Making every colonoscopy count: Ensuring quality in endoscopy

Abstract

The last few years have seen a burgeoning interest in the measurement of clinical performance and its impact on quality standards. The advent of the National Bowel Cancer Screening Program has highlighted the many deficiencies that exist in the provision of colonoscopy services in Australia. These include the absence of an agreed tool to measure the quality of colonoscopy on either an individual, departmental or regional basis and the absence of an endoscopic or colonoscopy training curriculum or an agreed standard for colonoscopy trainers. This review will discuss the current status and shortfalls of measuring quality in colonoscopy, highlight some recent initiatives by the Gastroenterological Society of Australia and articulate a direction for the future.

Introduction

The assurance that endoscopic procedures are completed with adherence to quality standards has become important and yet, at present, there are no consensus performance indicators or national systems of benchmarking. Administrative and governmental agencies increasingly recognize the importance of quality in health care and endoscopy,¹ patients are better informed and rightly demand safety and quality; the advent of the Australian National Bowel Cancer Screening Program in August 2006 has aligned these agendas. A screening program targeting asymptomatic individuals must aspire to the highest standards in accuracy and safety. Any failure in these parameters will seriously erode the net health benefit to the target population.

A high-quality endoscopy service pays heed to all facets of the patient journey from timely access to specialist consultation, pre-procedural consent and preparation, punctuality in scheduling, the procedure and its outcomes, and appropriate follow up. It also implies that the patient receives a correctly indicated procedure, that accurate and clinically relevant diagnoses are made (or excluded), that endoscopic therapy if required is correctly performed, and that this takes place with a minimum of risk, with careful consideration of the risk–benefit equation for each individual patient. For example, the risks of a significant complication following colonoscopic polypectomy of a small right colonic adenoma in an 80-year-old being investigated for iron deficiency anemia may far exceed any potential benefit, depending on the patient's comorbidities and life expectancy. In general, endoscopy is only indicated when the information that is likely to be gained, or the therapy provided will result in a net benefit for the patient without exposing them to excessive or undue risk.²

Quality benchmarking for endoscopy should encompass assessment of both the individual endoscopist and team of clinicians involved in providing the service, the environment and department within which they work including nursing and allied health staff, endoscopic equipment, anesthetic/sedation support, the physical facility, and the technical and staff resources available in the event of an unexpected emergency. Within this framework, assessment of quality can be divided into three broad phases for each individual procedure: pre-procedural, the procedure itself and after-care post-procedure. This review will focus on colonoscopy and primarily cover areas in quality improvement that require further refinement. Many of the concepts can be applied generally to all endoscopic procedures.

Specific quality indicators for colonoscopy

Colonoscopy is widely used in the diagnosis and treatment of colonic disorders. The information it provides is pivotal in the assessment of patients with bowel symptoms, rectal bleeding, iron deficiency anemia, and positive fecal occult blood tests (FOBTs). It is a vital surveillance tool for post-polypectomy, post-cancer resection, and chronic inflammatory bowel disease patients. Colonoscopy is now accepted as the most accurate method of screening the colon for neoplasia in patients over the age of 50 years³ and colonoscopic polypectomy is proven to reduce colorectal cancer incidence by up to 90% in both average and high-risk patient groups.^4,5 Development of and adherence to evidence-based quality standards is likely to further enhance the already proven efficacy of colonoscopy in diagnosis and management of colorectal disease.

Pre-procedural indicators

Definitions of the onset of the pre-procedure period vary from first contact with a member of the endoscopy service, to the more encompassing point at which the family doctor has determined that referral for consideration of colonoscopy is required. The latter definition seems most appropriate as this time interval may also be an important quality marker and determinant of outcome in certain situations such as positive FOBT (as part of a screening program), or serious rectal bleeding, particularly in a two-tiered health system.

