Performance measures for lower gastrointestinal endoscopy: a European Society of Gastrointestinal Endoscopy (ESGE) quality improvement initiative

Abstract

The European Society of Gastrointestinal Endoscopy and United European Gastroenterology present a short list of key performance measures for lower gastrointestinal endoscopy. We recommend that endoscopy services across Europe adopt the following seven key performance measures for lower gastrointestinal endoscopy for measurement and evaluation in daily practice at a center and endoscopist level: 1 rate of adequate bowel preparation (minimum standard 90%); 2 cecal intubation rate (minimum standard 90%); 3 adenoma detection rate (minimum standard 25%); 4 appropriate polypectomy technique (minimum standard 80%); 5 complication rate (minimum standard not set); 6 patient experience (minimum standard not set); 7 appropriate post-polypectomy surveillance recommendations (minimum standard not set). Other identified performance measures have been listed as less relevant based on an assessment of their importance, scientific acceptability, feasibility, usability, and comparison to competing measures.

Introduction

The European Society of Gastrointestinal Endoscopy (ESGE) and United European Gastroenterology (UEG) have identified quality of endoscopy as a major priority. We described our rationale for this priority in a recent manuscript that also addressed the methodology of the current quality initiative process.¹

Because of the variation in physicians’ performance and the introduction of nationwide colorectal cancer (CRC) screening programs, lower gastrointestinal (LGI) endoscopy was the first area of endoscopy to address quality.^2–4 Over more than a decade, several potential measures of quality in LGI endoscopy have been identified. In consequence, many professional societies have published recommendations on performance measures for LGI endoscopy.^5–7 These recommendations are, however, numerous (44 different performance measures),^5–7 country specific, and not always evidence based, which has limited their wider adoption in Europe.

The aim of the ESGE LGI working group was to identify a short list of key performance measures for LGI endoscopy that were widely applicable to endoscopy services throughout Europe. This list would ideally consist of performance measures with the following requirements: proven impact on significant clinical outcomes or quality of life; a well-defined, reliable, and simple method/approach for measurement; susceptibility for improvement; and application to all levels of endoscopy services.

This paper reports the agreed list of key performance measures for LGI endoscopy and describes the methodological process applied in the development of these measures.

Methodology

We previously described the multistep process for producing such performance measures.¹ In brief, at the United European Gastroenterology Week in 2014, we used a modified Delphi consensus process to develop quality measures in the following domains: pre-procedure, completeness of procedure, identification of pathology, management of pathology, complications, procedure numbers, patient experience, and post-procedure.^1,8,9 We decided to have one or two key performance measures for each quality domain.

In order to identify key performance measures, we first created a list of all possible performance measures for LGI endoscopy through email correspondence and teleconferences that took place between December 5, 2014 and February 7, 2015. All possible performance measures that were identified by this process were then structured using the PICO framework (where P stands for population/patient; I for intervention/indicator; C for comparator/control, and O for outcome) to inform searches for available evidence to support the performance measures. This process resulted in 38 PICOs. Detailed literature searches were performed by an expert team of methodologists and yielded results for 29 PICOs (see Supporting information, available online). Working group members also identified additional articles relevant for the performance measures in question.

The PICOs and the clinical statements derived from these were adapted or omitted during iterative rounds of comments and suggestions from the working group members during the Delphi process. The evolution and adaptation of the different PICOs and clinical statements during the Delphi process can be reviewed in the Supporting information. The domain addressing the competence of endoscopists’ quality (including procedure numbers), along with its associated PICOs and clinical statements, was moved for future initiatives.

In total, working group members participated in a maximum of three rounds of voting to agree on performance measures in predefined domains and their respective thresholds, as discussed below. Statements were discarded if agreement was not reached over the three voting rounds. The agreement that is given for the different statements refers to the last voting round in the Delphi process. The key performance measures were distinguished from the minor performance measures based on the ISFU criteria (importance, scientific acceptability, feasibility, usability, and comparison with competing measures), and expressed by mean voting scores.

The performance measures are displayed in boxes under the relevant quality domain. Each box describes the performance measure, the level of agreement during the modified Delphi process, the grading of available evidence (the evidence was graded according to the Grading of Recommendations Assessment, Development and Evaluation [GRADE] system),¹⁰ how the performance measure should be measured, and recommendations supporting its adoption. The boxes further list the measurement of agreement (scores), the desired threshold, and suggestions on how to deal with underperformance.

The minimum number needed to assess whether the threshold for a certain performance measure is reached can be calculated by estimating the 95% confidence intervals (CIs) around the predefined threshold for different sample sizes.^8,9,11 For the sake of practicality and to simplify implementation and auditing, we suggest that at least 100 consecutive procedures (or all, if <100 performed) should be measured to assess a performance measure. Continuous monitoring should however be the preferred method of measurement.

Performance measures for LGI endoscopy

The evidence derived by the literature search group and input from the working group members were used to formulate a total of 34 clinical statements addressing 27 potential performance measures grouped into eight quality domains. Over the course of two voting rounds, consensus agreement was reached for 18 statements regarding 14 potential performance measures (agreement in both voting rounds). The remaining 16 statements were again rephrased and subjected to a third and final voting round, with a further four statements being accepted. In total, 22 statements regarding 18 performance measures were accepted after three voting rounds. Over the course of voting, we decided that the quality domain on competence of endoscopists (including three accepted statements and three performance measures) would be discarded from these guidelines and left for future initiatives. Therefore, a final total of 15 performance measures (19 statements) attributed to seven quality domains were accepted for these guidelines (see Figure 1). The entire process of performance measure development can be reviewed in the Supporting information. The statement numbers correspond to those used in Supporting Information.

Figure 1.

The domains and performance measures chosen by the working group (N/A = not available).

We used the highest mean voting scores to identify one key performance measure for each of the seven quality domains (Figure 1). The remaining performance measures were considered minor performance measures. In the management of pathology domain, there were two performance measures (“appropriate polypectomy technique” and “tattooing resection sites”) that had similar voting scores. We decided to select “appropriate polypectomy technique” as the key performance measure for this domain, based on its wider usability and better feasibility.

All performance measures were deemed valuable by the working group members and were obtained after a rigorous process, as described above. From a practical viewpoint, it may however be desirable to implement the key performance measures first in units that are not monitoring any performance measures at this time. Once a culture of quality measurement (with the aim of improving practice, outcomes, and patient experience) is accepted and software is available, the minor performance measures may then further aid the monitoring of quality in LGI endoscopy. The use of appropriate endoscopy reporting systems is key to facilitate data retrieval on identified performance measures.¹²

All of the performance measures are presented below using the descriptive framework developed by the Quality Improvement Committee (QIC) and a short summary of the evidence for the ISFU criteria. The performance measures are listed according to the domain to which they were attributed (for a summary, see Figure 1).

