedation or insertion of the endoscope. Common issues for all endoscopic procedures during this period include: appropriate indication, informed consent, risk assessment, formulation of a sedation plan, management of prophylactic antibiotics and antithrombotic drugs, and timeliness of the procedure (38). Preprocedure quality indicators specific to performance of colonoscopy include the following: 1. Frequency with which colonoscopy is performed for an indication that is included in a published standard list of appropriate indications, and the indication is documented Level of evidence: 1C+ Performance target: >80% Type of measure: process The ASGE has published appropriate indications for colonoscopy (Table 3) (39). An appropriate indication should be documented for each procedure, and when it is a nonstandard indication, it should be justified in the documentation. When performing colonoscopy for average-risk CRC screening or colon polyp surveillance, endoscopists should specifically document whether the patient had a colonoscopy previously, date of the last colonoscopy (or document that the date of that procedure is not available), and any histologic findings from polyps removed during that colonoscopy.Table 3: Appropriate indications for colonoscopy (39)Discussion: In 2012, the ASGE updated its indications for endoscopic procedures (39). This list was determined by a review of published literature and expert consensus. Studies have shown that when colonoscopy is done for appropriate reasons, significantly more clinically relevant diagnoses are made (40, 41, 42). In these studies, which divided indications into appropriate, uncertain, and inappropriate and looked at high-volume European centers, 21 to 39% were classified as inappropriate. It is likely that this can be improved to a <20% inappropriate rate (43). The European Panel on the Appropriateness of Gastrointestinal Endoscopy Internet guideline is a useful decision support tool for determining the appropriateness of colonoscopy (43). The goal is to minimize the number of inappropriate procedures (44, 45, 46). 2. Frequency with which informed consent is obtained, including specific discussions of risks associated with colonoscopy, and fully documented Level of evidence: 1C Performance target: >98% Type of measure: process In addition to the risks associated with all endoscopic procedures, the consent should address the relevant and substantial adverse events pertaining to each specific colonoscopy procedure. Discussion: As with all other endoscopic procedures, consent must be obtained before the procedure from the patient or guardian (or as required by local law or per policy of the institution). It must include a discussion of the risks, benefits, and alternatives to the procedure. The most common risks of colonoscopy include bleeding, perforation, infection, sedation-related adverse events, missed lesions, and intravenous site adverse events. 3. Frequency with which colonoscopies follow recommended post-polypectomy and post-cancer resection surveillance intervals and 10-year intervals between screening colonoscopies in average-risk patients who have negative examination results and adequate bowel cleansing (priority indicator) Level of evidence: 1A Performance target: ≥90% Type of measure: process Discussion: For colonoscopy to be both effective and cost-effective and to minimize risk, the intervals between examinations should be optimized. Intervals between examinations can be effective in prevention of incident CRC only when the colon is effectively cleared of neoplasia. Therefore, detailed and effective examination of the colon, as discussed in the following, is critical to the effectiveness and safety of recommended intervals between colonoscopy. The recommended intervals assume cecal intubation, adequate bowel preparation, and careful examination. In the average-risk population (persons aged ≥50 years without other risk factors for CRC or who have only one first-degree relative with CRC and that cancer was diagnosed at age >60 years), colonoscopic screening is recommended in all past and current guidelines at 10-year intervals (15, 47, 48). A German case-control study found that a negative screening colonoscopy result was associated with >20 years of protection against colorectal cancer (49). In cohorts of average-risk persons who underwent an initial colonoscopy with a negative result, a repeat colonoscopy 5 years later had a very low yield (50, 51). Two studies of flexible sigmoidoscopy found a protective effect of endoscopy with polypectomy lasting 10 years and 16 years and could not exclude longer durations of protection (52, 53). Thus, although colonoscopy is not perfectly protective, its protective effect is prolonged. These data support the 10-year interval, but many American colonoscopists systematically perform screening colonoscopy at 5-year intervals in average-risk individuals (54). This practice is not cost-effective, exposes patients to excess risk, and cannot be justified. When performing colonoscopy for CRC screening, endoscopists should document under “indication