Pre-procedural indicators include that the endoscopic procedure is properly indicated and the patient has been appropriately selected, consented, and comprehensively assessed. Assessment implies that all necessary modifiable risk factors have been considered and addressed including management of anticoagulants and antiplatelet agents,⁶ consideration of prophylactic antibiotic administration, determination of sedation requirements,⁷ and timeliness to the performance of the procedure in respect of the clinical presentation.^8,9

When colonoscopy is appropriately indicated, studies have shown that significantly more clinically relevant diagnoses are made.^10,11 Minimizing inappropriate procedures is an important aspect of colonoscopic practice. Particularly important is the use of recommended post-polypectomy and post-cancer resection surveillance intervals. Adherence to established peer-reviewed guidelines minimizes unnecessary risk for patients and enhances cost effectiveness¹² (Fig. 1). The validity of this approach relies on effective clearing of the colon of all neoplasia and thus detailed and effective examination is critical to ensuring efficacy. Recommended intervals assume cecal intubation, adequate bowel preparation, and thorough examination on withdrawal. Surveys in the United States have shown that post-polypectomy surveillance colonoscopy is performed at intervals that are shorter than those recommended in guidelines.^13,14 Surgeons were more likely than gastroenterologists to use too short an interval,¹³ and this is also the case in Australia.¹⁵ These surveys underscore the importance of measuring intervals between examinations in continuous quality improvement programs. It is likely that minimizing over-servicing will shorten waiting lists and enhance timely access to colonoscopic services for new patients with sound indications (such as positive FOBT). Histological assessment of polyps should also inform surveillance recommendations. Small hyperplastic polyps restricted to the recto-sigmoid are of limited clinical relevance and should not influence surveillance intervals.¹⁶

Figure 1. National Health and Medical Research Council Clinical Practice Guidelines for the Prevention, Early Detection and Management of Colorectal Cancer (2005).

Risk stratification for endoscopy is an area that requires further refinement. One key is a sedation risk assessment, given that a large proportion of adverse events in endoscopy relate to sedation. Although several risk scoring systems exist, none specifically apply to endoscopy. Large prospective databases have shown that, although ostensibly crude, American Society of Anesthetists score (1 = completely healthy and 5 = critically ill and not expected to survive⁷) correlates with complications during endoscopy which are primarily sedation related.^17–19 Risk assessment for sedation then allows an individualized sedation plan to be determined. For example, in the case of an elderly patient with poor cardiorespiratory function, the use of narcotic sedatives may be avoided. In Australia, guidelines that define minimum staffing requirements, appropriate skill mix, and the necessary features of the physical facility for procedural sedation and/or analgesia for diagnostic and interventional medical and surgical procedures have recently been updated by the Tripartite Group (Gastroenterological Society of Australia, Australian and New Zealand College of Anesthetists, and Royal Australian College of Surgeons).⁷ These now represent the standard for safe endoscopic practice, to which adherence is expected.

Procedural indicators

Recognized quality indicators include quality of bowel preparation, success of cecal intubation, rates of adenoma detection, withdrawal times, and polyp retrieval rates. It is intuitively obvious, but yet unproven, that a coordinated and cohesive national approach to endoscopy information and reporting will result in a substantial increment in colonoscopy quality. Automating collection of intra-procedural quality data such as time to cecal intubation, withdrawal time, and adenoma detection rate will encourage endoscopists to achieve levels of performance above minimum quality indicators. In addition, drop-down menus that suggest appropriate surveillance intervals, such as National Health and Medical Research Council guidelines,¹² when planned follow up is being incorporated into the report, should eventually enforce adherence to these guidelines¹² (Fig. 1).

Competent endoscopists intubate the cecum in ≥ 95% of cases. All colonoscopy studies undertaken for screening report cecal intubation rates of ≥ 97%.^20–26 Photo documentation of the cecum within the colonoscopy report is a desirable standard. This can be used to verity cecal intubation rates for individual endoscopists as part of a continuous quality improvement program. In a setting of screening for asymptomatic individuals of 50 years of age and older, adenomas should be detected in ≥ 25% of men and ≥ 15% of women.