• 1. Domain: Pre-procedure

  Key performance measure	Rate of adequate bowel preparation
  Description	The percentage of patients with an adequately prepared bowel
  Domain	Pre-procedure
  Category	Process
  Rationale	It has been shown that the quality of bowel preparation affects the rates of cecal intubation and adenoma detection.
  	Inadequate bowel preparation results in increased costs and inconvenience as the examination has to be rescheduled or alternative investigations have to be organized.
  Construct	Denominator: Patients undergoing colonoscopy
  	Numerator: Patients in the denominator with adequate bowel preparation (assessed with a validated scale, preferably the Boston Bowel Preparation Scale (BBPS; score ≥6), Ottawa Scale (score ≤7), Aronchick Scale (excellent, good, or fair))
  	Exclusions: Emergency colonoscopies
  	Calculation: Proportion (%)
  	Level of analysis: Service and individual level
  	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a yearly audit of a sample of 100 consecutive LGI endoscopies
  Standards	Minimum standard: ≥90%
  	Target standard: ≥95%
  	Bowel preparation quality, assessed using a validated scale, such as the BBPS, the Ottawa Scale, or the Aronchick Scale, should be included in every colonoscopy report.
  	If the minimum standard is not reached, analysis of the factors influencing bowel preparation should be performed on a service level (information given to patients, dietary restrictions, cleansing agent used, colonoscopy timing).
  	After evaluation and adjustment, close monitoring should be performed with a further audit within 6 months.
  Consensus agreement	100%
  PICO	1.1–1.2 (see Supporting Information)
  Evidence grading	Moderate quality evidence

The acceptance of this performance measure is based on agreement with the following statements:

• In patients undergoing screening or diagnostic colonoscopy, bowel preparation quality should be recorded using a validated scale with high intraobserver reliability. (Statement number N1.1) Agreement: 100%

• A service should have a minimum of ≥90% procedures and a target of ≥95% procedures with adequate bowel preparation, assessed using a validated scale with high intraobserver reliability. (N1.2) Agreement: 100%

The quality of bowel preparation is important for the efficacy of colonoscopy. As pointed out in the ESGE guidelines on bowel preparation for colonoscopy,¹³ the quality of bowel preparation is associated with two other important performance measures for colonoscopy, namely adenoma detection rate (ADR) and cecal intubation rate.¹⁴ Suboptimal bowel preparation results in further costs and inconvenience because the examination has to be repeated or an alternative examination has to be arranged.¹⁵

To determine the scientific acceptability of measuring bowel preparation quality, we focused on the performance of different bowel preparation scales and the quantification of adequacy of bowel preparation. There were no direct comparisons of performance between the bowel preparation scales (see Supporting information). Three bowel preparation scales have undergone comprehensive validation and have shown sufficient validity and reliability: the Boston Bowel Preparation Scale (BBPS),¹⁶ the Ottawa Scale,¹⁷ and the Aronchick Scale.¹⁸ The BBPS is the most thoroughly validated scale and should be the preferred one.¹⁹ There were no significant differences between intermediate and high quality bowel preparation (regardless of the scale used) in terms of the detection rates for adenomas or advanced adenomas (see Supporting information).²⁰ Therefore, adequate bowel preparation may be defined as: BBPS ≥6, Ottawa Scale ≤7, or Aronchick Scale excellent, good, or fair. The adoption of validated scales for bowel preparation quality assessment has been proven to be feasible in routine practice.²¹

The proposed minimum (≥90%) and target standard (≥95%) rates of adequate bowel preparation were based on values reported in recent population-based studies,^22–24 and on randomized clinical trials of split-dose bowel cleansing regimens,^25,26

Minor performance measure	Time slot allotted for colonoscopy
Description	Time allotted for each colonoscopy in daily schedule
Domain	Pre-procedure
Category	Structure
Rationale	Colonoscopy needs adequate time allocated for the entire procedure (including discussion with the patient, sedation, insertion, withdrawal, and therapy).
	Time pressure due to inadequate time slots may impair colonoscopy quality.
Construct	Denominator: Number of colonoscopies scheduled in an outpatient colonoscopy list (session)
	Numerator: Outpatient colonoscopy list (session) working hours
	Exclusions: Emergency colonoscopy
	Calculation: Average time length (minutes)
	Level of analysis: Service level
	Frequency: Two-yearly check of booking log
Standards	Minimum standard: 30 min for clinical and primary screening colonoscopy; 45 min for colonoscopy following positive fecal occult blood testing
	Target standard: no target standard set
	If the minimum standard is not reached, a systematic approach to schedule modification should be applied.
Consensus agreement	100%
PICO	1.3 (see Supporting Information)
Evidence grading	No evidence

The acceptance of this performance measure is based on agreement with the following statement:

• Colonoscopy needs adequate time allocated for insertion, withdrawal, and therapy. Routine colonoscopy should be allocated a minimum of 30 min. Colonoscopies following positive fecal occult blood testing should be allocated a minimum of 45 min to allow for therapeutic intervention. (N1.3) Agreement: 100%

There is some evidence that productivity pressure may negatively affect the quality of colonoscopy.²⁷ Although it has been shown that working behind schedule is not associated with lower ADRs,²⁸ the effect of a very tight schedule on colonoscopy performance is unknown (see Supporting information). The working group members suggested that 30 min and 45 min are minimum times that should be allotted for routine colonoscopy and colonoscopy after positive fecal occult blood testing (longer time to accommodate high prevalence of large polyps), respectively. These values correspond well with mean total procedure times for colonoscopy reported in recent studies.^29,30

The acceptance of this performance measure is based on agreement with the following statement:

• For audit purposes, the colonoscopy report should include an explicit indication for the procedure, categorized according to existing guidelines on appropriateness of colonoscopy use. (N1.4) Agreement: 93.8%

Appropriate referrals for colonoscopy may help to optimize the use of limited resources and protect patients from the potential harms of unnecessary invasive procedures. Colonoscopies with an appropriate indication are associated with significantly higher diagnostic yields for cancer and other relevant lesions than colonoscopies without an appropriate indication.^31–34 The American Society for Gastrointestinal Endoscopy (ASGE) and the European Panel on the Appropriateness of Gastrointestinal Endoscopy (EPAGE) II guidelines, on the appropriateness of colonoscopy use,^35,36 consistently show 67%–96% sensitivity and 13%–40% specificity for the detection of relevant findings (see Supporting information).^31–34

The proposed minimum standard of appropriate indication for colonoscopy (≥85%) was based on values achieved in studies from academic and non-academic centers over the last 5 years.^32,33,37 The use of appropriate endoscopy reporting systems with a drop-down menu for indication is key to facilitate data acquisition for this performance measure.¹²