for procedure” whether the patient previously had a colonoscopy, date of the last colonoscopy, and any histologic findings from polyps removed during that colonoscopy if that information is available. This documentation should demonstrate that colonoscopy for CRC screening or colon polyp surveillance is being performed at an appropriate interval. Evidence from surveys indicates that post-polypectomy surveillance colonoscopy in the United States is frequently performed at intervals that are shorter than those recommended in guidelines (55, 56, 57, 58, 59, 60), that knowledge of guideline recommendations is high, and lack of guideline awareness is unlikely to account for overuse of colonoscopy. Assessments of actual practice identified both overuse of surveillance examination in low-risk patients and underuse in high-risk patients (61). An assessment of Medicare colonoscopy codes demonstrated systematic overuse of colonoscopy for screening and post-polypectomy surveillance by some physicians (54). These surveys underscore the importance of measuring intervals between examinations in continuous quality improvement programs. Surgeons were more likely to use short intervals than were gastroenterologists (55), emphasizing the need for all specialties practicing colonoscopy to participate in quality programs. Primary care and other referring physicians can reasonably expect surveillance recommendations to reflect post-polypectomy surveillance recommendations or to be accompanied by an explanation indicating why the recommended interval differs from the guideline. Intervals between examinations are recommended based on the best available evidence and experience that indicates a balance between the protective effect of high-quality clearing colonoscopy with the risks and cost of colonoscopy. Intervals are determined by the numbers, size, and specific histology of precancerous lesions (15). Patients with sessile polyps >2 cm in size that are removed piecemeal have a high risk for residual polyp at the polypectomy site and require short-term follow-up at 3 to 6 months (15) and a second examination a year later to rule out a late recurrence of polyp at the site (62). Recommended post-polypectomy surveillance intervals for sessile serrated polyps (also called sessile serrated adenomas) and traditional serrated adenomas currently are based on limited evidence and will be subject to updating as new evidence appears (15). Serrated lesions include hyperplastic polyps, sessile serrated polyps, and traditional serrated adenomas. Serrated lesions, particularly the sessile serrated polyp, are considered the precursors of a substantial group of CRCs that arise predominantly in the proximal colon. At this time, consensus post-polypectomy surveillance intervals for sessile serrated polyps are similar to recommended intervals for adenomas and are based on size and number of lesions (15). Serrated lesions of all types should be counted to identify patients who meet the criteria for serrated polyposis, formerly known as hyperplastic polyposis syndrome, for which colonoscopy is recommended at 1 to 2-year intervals (15). Patients who have suspected colon bleeding after a negative colonoscopy result may require repeat examinations at intervals shorter than those recommended. However, the use of fecal occult blood testing by using guaiac-based tests for the first 5 years after a colonoscopy is inappropriate because the positive predictive value of guaiac-based fecal occult blood testing during that interval is extremely low (63). Additional study of fecal immunochemical testing for blood in this setting as an adjunct to colonoscopy is warranted (64). Colonoscopies performed for screening or surveillance at intervals shorter than those recommended in guidelines and without an appropriate explanation for the shortened interval should be considered to have an inappropriate indication. 4. Frequency with which ulcerative colitis and Crohn's colitis surveillance is recommended within proper intervals Level of evidence: 2C Performance target: ≥90% Type of measure: process Discussion: In ulcerative colitis and Crohn's colitis, surveillance refers to interval examinations in patients with long-standing disease who have undergone an initial examination in which dysplasia was not detected. The term also is used when patients who are asymptomatic are prospectively entered into interval colonoscopy programs based on the duration of disease. Surveillance does not refer to diagnostic examinations or examinations in previously diagnosed patients to assess symptoms. Both ulcerative colitis and Crohn's colitis of long duration are associated with an increased risk of colorectal cancer (65, 66). Surveillance colonoscopy often is recommended beginning 7 to 10 years after the onset of symptoms when ulcerative colitis extends beyond the rectum or in Crohn's disease when more than one third of the colon is involved. There are no randomized trials to support the effectiveness of surveillance colonoscopy in ulcerative colitis or Crohn's colitis, but case-control studies in ulcerative colitis