Despite its proven efficacy, studies over the last 10 years have confirmed that colonoscopy is an imperfect test. Tandem colonoscopy studies performed in the mid-1990s demonstrated, for adenomas ≥ 1 cm in size, miss rates of up to 6%, 12–13% for 6–9 mm adenomas and 15–27% for adenomas ≤ 5 mm.^27,28 Studies demonstrating variable sensitivity for adenoma and advanced neoplasia detection among endoscopists constitute the strongest evidence that sub-optimal performance is an important factor favoring colonoscopy to identify and prevent colorectal cancer. However, there is enormous variability in adenoma detection rates among colonoscopists. A private practice group of 12 US gastroenterologists performing screening colonoscopy on patients 50 years and older described an adenoma detection rate > 100 adenomas per 100 colonoscopies for the highest performer, down to < 10% of this for the poorest performer.²⁹ Persons who spent longer than 6 min detected significantly more adenomas. Others have also demonstrated a wide variability in adenoma detection rates and a positive association with longer procedure duration.³⁰

The optimal duration of the withdrawal phase with white light colonoscopy is not yet known, and whether this should simply be a time-based phenomenon or rather an algorithmic or protocolized withdrawal that includes careful inspection of critical ‘blind spots’ has yet to be determined. In the author's opinion, the future will likely see the advent of a standardized colonoscopy withdrawal protocol that specifically targets areas that may often be difficult to visualize such as the region beneath the ileocaecal valve, the cecum, sigmoid colon, and the major flexures (Table 1).

Table 1.  Suggested features of a standardized colonoscopy withdrawal protocol

Withdrawal time of ≥ 6 min and direct visualization behind all folds

Ileal intubation with careful inspection of the region immediately beneath the ileo-cecal valve, and between the valve and the appendiceal orifice

Retroflexion in the ascending colon and examination of the proximal aspect of folds up to the hepatic flexure

Two pass examination of the hepatic flexure, i.e. reinsertion of the scope beyond the flexure and an additional withdrawal examination of this area once the first withdrawal examination has been completed

Two pass examination of the splenic flexure

Two pass examination of the sigmoid colon

Retroflexion in the rectum and inspection of the anorectal junction and distal rectum

Change of position to enhance visualization in different regions of the colon may also become part of the withdrawal protocol as further scientific evidence comes to hand; for example, the supine position may provide superior views of the transverse colon.

Standardized documentation of bowel preparation quality is yet to be achieved, but poor bowel preparation is a major impediment to the effectiveness of colonoscopy. It prolongs intubation and withdrawal time and reduces detection of both small and large polyps.^31,32 The economic burden of repeating examinations because of inadequate bowel preparation is substantial. Practitioners should record prospectively the number of repeated examinations that are required because of inadequate bowel preparation. An acceptable level of this outcome has not yet been determined; however, the responsibility largely rests with the endoscopist and their ancillary staff.

Currently unanswered questions include: how can one improve the performance of physicians with either low adenoma detection rates or low cecal intubation rates? Are there teaching methods that will enhance performance?

Post procedure

The post-procedure period commences at the completion of the procedure and extends to subsequent follow up. It encompasses post-procedural instructions, documentation of the procedure, follow up of pathology, and recognition and documentation of complications. It is critical that any endoscopy service has an established system by which all patients, but particularly high-risk patients, are contacted at a defined time interval after their procedure to determine whether any delayed complications have occurred. A system by which these are reported and evaluated should be in place so that repetitive or systematic errors can be discovered and rectified.