Minor performance measure	Indication for colonoscopy
Description	The colonoscopy report should include an explicit indication for the procedure, categorized according to existing guidelines on appropriate use of colonoscopy (the ASGE or the EPAGE II guidelines)
Domain	Pre-procedure
Category	Process
Rationale	Colonoscopies with an appropriate indication are associated with higher diagnostic yield for relevant lesions than colonoscopies without an appropriate indication.
Construct	Denominator: All colonoscopies performed
	Numerator: Colonoscopies with appropriate and ‘uncertain’ indication (according to ASGE or EPAGE II)
	Exclusions: None
	Calculation: Proportion (%)
	Level of analysis: Service level
	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a yearly audit of a sample of 100 consecutive LGI endoscopies.
Standards	Minimum standard: ≥85%
	Target standard: ≥ 95%
	All reports from colonoscopies performed should include an appropriate indication according to the ASGE or EPAGE II guidelines.
	When performed for screening, the colonoscopy report should state this and it must be ensured that the subject meets the criteria for screening.
	A colonoscopy reporting system with a drop-down menu for indication is ideal to ensure proper recording of the indication and later auditing.
	If the minimum standard is not met, a systematic approach to validate the appropriateness of colonoscopies should be applied (i.e. validation of appropriateness before colonoscopy scheduling).
	After evaluation and adjustment, close monitoring should be performed with a further audit within 6 months.
Consensus agreement	93.8%
PICO	1.4 (see Supporting Information)
Evidence grading	Moderate quality evidence

• 2. Domain: Completeness of procedure

  Key performance measure	Cecal intubation rate
  Description	The percentage of colonoscopies reaching and visualizing the whole cecum and its landmarks
  Domain	Completeness of procedure
  Category	Process
  Rationale	Whole bowel examination is a prerequisite for complete and reliable inspection of the mucosa in search of lesions.
  	A low cecal intubation rate is associated with an increased risk of interval colorectal cancer.
  	Incomplete colonoscopy leads to increased costs and inconvenience as the examination has to be repeated.
  Construct	Denominator: All screening or diagnostic colonoscopies
  	Numerator: Procedures in the denominator that report reaching the cecum (documented in written form and by photo/video)
  	Exclusions:
  	– Therapeutic procedures with no indication to reach the cecum
  	– Emergency colonoscopies
  	Calculation: Proportion (%)
  	Level of analysis: Service and endoscopist level
  	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a yearly audit of a sample of 100 consecutive LGI endoscopies.
  Standards	Minimum standard: ≥90%
  	Target standard: ≥95%
  	Cecal intubation, meaning complete visualization of the whole cecum and its landmarks, should be documented in a written report, as well as with photo or video documentation.
  	If the minimum standard is not reached for an individual endoscopist, additional training should be offered.
  	If the minimum standard is not reached on a service level, an audit to determine the cause should be performed.
  	After evaluation and adjustment, close monitoring should be performed with a further audit within 6 months.
  Consensus agreement	97.9%
  PICO	2.1–2.3 (see Supporting Information)
  Evidence grading	Moderate quality evidence

The acceptance of this performance measure is based on agreement with the following statements:

• Complete colonoscopy requires cecal intubation with complete visualization of the whole cecum and its landmarks. (N2.1) Agreement: 100%

• A service should have a minimum unadjusted cecal intubation rate of ≥90% and a target rate of ≥95% as a measure of the completeness of colonoscopy examination. (N2.2) Agreement: 93.8%

• Complete colonoscopy (cecal intubation) should be documented both in written form and in a photo or video report. (N2.3) Agreement: 100%

Cecal intubation is a prerequisite for complete visualization of the colorectum. Cecal intubation must be confirmed with photo or video documentation. Clear cecal image documentation is associated with a higher polyp detection rate (PDR).³⁸ For the purpose of colorectal neoplasia detection, terminal ileum intubation is useful only to confirm completion of the colonoscopy when classic cecal landmarks are not confidently seen.³⁹

Failed cecal intubation results in further costs and inconvenience as the examination must be rescheduled or an alternative investigation organized. A cecal intubation rate <80% is associated with significantly higher risks of proximal and distal interval CRCs when compared with higher completion rates.⁴⁰ Adjustment of the cecal intubation rate for inadequate bowel preparation or impassable strictures makes the measurement less feasible and harbors the risk of gaming. In recent large population-based studies, unadjusted cecal intubation rates always exceeded 90% and were usually above 95%.^22,41–45 The effect of raising the target standard beyond the minimum of 95% is uncertain.

• 3. Domain: Identification of pathology

  Key performance measure	Adenoma detection rate (ADR)
  Description	Percentage of colonoscopies with at least one adenoma identified
  Domain	Identification of pathology
  Category	Process
  Rationale	ADR reflects adequate inspection of the bowel mucosa.
  	ADR is associated with interval CRC and CRC death, with improvement in the ADR lowering the risk for CRC and CRC death.
  Construct	Denominator: All colonoscopies in patients aged 50 years or older
  	Numerator: Procedures in the denominator in which at least one adenoma was identified
  	Exclusions:
  	– Emergency colonoscopy
  	– Endoscopy with a specific therapeutic indication, including work-up of a previously detected lesion or follow-up of disease activity in inflammatory bowel disease
  	Calculation: Proportion (%)
  	Level of analysis: Service and endoscopist level
  	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a yearly audit of a sample of 100 consecutive LGI endoscopies.
  Standards	Minimum standard: ≥25%
  	Target standard: no current target standard defined
  	ADR should be monitored in all settings (screening and out-patient), which requires routine access to histopathology reports.
  	If the minimum standard is not met by an individual endoscopist, appropriate feedback followed by a competence assessment (with special consideration of withdrawal time and technique) should be given.
  	If the minimum standard is not met on a service level, comprehensive training for the center leader should be considered.
  Consensus agreement	100%
  PICO	3.1–3.4 (see Supporting Information)
  Evidence grading	Moderate to high quality evidence

The acceptance of this performance measure is based on agreement with the following statement:

• Adenoma detection rate should be used as a measure of adequate inspection at screening or diagnostic colonoscopy in patients aged 50 years or more. (N3.1) Agreement: 100%

The detection and removal of adenomas, which are major precursor lesions for CRC, is seen as a key aspect of CRC prevention. However, there is a wide variation between endoscopists in terms of their skills at detecting adenomas, expressed as the ADR.^{22,43,46–48} ADR has been inversely associated with the risk of interval CRC and CRC death.^46,47 A similar relationship with the incidence of distal interval CRC was confirmed for flexible sigmoidoscopy screening.⁴⁹ Of note, the detection rate of serrated polyps has been shown to strongly correlate with the ADR.⁴³ Although ADR is considered a surrogate for meticulous inspection of the colorectal mucosa, the correlation with other important, but non-neoplastic, findings has never been studied.

Several interventions, including education, creating awareness, feedback, and benchmarking on colonoscopy quality, have all helped to improve the ADR.^50–53 Recently, it has been shown that an improved ADR translates to risk reductions for interval CRC and death, which closes the quality improvement loop.⁵⁴

It has been postulated that ADR has an inherent limitation of not measuring the total number of adenomas detected.⁴¹ A potentially more accurate measure, namely number of adenomas per colonoscopy, has been proposed, but this was proven not to be superior to ADR in a recent study.⁵⁵

It is challenging to set the standards for ADR, especially in populations enriched with fecal occult blood test (FOBT)-positive patients. In a primary colonoscopy screening setting, a 1% increase in ADR predicted a 3% decrease in the risk of interval CRC within the observed ADR range 7.35%–52.5%.⁴⁷ In another study, an ADR above 24.6% was associated with a reduced risk of interval CRC and subsequent death.⁵⁴ In recent population-based studies, a proposed minimum standard ADR of 25% was met by the majority of endoscopists.^22,47,51 In fecal immunochemical test (FIT) positive-enriched populations, the minimum standard may need to be higher; however, the exact value is yet to be established.