indicate a survival benefit for patients who participate in surveillance (67, 68). Surveys of practitioners in the United States (69) and the United Kingdom (70) demonstrate that many practitioners are not familiar with surveillance recommendations, have a poor understanding of dysplasia, and make inappropriate recommendations in response to findings of dysplasia (69, 70). Patients should undergo surveillance colonoscopy, which has emerged as a standard of medical care in the United States. The onset of disease is considered to be the onset of symptoms for the purpose of initiating surveillance for both ulcerative colitis and Crohn's colitis. Because the yield of dysplasia or cancer during ulcerative colitis surveillance is relatively low and not cost-effective (71), it is important to avoid overuse of surveillance colonoscopy during the first 20 years (72). At between 7 and 20 years of disease, intervals of 2 to 3 years are generally adequate, assuming the absence of primary sclerosing cholangitis and a colon that is without severe scarring (71). Indeed, recent studies continue to indicate that the risk for CRC in chronic ulcerative colitis has been overestimated in previous decades (18, 73). Shorter intervals between examinations are indicated for patients with long-duration disease and may be initiated earlier in the course of disease in patients with established risk modifiers, such as a family history of CRC or a personal history of primary sclerosing cholangitis (71). Persons with primary sclerosing cholangitis who are discovered to have asymptomatic ulcerative colitis should begin surveillance at the time ulcerative colitis is diagnosed. Patients with endoscopically abnormal colons (eg, endoscopic scarring, pseudopolyp formation or cobblestoning, chronic severe inflammation) are at increased risk for development of cancer, compared with patients with endoscopically normal colons (74). Thus, patients with endoscopically normal colons, or with only mild abnormalities, can be triaged to longer intervals of surveillance of at least 2 to 3 years, at least during the interval from 7 to 20 years after the onset of symptoms, and assuming the absence of primary sclerosing cholangitis (74). Preprocedure research questions Why do physicians fail to follow recommended guidelines for screening and surveillance intervals? Are they concerned about missed lesions? Is there fear of litigation? What interventions will maximize adherence to guideline recommendations? Which serrated lesions in the proximal colon are clinically important? What are cost-effective intervals for follow-up after removal of sessile serrated polyps and large (>10 mm) hyperplastic polyps? Does the incidence of splenic injury during colonoscopy warrant inclusion in the informed consent process? What is the current understanding among clinicians of surveillance guidelines for ulcerative colitis and Crohn's colitis? How will new reimbursement models affect compliance with recommended surveillance intervals? Can and should surveillance interval recommendations be adjusted for endoscopists with high-level versus low-level baseline ADRs? Does the presence of 3 small adenomas warrant high-risk surveillance for endoscopists with high ADRs? Intraprocedure quality indicators Quality evaluation of the colon consists of intubation of the entire colon and a detailed mucosal inspection. Cecal intubation improves sensitivity and reduces costs by eliminating the need for radiographic procedures or repeat colonoscopy to complete the examination. Careful mucosal inspection is essential to effective CRC prevention and reduction of cancer mortality. The detection of neoplastic lesions is the primary goal of most colonoscopic examinations. Cost-benefit analyses of colonoscopy for the detection of neoplastic lesions are well within acceptable rates (approximately $20,000 per year of life saved) (75). However, adverse events, repeat procedures, and inappropriate surgical intervention for endoscopically removable polyps can reduce this benefit significantly. It is incumbent on endoscopists to evaluate their practices and make improvements wherever possible to reduce the costs associated with neoplasia detection. The intraprocedure period extends from the administration of sedation, or insertion of the endoscope when no sedation is given, to the removal of the endoscope. This period includes all the technical aspects of the procedure including completion of the examination and of therapeutic maneuvers. Common to most endoscopic procedures is the provision of sedation and need for patient monitoring (38). Intraprocedure quality indicators specific to performance of colonoscopy include the following: 5. Frequency with which the procedure note documents the quality of preparation Level of evidence: 3 Performance target: ≥98% Type of measure: process Quality of bowel preparation is based on ability to visualize the mucosa after retained stool and fluid have been suctioned away. Discussion: The endoscopist should document the quality of the bowel preparation in each colonoscopy (76, 77). Terms commonly used