Colonoscopy-related perforation may not be apparent until several days after the patient has left the facility. There is a small incidence of mortality; older data suggesting this may be up to 5%.³³ Considering all the available data, and including therapeutic procedures, colonoscopy perforation rates should be no greater than 1:1000 procedures. Major complications such as perforation mandate a detailed review of the circumstances around the event to determine whether any modifiable risk factors existed so that the problem can be avoided in the future, such as incorrect diathermy settings for endoscopic mucosal resection in the right colon.^34,35 Individual departments should undertake prospective data collection and careful clinical follow up to identify major complications or patterns of adverse events. These then can be logged in an endoscopic complications registry. These data can then be used in a regular audit of outcomes and in morbidity analysis meetings as part of the continuous quality improvement program. Only with a complete understanding of the entire procedure and its clinical outcomes on both an individual and departmental level are incremental gains in safety and quality likely to occur.

Improving outcomes in complex and advanced lesions

There is now established evidence that all gastrointestinal mucosal neoplasia without submucosal invasion can be treated effectively by endoscopic means, irrespective of its location in the gastrointestinal tract.^36–39 This should be the preferred treatment approach with surgery reserved for failures. Advanced, complex, or large sessile lesions generally require subspecialty endoscopic management to achieve the optimal clinical outcome. Referral pathways for endoscopic management of complex colonic lesions are yet to be formally established, although anecdotally they exist. Large colonic polyps greater than 20–30 mm in maximum dimension, particularly sessile lesions are relatively infrequent, so they represent an uncommon challenge for individual endoscopists when they are encountered. They require advanced skills, additional equipment, and procedural time, and thus should be managed by a team of specialists who deal with this type of pathology regularly. A separate more detailed consent process is also necessary. When properly assessed by the referring endoscopist, very few of these lesions will be found to have invasive neoplasia and the majority can be safely removed by a tertiary referral polypectomy and mucosal resection service.^40–42

The few published data that exist indicate that many benign adenomas detected during colonoscopy, if large, will be treated surgically. This carries associated expense, morbidity, and potential mortality in patients who are frequently elderly with significant comorbidity. For example, in the first regional publication for the French National Bowel Cancer Screening Program, more than 10% of benign polyps were treated surgically;⁴³ this is difficult to understand or justify. Large US studies have shown that 30-day mortality from colectomy is approximately 5%, and similar mortality data have been reported for colonic surgery in Australia.^44,45 The mortality rate is largely independent of hospital volume; it is 4.5% for high-volume centers (> 124 procedures/year), and 5.6% for low-volume centers (< 33 procedures/year). This modest reduction in mortality with greater surgical volume is at variance with the results for more complex surgery, such as oesophagectomy and pancreatic resection, where 30-day mortality in the low-volume centers was three to five times higher (at up to 20%) than that recorded for the high-volume centers.⁴⁴ Thus, for colectomy, mortality is largely independent of unit work volume. This is consistent with patient-related factors being the primary determinant of outcome in patients undergoing colectomy. Clearly patients with difficult, yet benign, mucosal pathology require the least invasive intervention that will cure their problem. Such an approach is likely to yield substantial cost savings and clinical gains.

A large Australian study has shown that when difficult or advanced lesions are managed by a tertiary referral team, as compared with surgery, substantial cost savings in the order of AUD11 000 per patient can be realized with limited morbidity and no mortality. In this study, 174 patients with difficult refractory polyps, with a mean size of 30 mm, were referred by 52 independent specialist endoscopists and managed on an outpatient basis. Procedural success was 95% and ultimately in 90% of instances this avoided the need for surgery. A total of 20 bed-days were utilized due to endoscopic complications, but there were no perforations. The total cost saving for the entire cohort was 1.87 million AUD.⁴¹

In the management of complex lesions this tertiary referral approach delivers major cost savings and avoids the protracted recovery and potential complications of colonic surgery for the majority of patients. This type of clinical pathway should be developed to enhance patient outcomes and reduce health-care costs. Regional health services and governmental agencies should examine ways to formalize such pathways.