Minor performance measure	Withdrawal time
Description	Time spent on withdrawal of the endoscope from cecum to anal canal and inspection of the entire bowel mucosa at negative (no biopsy or therapy) screening or diagnostic colonoscopy
Domain	Identification of pathology
Category	Process
Rationale	A mean withdrawal time of 6 min or longer was associated with higher ADRs and lower interval cancer rates as compared to shorter withdrawal times.
Construct	Withdrawal time is measured from cecum to anal sphincter
	Denominator: Number of negative (no biopsy/therapy) screening or diagnostic colonoscopies
	Numerator: Sum of withdrawal time in colonoscopies included in the numerator
	Exclusions:
	– Emergency colonoscopy
	– Incomplete colonoscopy
	Calculation: Mean time in minutes
	Level of analysis: Endoscopist level
	Frequency: Measured only if the ADR is insufficient, using a sample of 100 consecutive colonoscopies
Standards	Minimum standard: mean 6 min
	Target standard: mean 10 min
	Time can be measured by different methods: stopwatch operated by a nurse, time stamp on photodocumentation of the cecum and rectum, length of video recording, or external device (this requires inclusion of the withdrawal time in the colonoscopy report).
	Withdrawal time should be measured only when the ADR is insufficient.
	Feedback on mean withdrawal time should be given to endoscopists.
Consensus agreement	87.5%
PICO	3.6 (see Supporting Information)
Evidence grading	Moderate quality evidence

The acceptance of this performance measure is based on agreement with the following statement:

• A mean withdrawal time of at least 6 min should be used as a supportive measure of adequate identification of pathology at negative screening or diagnostic colonoscopy. (N3.6) Agreement: 87.5%

Colonoscope withdrawal time provides information about the time that endoscopists spend identifying pathology. A mean withdrawal time of >6 min has been associated with higher ADRs.⁵⁶ Although the association between withdrawal time and ADR was not observed in all studies,⁵⁷ a recent large population-based analysis confirmed the positive relation between these two measures, with a 3.6% absolute increase in ADR per minute increase in withdrawal time.²⁴ Importantly, the latter study also showed an inverse association between mean withdrawal time and the incidence of interval CRC.²⁴ The observed association was not linear and the risk of interval CRC leveled off at a mean withdrawal time of 8 min (the most significant difference was observed for the 6-minute cut-off). In another study, an increase in mean withdrawal time beyond 10 min had minimal effect on ADR.⁵⁸ Therefore, the minimum standard mean withdrawal time of 6 min and the target standard of 10 min are quite well defined.

Monitoring withdrawal time or institution policy on withdrawal time above a certain threshold showed inconsistent effects on ADRs.^59–61 The explanation could be that the variation in withdrawal technique is more important than the withdrawal time.⁶² Therefore, it appears that the withdrawal time is particularly useful as a supportive tool when the observed ADR is less than the minimum standard of 25%.⁶³

Minor performance measure	Polyp detection rate (PDR)
Description	Percentage of colonoscopies in patients aged 50 years or older in which at least one polyp was identified
Domain	Identification of pathology
Category	Process
Rationale	PDR reflects adequate inspection of bowel mucosa.
	PDR correlates with ADR and polypectomy rate is weakly associated with interval CRC risk.
Construct	Denominator: All screening and diagnostic colonoscopies in patients aged 50 years or older
	Numerator: Procedures in the denominator with at least one polyp identified
	Exclusions:
	– Emergency colonoscopy
	– Endoscopy with a specific therapeutic indication, including work-up of a previously detected lesion or follow-up of disease activity in inflammatory bowel disease
	Calculation: Proportion (%)
	Level of analysis: Service and endoscopist level
	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a yearly audit of a sample of 100 consecutive LGI endoscopies.
Standards	Minimum standard: 40%
	Target standard: no current target standard defined
	PDR is an approximation of ADR and should only be used when there is limited access to histopathology reports; however, caution is needed because PDR is susceptible to gaming.
	If the minimum standard is not met, there should be an attempt to obtain histopathology reports and calculate the ADR.
Consensus agreement	84.6%
PICO	3.1 (see Supporting Information)
Evidence grading	Low quality evidence.

The acceptance of this performance measure is based on agreement with the following statement:

• Polyp detection rate should be used as a measure of adequate inspection at screening or diagnostic colonoscopy in patients aged 50 years or more. (N3.5) Agreement: 84.6%

PDR is a surrogate for ADR and is more feasible to measure as it does not require histological verification. In some studies, PDR has been shown to correlate well with ADR;^64–66 however, in others the correlation was poor for polyps in the distal colorectum.^67,68 In one study, polypectomy rates of at least 25% were associated with a significantly lower risk of proximal interval CRC.⁴⁰ In a recent study, PDR was found to be non-inferior to ADR in predicting the risk of interval CRC.⁵⁵ With an average adenoma to polyp detection quotient of 0.64, the minimum standard PDR was estimated at 40%, which corresponds with an ADR of 25%.⁶⁶ The detection of adenomas and non-neoplastic polyps are however associated, which may inflate the PDR.⁶⁷ The use of PDR instead of ADR could therefore be considered if there is limited availability of histopathology data, accepting the potential risks of gaming. We note that the increased pressure on quality may force endoscopists to detect and remove non-neoplastic lesions that would otherwise be undetected so as to inflate the rate of detection of “so-called” polyps.

• 4. Domain: Management of pathology

  Key performance measure	Appropriate polypectomy technique
  Description	Adequate resection technique of colorectal polyps includes biopsy forceps removal of polyps ≤3 mm in size, and snare (cold or with diathermy) polypectomy for larger polyps. Polyp size estimated by endoscopists has to be included in the endoscopy report.
  Domain	Management of pathology
  Category	Process
  Rationale	Inappropriate polypectomy technique increases the risk of incomplete polyp removal.
  	Incomplete polyp removal leads to further costs and inconvenience as the examination has to be repeated.
  	Incomplete polyp removal is also considered to contribute to the development of interval CRCs.
  Construct	Denominator: Polyps >3 mm in size removed at colonoscopy (polyp size estimated by endoscopist)
  	Numerator: Polyps in the denominator removed with snare polypectomy (cold or with diathermy)
  	Exclusions: None
  	Calculation: proportion (%)
  	Level of analysis: Service and endoscopist
  	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a yearly audit of a sample of 100 consecutive LGI endoscopies.
  Standards	Minimum standard: ≥80%
  	Target standard: ≥90%
  	Colonoscopy reports must include information on polyp resection technique
  	If the minimum standard is not met, the rate of complete polyp resection should be measured and feedback should be given to the endoscopist or service. Additional training on basic polypectomy technique should be considered.
  	After evaluation and adjustment, close monitoring should be performed with a further audit within 6 months.
  Consensus agreement	93.3%
  PICO	4.6 (see Supporting Information)
  Evidence grading	Low quality evidence

The acceptance of this performance measure is based on agreement with the following statement:

• Adequate resection technique of small and diminutive colorectal polyps includes biopsy forceps removal of polyps ≤3 mm in size and snare polypectomy for larger polyps. (N4.6) Agreement: 93.3%

Incomplete polypectomy is considered the cause for up to 25% of interval CRCs.^69,70 Incomplete resection of polyps 5–20 mm in size varies from 6.5% to 22.7% among endoscopists;⁷¹ however, completeness of polyp resection is considered challenging to measure, and statements regarding this topic have not reached agreement in the current Delphi process (see Supporting information).