to characterize bowel preparation include excellent, good, fair, and poor. In clinical practice, these terms do not have standardized definitions (78). They are given standardized definitions in clinical trials of bowel preparation (79), but these trials often take into account retained fluid, which is of little interest to the examination because it can be readily suctioned. Some practitioners use the terms adequate or inadequate. The ASGE/ACG task force recommends that the examination be considered adequate if it allows detection of (within the technical limitations of the procedure) polyps >5 mm in size (80). Another option is to use independently validated preparation scores, such as the Boston Bowel Preparation Scale (81) or the Ottawa Bowel Preparation Scale (82). However, the Ottawa scale also takes into account retained material that can be removed before examination. Regardless of the scoring system used, endoscopists should document the quality of bowel preparation based on ability to identify polyps after retained fluid or stool has been suctioned. If bowel cleansing is inadequate to identify polyps >5 mm in size, and the procedure is being performed for CRC screening or colon polyp surveillance, then the procedure should be repeated in 1 year or less (15). Adequate preparation carries the implication that the recommended interval before the next colonoscopy will be consistent with guidelines (15). Poor bowel preparation is a major impediment to the effectiveness of colonoscopy. Poor preparation prolongs cecal intubation time and withdrawal time and reduces detection of both small (4) and large (4, 5, 83) polyps. In every colonoscopic practice, some colonoscopies must be repeated at intervals shorter than those recommended (15, 84) based on inadequate preparation. The economic burden of repeating examinations because of inadequate bowel preparation is substantial (6). 6. Frequency with which the bowel preparation is adequate to allow the use of recommended surveillance or screening intervals Level of evidence: 3 Performance target: ≥85% of outpatient examinations Type of measure: process We recommend that the percentage of outpatient examinations with inadequate bowel preparation that require repeat colonoscopy in ≤1 year should not exceed 15% (5). Measurement of an individual practitioner’s percentage of examinations requiring repetition because of inadequate preparation is recommended. Endoscopists who have >15% of examinations with inadequate bowel preparation should re-examine their bowel preparation protocols, including patient education, choice of purgative, and protocol for administering the purgative, including use of the split-dose protocol. Recent clinical trials of even low-volume preparations (which have lower effectiveness than 4-liter preparations) suggest that these rates of adequate preparation are readily achievable in outpatients by using split-dose preparation (85, 86). Socioeconomic factors and language barriers in some patient populations may require increased educational efforts before the colonoscopy to achieve this level of success. The most important determinant of preparation quality is the interval between the end of the preparation ingestion and the start of the procedure (87). Quality diminishes as the interval increases, and the right side of the colon is particularly affected. We recommend that all patients be prescribed split-dosing of bowel preparations, meaning that half the preparation is given on the day of the examination (87). For afternoon colonoscopies, the entire preparation can be ingested on the day of examination (88). According to fasting guidelines of the American Society of Anesthesiologists, patients should have nothing by mouth for 2 h after ingestion of clear liquids (89). We recommend that rule be followed for ingestion for split-dose and same-day preparations. This recommendation is supported by prospective observational studies that demonstrate that residual volume of liquid in the stomach is minimal (<25 ml) and similar whether patients split the bowel preparation or consume all of the bowel preparation on the evening before the procedure (90). However, because this study (90) excluded patients with gastroparesis, longer intervals may be prudent in those with conditions such as gastroparesis or achalasia (increased risk of larger volumes of retained fluid), those with central nervous system dysfunction that might be more inclined to aspirate, or in those with cardiac, pulmonary, or immunologic disease in whom a small aspiration event might be devastating. Patients should receive instruction to begin the second half of split-dose preparations 4 to 5 h before their scheduled procedure start time, and they should be finished with ingestion by at least 2 h before that time (89). Because the quality of preparation deteriorates as the preparation-to-procedure interval increases, patients scheduled in the early morning (before 9 AM) who refuse to begin ingestion 4 to 5 h before the scheduled time can begin ingestion of the second half of the preparation late on the evening before (aft