Lesion assessment

Large sessile or flat lesions of the colon are best described according to the Paris system of endoscopic classification of superficial neoplastic lesions.^46,47 This classification system is now widely used in Japan. Application of a widely accepted reproducible classification has numerous advantages, including standardizing descriptions of polyp morphology and improved lesion categorization. By improving lesion categorization, a better understanding of variations in their characteristics and biology is achieved, and lesion-specific treatment and surveillance algorithms can be developed.

Within the Paris system, polypoid lesions may be pedunculated (0-Ip), sessile (0-Is), or have a mixed pattern (0-Isp). Non-polypoid lesions are either slightly elevated, termed 0-IIa (elevation < 2.5 mm above the level of the mucosa), completely flat (0-IIb), or (uncommonly) be slightly depressed (0-IIc). In addition, 0-IIa and 0-IIb lesions > 10 mm in diameter with a low vertical axis but extending laterally along the interior luminal wall are termed laterally spreading tumors. The latter are further subclassified according to their surface morphology as either granular (G) (Figs 2–5) or smooth non-granular (NG) (Figs 6,7), a subclassification that has important implications for the risk of submucosal invasion. Granular lesions, particularly of 0-IIa morphology, rarely contain invasive malignancy, irrespective of size, while non-granular lesions have a 15–30% risk of submucosal invasion which is size dependent.⁴⁸ Currently western endoscopists tend to group 0-Is, 0-Isp, and 0-IIa lesions together and simply describe them as ‘sessile’. 0-IIb lesions may also be included in this group, or described as being ‘flat’. Little attention is paid to the surface morphology, although large granular laterally spreading tumors may be described as ‘carpet lesions’. This sort of simplification and unified grouping of all these different types of lesions with their differing tumor biology has retarded progress in understanding in the West.⁴⁹ Other valuable assessment techniques such as magnification with pit pattern analysis, which has been shown to be a reliable predictor of submucosal invasion, are also not commonly used.⁵⁰

Figure 2. Large granular (80 mm maximum dimension) laterally spreading tumor of the descending colon—Paris classification 0-IIa.

Figure 3. Figure 2 after endoscopic mucosal resection.

Figure 4. Large 50-mm granular laterally spreading tumor of the cecum and ileo-cecal valve—Paris classification 0-IIa.

Figure 5. Lesion in Figure 4 after endoscopic mucosal resection.

Figure 6. 30-mm non-granular laterally spreading tumor of the distal transverse colon—Paris classification 0-IIa—invasive malignancy found after resection.

Figure 7. 35-mm non-granular laterally spreading tumor of the distal transverse colon—Paris classification 0-IIa—invasive malignancy found after resection.

Improving lesion description can be achieved by endoscopy-reporting tools that contain drop-down menus, and educational tools that allow endoscopists to choose the appropriate morphology for the lesion they are reporting. Mandating adherence to minimum descriptive standards (e.g. Paris classification) is likely to have a substantial impact on the quality of clinical decision-making. For example, Paris 0-IIa lesions, even if > 50 mm (often described as ‘carpet lesions’ outside of Japan), rarely contain invasive malignancy in the absence of a 0-Is component.^40,48 These patients should be managed endoscopically, although currently outside of specialist centers they are frequently referred for surgery.

Training and re-certification

Underpinning any national quality initiative is the assurance that young endoscopists will be appropriately trained. To achieve this, those teaching them must have the necessary endoscopic knowledge base and teaching competencies to impart their technical expertise and skills to the trainee endoscopist. In contrast to contemporary practice, it is the author's view that the acquisition of cognitive skills in endoscopy far exceeds the importance of acquiring technical skills. Currently, too much teaching is passive and not deconstructed. The trainee needs to be guided through a careful progressive stepwise acquisition of basic endoscopic skills and their variations. Conscious competence among trainers is a critical aspect. Those teaching endoscopy and colonoscopy must be able to precisely articulate both the cognitive aspects and the technical elements of fundamental and more advanced technique, and the reasons why various approaches are adopted. Currently, trainees are allowed to ‘practice on patients’ with limited input from the trainer. This ultimately results in a ‘trained’ endoscopist who can get to the cecum and back, but is unable to offer any real technical insight on how this was achieved or what was encountered along the way. This type of endoscopist will be a poor teacher and will struggle when difficult endoscopic situations arise.