Biopsy forceps resection of polyps 4–5 mm in size or larger has been shown to be inferior to snare techniques, with regard to completeness of resection.^72,73 Therefore, the appropriate resection technique for colorectal polyps includes biopsy forceps removal of polyps ≤3 mm in size, and snare (cold or with diathermy) polypectomy for larger polyps. Despite this, in a recent large cohort study, it was demonstrated that 28.2% of lesions ≥5 mm in size were resected using biopsy forceps instead of a snare technique.⁷⁴ Contrary to this, in a large study from the UK, over 90% of polyps larger than 3 mm in size were removed using a snare.⁷⁵

There are insufficient data to set the minimum and target standards reliably, but the proposed values for the use of appropriate polypectomy techniques of ≥80% and ≥90%, respectively, seem relatively easy to achieve.

Minor performance measure	Tattooing resection sites
Description	In patients undergoing removal of colorectal non-pedunculated lesions 20 mm in size or larger, or with suspicious macroscopic features regardless of size, the resection site should be tattooed to improve future re-location of the resection site.
Domain	Management of pathology
Category	Process
Rationale	Facilitates detection of the post-polypectomy site at surveillance colonoscopy or surgical resection
Construct	Tattooing the resection site of the abovementioned lesions should be applied in all cases. A service must provide appropriate equipment.
	Denominator: Colonoscopies with removal of non-pedunculated lesions 20 mm in size or larger, or with suspicious macroscopic features regardless of size
	Numerator: Procedures in the denominator where the resection site was marked with a tattoo
	Exclusions: None
	Calculation: Proportion (%)
	Level of analysis: Service level
	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a 3-yearly audit of all colonoscopies performed over a 3-month period.
Standards	Minimum standard: Unknown
	Target standard: 100%
	Every endoscopy report for procedures where removal of the abovementioned lesions was performed should include written information on tattooing the resection site.
	If tattooing is not performed in all cases, feedback should be given to the service and all endoscopists.
Consensus agreement	93.3%
PICO	4.5 (see Supporting Information)
Evidence grading	Very low quality evidence

The acceptance of this performance measure is based on agreement with the following statement:

• In patients undergoing removal of colorectal lesions with a depressed component (0-IIc, according to the Paris classification) or non-granular or mixed-type laterally spreading tumors, located between the ascending and the sigmoid colon, the resection site should be tattooed to improve future re-location of the resection site. (N4.1) Agreement: 93.3%

Colorectal lesions with a depressed component and non-granular or mixed-type laterally spreading tumors (LSTs) harbor an increased risk of malignancy.^76–78 Therefore, the site of endoscopic removal of these lesions often needs to be re-located to identify recurrence or to guide surgical management. It has been shown that tattooing significantly shortens the time to re-locate the resection site on endoscopy.⁷⁹ There is however no evidence that tattooing the resection site increases the rate of re-location of lesions (see Supporting information). Preoperative tattooing using prepacked kits was proven to be a very effective method of tumor localization in laparoscopic surgery.⁸⁰ Moreover, some studies have shown that tattooing improves lymph node yield and facilitates the harvesting of suspicious lymph nodes during colorectal surgery.^81,82

Although the accepted statement focused only on lesions with an increased risk of malignancy, for audit purposes it will be much more feasible to track the tattooing of resection sites for all lesions larger than 20 mm in size. These lesions are frequently removed piecemeal, which increases the risk of recurrence,⁸³ and have a considerable risk of malignancy.⁸⁴ The minimum standard for tattooing resection sites is unknown.

Minor performance measure	Polyp retrieval rate
Description	Percentage of polyps removed that were retrieved for histopathology
Domain	Management of pathology
Category	Process
Rationale	The retrieval of polyps is required for histopathological diagnosis and is a prerequisite for recommendations on proper post-polypectomy surveillance interval.
Construct	Denominator: Polypectomies of polyps >5 mm
	Numerator: Polyps in the denominator that were retrieved for histopathology examination
	Exclusions: Removal of diminutive polyps (≤5 mm)
	Calculation: Proportion (%)
	Level of analysis: Service and endoscopist level
	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a yearly audit of a sample of 100 consecutive LGI endoscopies.
Standards	Minimum standard: ≥90%
	Target standard: ≥95%
	Colonoscopy reports must include information on non-retrieval of non-diminutive polyps.
	If the minimum standard is not reached, feedback should be given on the importance of this performance measure.
Consensus agreement	86.7%
PICO	No PICO (see Supporting Information)
Evidence grading	Very low quality evidence

The acceptance of this performance measure is based on agreement with the following statement:

• The non-diminutive polyp retrieval rate should be monitored. A service should have a polyp retrieval rate of ≥90%. (N4.2) Agreement: 86.7%

The retrieval of polyps after endoscopic resection is a “sine qua non” requirement for histopathology examination. Histopathology examination guides further management including post-polypectomy surveillance. Diminutive polyps (≤5 mm in size) harbor a very low risk of cancer or advanced histology and are considered amenable for a resect-and-discard policy following in vivo optical diagnosis under strictly controlled conditions.⁸⁵ Furthermore, diminutive polyps are frequently removed using biopsy forceps, which makes their retrieval quite straightforward.

It has therefore been decided to monitor only the retrieval of polyps larger than 5 mm in size. Their retrieval is not only more important from the clinical perspective but also technically more difficult because it requires the transected polyp to be suctioned into a trap, ensnared, or grasped using a Roth net, so that it can be removed together with the endoscope.^86,87 Even though the need for polyp retrieval seems obvious, it is unknown what the effect of substandard retrieval is on repeat colonoscopy rates or the appropriateness of recommended post-polypectomy surveillance.

The proposed minimum standard (≥90%) and target standard (≥95%) for polyp retrieval rate were based on values reported in recent large studies.^41,45,88,89 Polyp retrieval rate seems feasible to measure and is amenable for improvement through education and competitive feedback.⁹⁰

Minor performance measure	Advanced imaging assessment
Description	In patients undergoing removal of colorectal lesions with a depressed component (0-IIc, according to the Paris classification) or non-granular or mixed-type laterally spreading tumors (LSTs), conventional or virtual chromoendoscopy should be used to improve delineation of the lesion margins and to predict the potential depth of invasion.
Domain	Management of pathology
Category	Process
Rationale	Polyps with a depressed component (0-IIc) and non-granular or mixed type LSTs harbor a higher risk of submucosal invasion.
	Such polyps frequently have indistinct borders; therefore better margin delineation is warranted.
	Improved delineation and prediction of deep invasion may optimize management of these lesions.
Construct	Advanced imaging assessment should always be used before an attempt to remove the abovementioned lesions. A service offering removal of these types of lesions must provide dedicated equipment.
	Denominator: Colonoscopies with removal of lesions with a depressed component (0-IIc) or non-granular or mixed-type LSTs
	Numerator: Procedures in the denominator where virtual or conventional chromoendoscopy was used to improve delineation of the lesion margins (described in the report)
	Exclusions: None
	Calculation: Proportion (%)
	Level of analysis: Service and endoscopist
	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a 3-yearly audit of all colonoscopies performed over a 3-month period
Standards	Minimum standard: Unknown
	Target standard: 100%
	If the target standard is not met, feedback on the appropriate use of advanced imaging assessment is warranted.
	At a service level, the availability of equipment should be analyzed and facilitated.
	After evaluation and adjustment, close monitoring should be performed with a further audit within 6 months.
Consensus agreement	93.3%
PICO	4.4 (see Supporting Information)
Evidence grading	No evidence

The acceptance of this performance measure is based on agreement with the following statement:

• In patients undergoing removal of colorectal lesions with a depressed component (0-IIc, according to the Paris classification) or non-granular or mixed-type laterally spreading tumors, conventional or virtual chromoendoscopy should be used to improve delineation of lesion margins and predict potential depth of invasion. (N4.4) Agreement: 93.3%

In 2014, the ESGE issued guidelines on advanced endoscopic imaging for the detection and differentiation of colorectal neoplasia in which it suggested the use of advanced endoscopic imaging for margin assessment and prediction of deep submucosal invasion in lesions with a depressed component (0-IIc) or non-granular or mixed-type LSTs.⁸⁵ The quality of evidence supporting these recommendations was considered very low and moderate for margin delineation and assessment of depth of submucosal invasion, respectively. Since then no new evidence with clinically relevant endpoints for the patients (incomplete resection, interrupted procedure, cancer detection) has been published to further support its use (see Supporting information).

The availability, feasibility, and minimum standard of advanced imaging use, particularly in the community setting, are unknown. Colonoscopy services should set up structured monitoring and initiate audit to generate further evidence for advanced imaging.

Minor performance measure	Adequate description of polyp morphology
Description	The Paris classification should be routinely used to describe the morphology of non-pedunculated lesions identified at colonoscopy.
Domain	Management of pathology
Category	Process
Rationale	The Paris classification is a helpful tool to assess the risk of invasion.
	When polyp description is adequate, removal of polyps harboring suspicious features is likely to be avoided.
Construct	Denominator: Colonoscopies with removal of non-pedunculated lesions
	Numerator: Procedures in the denominator where the Paris classification was used to describe lesions
	Exclusions: None
	Calculation: Proportion (%)
	Level of analysis: Service and endoscopist
	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a 3-yearly audit of all colonoscopies performed over a 3-month period.
Standards	Minimum standard: Unknown
	Target standard: 100%
	Written colonoscopy reports should include a lesion description based on the Paris classification.
	If the target standard is not met, feedback on adequate description of polyp morphology is warranted.
	After evaluation and adjustment, close monitoring should be performed with a further audit within 6 months.
Consensus agreement	84.6%
PICO	3.9 (see Supporting Information)
Evidence grading	Very low quality evidence

The acceptance of this performance measure is based on agreement with the following statement:

• The Paris classification should be routinely used to describe the morphology of non-polypoid lesions identified at colonoscopy. (N4.5) Agreement: 84.6%

The Paris classification was developed with the aim of standardizing the terminology of superficial colorectal lesion morphology.⁷⁶ It divided lesions into two main groups: polypoid and non-polypoid, further defining four subtypes of the latter. Although its use is widely endorsed, it has never been fully validated. Recent studies have shown only moderate interobserver agreement for the Paris classification, even among experts.^91,92 More importantly, short training sessions are not sufficient to improve the agreement, suggesting that refinement of the classification is needed.⁹¹ Adoption of the classification in the community setting is unknown. The introduction of the Paris classification did however have two important effects: it raised awareness of subtle colorectal lesions among Western endoscopists,⁹³ and helped to predict submucosal invasion of colorectal lesions before their removal.^78,93

In light of the lack of better classifications, the Paris classification should be routinely used to describe the morphology of non-polypoid lesions identified at colonoscopy and its usage should be monitored. No minimum standard for this key performance measure was defined because of lack of evidence.

• 5. Domain: Complications

  Key performance measure	Complication rate
  Description	Percentage of patients in which complications (immediate, 7-day readmission rate, and 30-day mortality rate) occur after screening, diagnostic, or therapeutic colonoscopy
  Domain	Complications
  Category	Outcome
  Rationale	Monitoring the rate of complications after screening, diagnostic, and therapeutic colonoscopy is important to assess the safety of procedures, to identify possible targets for improvement, and to allow accurate informed consent of patients.
  Construct	Record the following parameters:
  	– Early complications, adverse events, and harms
  	– 7-day readmission rate (30-day readmission rates, where there are reliable registries and sufficient resources)
  	– 30-day mortality rate
  	Assessment should be done using a reliable method that allows identification of immediate and delayed complications, such as:
  	– Direct contact (e.g. telephone call) with the patient
  	– Analysis of hospital records (readmission rate)
  	– Analysis of registries (readmission rate and mortality rate)
  	Denominator: All colonoscopies
  	Numerator: Procedures in the denominator with a complication registered (separately for early, 7-day readmission (30-day readmission, where there are reliable registries and sufficient resources), and 30-day mortality)
  	Exclusions: None
  	Calculation: Proportion (%) (separate for each parameter)
  	Level of analysis: Service
  	Frequency: Yearly for all colonoscopies performed at a service level
  Standards	Minimum standard: ≤0.5% for 7-day readmission rate, standards not set for 30-day mortality rate or immediate complication rate
  	Target standard: no target standard set
  	Endoscopic reporting systems should allow the reporting of early (in-hospital) complications, including the type of complication, description of any action relating to the complication (need for transfusion, hospitalization, or prolonged hospitalization; surgery; death; need for endoscopic re-intervention), and time from endoscopic procedure to onset of the complication.
  	Regular morbidity and mortality conferences are encouraged to assess the causes of any complications and to discuss solutions to avoid them.
  Consensus agreement	93.8%
  PICO	5.1–5.2 (see Supporting Information)
  Evidence grading	Low quality evidence

The acceptance of this performance measure is based on agreement with the following statement:

• In patients undergoing colonoscopy, a 6-day readmission rate and 30-day mortality rate should be monitored using a reliable system. (N5.1) Agreement: 93.8%

The rate of complications, adverse events, and harms are important outcome measures of colonoscopy performance. Some studies and guidelines have reported rates for specific complications such as perforation, bleeding, or sedation-related cardiopulmonary adverse events.^6,45,94–96 These specific outcomes are however difficult to compare across services because they are infrequent, have variable definitions, and depend on case mix. For feasibility reasons, we propose to measure adverse outcomes, as defined in previous studies,^97–100 to give an overall rate of complications and to drill down into specific outcomes only if the standard is not met.