The Gastroenterological Society of Australia and through its subsidiary, The Australian Gastrointestinal Endoscopy Association, have undertaken to correct these deficiencies in the current endoscopic training program. A full discussion of the initiatives is beyond the purview of this paper; however, the approach is multi-facetted and comprehensive. In brief, coordinated national training initiatives have been developed for both endoscopy trainees (both medical and surgical) and those involved in teaching endoscopy. The former has been branded as the National Endoscopy Training Initiative. It commenced in 2006 and involves a series of standardized small group workshops with hands-on exposure, dealing with the fundamentals of colonoscopy, including optimal insertion and withdrawal techniques, the basic principles of endoscopic electrosurgery, basic polypectomy techniques, and overcoming common cecal intubation challenges such as severe stenosing diverticular disease. For the colonoscopy trainers, an interactive trainee-focused small group workshop curriculum entitled ‘Train the Colonoscopy Trainer’ has been developed. Ultimately, it is envisaged that all major endoscopy teaching centers will have a member of staff designated as the Director of Endoscopic Training, that dedicated endoscopy training lists will be developed within individual departments, and that all those involved in teaching of endoscopy in Australia will attend a ‘Train the Trainer’ course regularly.

As a foundation to these initiatives, an endoscopy curriculum has been developed entitled ‘A Primer on Gastrointestinal Endoscopy’. This syllabus is intended to introduce new endoscopy trainees to the core endoscopic knowledge that is required to train as a gastroenterologist, and to provide a framework around which they can develop their endoscopic knowledge base with the assistance of their mentors, core texts, and the literature.

Ongoing accreditation is an important aspect of providing a high-quality endoscopy service at a national level. The development of a formal process to ensure the ongoing accredited competency of proceduralists through a nationally recognized mandatory certification and eventual re-certification system is required. A national endoscopy reporting system and information tool would make such a process seamless. Agreed minimum quality standards could be routinely documented for all endoscopic proceduralists throughout the country. This could easily link with certification and training processes through a national endoscopy information system.

Pulling it all together: a national endoscopy reporting system and information tool

The advent of the National Bowel Cancer Screening Program has highlighted the numerous deficiencies that exist in the reporting of quality indicators for colonoscopy in Australia. Although colonoscopy services in Australia in both the public and private sector either directly or indirectly are primarily funded by the Federal Government, a coordinated national system to report on activity, productivity, and quality indicators does not exist. At individual institutions a number of diverse and disparate endoscopy reporting systems are currently in use and these range from the handwritten to site-specific electronic systems. The range of data captured by these diverse systems is neither uniform, nor standardized, and nor is it reported in any meaningful way to any responsible body.

In the absence of a coordinated national reporting system, meaningful assessments against critical safety, quality, productivity, and accuracy targets and benchmarks are not possible. In 2005/2006 at least 445 000 colonoscopies were performed in Australia at a cost of $1259 per procedure,^41,51 in a population of less than 22 000 000, placing the nation among the highest colonoscopy per capita rates in the world. Annual colonoscopy volume continues to grow at a rate of approximately 8%, far outstripping population growth. The indications for, or the outcomes of these nearly half a million procedures are currently largely unknown on a national basis. The financial burden of colonoscopy services in Australia in the 12 months of 2005/2006 was at least AUD560 000 000.

In a time of scarce health funding and resources, it is absolutely vital that each health dollar is spent wisely. If a national endoscopy information system and reporting tool is not developed, then no meaningful analysis of this burgeoning national economic burden can be undertaken. It could be that the majority of colonoscopy undertaken in Australia is appropriate, well targeted and of high quality, but without a coordinated national reporting system and information tool we will never know.