The definitions of complications are of paramount importance because the differences between major and minor complications or between minor complications and routine events encountered during the course of the procedure can be vague. The all-cause 30-day mortality rate is certainly well defined and important to measure. In large clinical or administrative databases, the rate of all-cause 30-day mortality has been estimated at 0.07% (1 in 1500),^{95–97,100–102} and the colonoscopy-specific mortality at more than 10 times lower (1 in 15 000 or lower).^{95,96,102,103} Although all-cause 30-day mortality rates would be impossible to compare across services, all deaths should be discussed during morbidity and mortality conferences.¹⁰⁴ The LGI working group members decided that, although the accepted statement focused on the 6-day readmission rate, this should be changed to a 7-day readmission rate in order to make it more comparable with the published literature. The 7-day or 30-day hospital admission/readmission rate is a well-defined and objective way to track late complications of colonoscopy.^{95–97,99,100}

Late complications represent over half of all colonoscopy-associated complications.⁹⁸ Furthermore, the 6-day readmission rate was shown to predict 30-day all-cause mortality.⁹⁹ The reported all-cause 7-day and 30-day hospital admission/readmission rates were 0.5% and 1.1%–3.8%,^95,97,99,100 respectively (0.5% for colonoscopy-specific readmission rates).⁹⁵ Therefore, the minimum standard of 0.5% seems acceptable for 7-day overall or 30-day colonoscopy-specific readmission rates.

The early complication rate (diagnosed immediately during the procedure or before patient discharge) is relatively easy to measure using appropriate endoscopy reporting systems.¹² The definition of an early complication is however more challenging and, in the view of the working group, should only include complications that result in one of the following: (i) lengthening of the hospital stay; (ii) unscheduled further endoscopic procedure; or (iii) emergency intervention, including blood transfusion or surgery.⁶

Reliable recording of all colonoscopy complications is a major concern.⁹⁸ A direct telephone call with a patient,¹⁰¹ analysis of hospital records,¹⁰⁰ and analysis of administrative data claims have all been used for this purpose,^97,100 but it is uncertain which method is the most feasible and reliable (see Supporting information).⁹⁸

• 6. Domain: Patient experience

  Key performance measure	Patient experience
  Description	Patient experience during and after colonoscopy and sigmoidoscopy should be routinely measured and self-reported by patients using validated scales.
  Domain	Patient experience
  Category	Outcome
  Rationale	Colonoscopy can be an unpleasant experience. Moreover, there are considerable differences between endoscopists and between different sedation modalities with regards to patient-reported pain and discomfort.
  	Patient experience and its improvement is crucial for the acceptance of procedures.
  Construct	Denominator: All colonoscopies
  	Numerator: Procedures in the denominator in which patient experience was measured using a validated scale (the Global Rating Scale, the Gastronet, or others)
  	Exclusions: Emergency colonoscopies
  	Calculation: Proportion (%)
  	Level of analysis: Individual endoscopist and service
  	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a yearly audit of a sample of 100 consecutive LGI endoscopies.
  Standards	Minimum standard: Unknown
  	Target standard: ≥90%
  	Currently there is no standard approach to measuring patient experience: different questionnaires are available and their comparative performance is unclear. Ideally, patient experience should be self-reported using a standardized and validated reporting method.
  	Audits should be performed on both service and individual endoscopist level to assess patient-reported outcomes.
  	In case of substandard results (for example if one endoscopist performs worse than others in the same service), additional training and feedback should be considered.
  Consensus agreement	93.8%
  PICO	7.1–7.4 (see Supporting Information)
  Evidence grading	Very low quality evidence

The acceptance of this performance measure is based on agreement with the following statements:

• Patient experience during and after unsedated or moderately sedated colonoscopy or sigmoidoscopy should be routinely measured. (N7.1) Agreement: 93.8%

• Patient experience with colonoscopy or sigmoidoscopy should be self-reported by a patient using a validated scale. (N7.2) Agreement: 93.8%

Colonoscopy may be perceived to be a painful and embarrassing procedure and this perception hampers patient participation in screening programs, adherence to surveillance recommendations, and even diagnostic work-up for large bowel symptoms.^105–107 Although sedation may decrease pain during colonoscopy, it does not eliminate it,¹⁰⁸ has little effect on post-procedure pain,²² and increases the risk of complications.¹⁰⁹ Therefore, monitoring patient experience, including intra- and post-procedure pain levels, is crucial.

Monitoring patient experience is feasible, yet it is not universal and no standardized approach exists. The two most widely used and validated questionnaires for assessing patient experience are the Global Rating Scale and the Gastronet.^{22,108,110–115} Patient coverage and response rates varied across services from less than 80% to over 90%,^22,116,117 and sustained compliance is a concern.¹¹⁶ Of note, there is poor to moderate correlation between physician- or nurse-recorded and patient-reported pain levels, therefore the latter measure should be the preferred one.¹¹⁸ The two main validated scales for pain assessment are a Visual Analog Scale and 4-point Verbal Rating Scale. Three studies have shown similar sensitivities for these scales (see Supporting information).^119–121

• 7. Domain: Post-procedure

  Key performance measure	Appropriate post-polypectomy surveillance recommendations
  Description	Adherence to post-polypectomy surveillance recommendations should be monitored and the reason for deviation from national/European guidelines should always be provided.
  Domain	Post-procedure
  Category	Process
  Rationale	Post-polypectomy surveillance recommendations reflect the best evidence-based balance between benefit and harm.
  	Too frequent surveillance wastes resources and exposes patients to complications of an invasive procedure.
  	Too infrequent surveillance may limit the effectiveness of surveillance.
  Construct	This performance measure takes into account not only patients’ adherence to the recommendations but also whether there were any written recommendations (letter to the patient or the patient’s general practitioner).
  	Denominator: Patients who underwent colorectal polypectomy
  	Numerator: Patients in the denominator who received proper (national or European) surveillance recommendations
  	Exclusions: Reason provided for deviation from the actual surveillance recommendations
  	Calculation: Proportion (%)
  	Level of analysis: Service and individual endoscopist
  	Frequency: Continuous monitoring using novel endoscopy reporting systems should be the preferred approach;12 an alternative approach is a yearly audit of a sample of 100 consecutive LGI endoscopies.
  Standards	Minimum standard: no standard defined
  	Target standard: ≥95%
  	All endoscopists should follow national or European guidelines for post-polypectomy surveillance and any deviation from these guidelines should be clearly stated.
  	When no written recommendation is given, this should be treated as a missing recommendation.
  	Endoscopic reporting systems should contain data about surveillance recommendations issued to the patient.
  	If there is suboptimal performance, an automated system that issues surveillance recommendations from the endoscopy database and reminders to the patients should be considered.
  Consensus agreement	93.8%
  PICO	No PICO (see Supporting Information)
  Evidence grading	Low quality evidence

The acceptance of this performance measure is based on agreement with the following statement:

• Adherence to post-polypectomy surveillance recommendations should be monitored. The reason for deviation from national/European guidelines should always be provided. (N8.1) Agreement: 93.8%

Patients who have had adenomas removed are believed to be at increased risk of developing new adenomas or cancer in the future.^122–124 In order to mitigate this risk, professional societies recommend patients undergo colonoscopy surveillance depending on age, comorbidity, and adenoma characteristics.^125,126 Surveillance intervals recommended in the guidelines represent the best evidence-based balance between the benefits (protection against CRC) and harms (too frequent invasive examinations) of subsequent colonoscopies.

Adherence to these recommendations is key to the efficacy and efficiency of colonoscopy surveillance. Unfortunately, studies from the Netherlands and Canada have shown that less than 30% of patients who have undergone adenoma removal receive appropriate surveillance.^127,128 One of the key reasons for inappropriate surveillance is inappropriate recommendations given by gastroenterologists, surgeons, or primary care physicians.^129,130 The adherence of physicians to the post-polypectomy surveillance recommendations could be relatively easily monitored using modern endoscopy reporting systems.¹² Any deviation from guideline recommendations should be clearly stated in the reporting system, with the rationale for this provided.

No minimum standard for this key performance measure was defined because of lack of evidence.

General conclusions, research priorities, and future prospects

This paper describes a short list of key performance measures for LGI endoscopy that have the best evidence-based impact on clinical outcomes, while being feasible to measure and susceptible to improvement.

The systematic process of development of these key performance measures revealed broad variation in the available evidence between the performance measures in different quality domains. Although the domains of completeness of procedure, identification of pathology, and pre-procedure have relatively robust scientific support, others, such as management of pathology and patient experience, are rather understudied. Indeed, these two quality domains were listed among the key research priorities by the ESGE research committee and are considered key research questions by the LGI working group (see Table 1).¹³¹

Table 1.

Areas for further research.

Domain	Key research questions
1 Pre-procedure	What kind of intervention improves the rate of adequate bowel preparation?
	What is the appropriate time that should be allotted for screening and diagnostic colonoscopies?
2 Completeness of procedure	What is the diagnostic yield (and interval cancer rate) relative to increasing cecal intubation rate?
	What is the benefit of cecal intubation documented within a written report only or within a written and photo report?
3 Identification of pathology	What is the target standard for adenoma detection rate?
	What performance measure reflects the identification of pathology outside the CRC screening/surveillance setting?
4 Management of pathology	What is the most reliable and feasible method of measuring completeness of polyp removal?
	What is the effectiveness of add-on techniques/scales (chromoendoscopy/Paris classification/tattooing resection sites) in the management of pathology?
5 Complications	What is the most reliable and feasible method to monitor complication rates?
	Does monitoring help to reduce complication rates?
6 Patient experience	What is the most reliable and feasible method to monitor patient experience?
	How can patient experience with colonoscopy be optimized?
7 Post-procedure	What are the optimal surveillance intervals following removal of colorectal polyps?
	What is the effect of monitoring appropriate post-polypectomy surveillance recommendations on adherence to surveillance colonoscopy?

CRC, colorectal cancer.

The other notable feature of the identified performance measures is that the evidence behind them comes almost exclusively from the field of CRC prevention and early detection. Although performance measures from the pre-procedure and completeness of procedure domains are largely universal, performance measures within the identification of pathology, management of pathology, and post-procedure domains are not applicable outside of the CRC screening/surveillance setting. Further research on these topics is warranted (see Table 1).

The first step now is to implement these key performance measures in endoscopy practice throughout Europe. We encourage individual endoscopists, as well as heads of endoscopy units, to start implementation of the performance measures without delay. Implementing performance measures is important to identify services and individual endoscopists with substandard levels of performance. The aim is not to penalize these endoscopists or services but to have a tool to improve the quality of endoscopy. Feedback and benchmarking of colonoscopy performance measures are usually sufficient to positively influence the overall quality of colonoscopy.^54,132 If the provision of such information turns out to be insufficient to promote improvement, the next step is to provide assistance and additional training.^50,52

At a service level, the implementation of key performance measures may well require investment in hardware to accommodate a more efficient auditing process. We want to encourage hospital management to support the implementation of these performance measures in their endoscopy services. We think that, in an era where general hospital accreditation has become increasingly important, hospital administrations will be more susceptible to support such actions. Moreover, we owe it to our patients to overcome individual or financial barriers to ensure that endoscopy services are of the highest quality and to set research priorities to gather data that will inform the next generation of performance measures.

Declaration of Conflicting Interests

M. Kaminski receives speaker’s and teaching fees from Olympus Poland. M. Bretthauer receives funds from Thieme Verlag for editorial work for Endoscopy. C. Rees’s department receives research funding from Olympus Medical, ARC Medical, Aquilant Endoscopy, Almirall, and Cook (from 2010 to present). E. Dekker’s department has received research support and loan equipment from Olympus Europe (for the last 10 years). J.E. East has received research support and speaker’s fee from Olympus (from June 2014 to present); research support and consultancy fees from Cosmo Technologies (from January 2014 to present). C. Bennett owns and works for Systematic Research Ltd; and received a consultancy fee from ESGE to provide scientific, technical, and methodological expertise for the present project. C. Senore’s department receives PillCam Colon devices from Covidien-Given for study conduct, and loaner Fuse systems from EndoChoice. R. Bisschops has received: speaker’s fees from Covidien (2009–2014) and Fujifilm (2013); speaker’s fee and hands-on training sponsorship from Olympus Europe (2013–2014); speaker’s fee and research support from Pentax Europe; and an editorial fee from Thieme Verlag as co-editor of Endoscopy. R. Valori is a director of Quality Solutions for Healthcare, a company providing consultancy for improving quality in healthcare, and of AnderVal Ltd., a company providing endoscopy skills training. C. Spada has received training support from Given Imaging (2013 and 2014). C. Hassan has received equipment on loan from Fujinon, Olympus, EndoChoice, and Medtronic; consultancy fees from Medtronic, Alpha-Wasserman, Norgine, and EndoChoice. M. Dinis-Ribeiro receives funds from Thieme Verlag for editorial work for Endoscopy; his department has received support from Olympus for a teaching protocol (from August 2014 to July 2015). M. D. Rutter’s department receives research funding from Olympus for a colitis surveillance trial (2014 to present). J. Anderson, M. Bugajski, D. Domagk, M. Ferlitsch, K. Garborg, G. Hoff, R. Hultcrantz, R. Jover, E. J. Kuipers, I. Racz, S. Thomas-Gibson, T. Rösch, M. Rupinski, B. Seip, and S. Suchanek have no competing interests.

Funding

Michal F. Kaminski, Marek Bugajski, Michael Bretthauer, Kjetil Garborg, and Geir Hoff are supported by a grant (grant number Pol-Nor/204233/30/2013) from the Polish–Norwegian Research Program. Michael Bretthauer is supported by Top Researcher Grants of the Norwegian Cancer Society and the Norwegian Research Council. UEG supplied co-funding and additional project governance to this endeavor.

Supporting information

The detailed literature searches performed by an expert team of methodologists, as well as evolution and adaptation of the different PICOs and clinical statements during the Delphi voting process can be viewed in Supporting Information on the ESGE website.

Online content viewable at: http://www.esge.com/performance-measures-for-lower-gastrointestinal-endoscopy.html