Effectiveness of Colonoscopy Screening for Colorectal Cancer
Effectiveness of Colonoscopy Screening for Colorectal Cancer
Screening programmes are ambitious, attempting to prevent disease in those that appear not to have any. Many will require evaluation to prevent disease in a few, and so most that submit to screening can never benefit.
It is in this light, that quality assurance programmes for colonoscopy are incredibly important. First, small decrements in the safety or quality of colonoscopy programmes could have significant impact on the net gains from any large-scale CRC screening programme. For example, injuring just a small fraction of those undergoing the test will erode any benefits of the programme. Further, from the quality perspective, of the many exams performed, only a few could actually detect or prevent a cancer. Missing that opportunity because the caecum was not reached, or the prep was poor, directly impacts the small margin of success that most screening programmes operate under. Finally, quality assurance programmes need to maximise the patient experience. While a single colonoscopy exam may be quite technically adept, if good patient experience is lacking, important follow-up may be compromised. A positive experience from the patient perspective both enhances the likelihood that the individual will return for needed follow-up and, through word of mouth, that larger numbers of individuals will submit to screening.
CRC screening is organised differently in different parts of the world. The European Union recommends organised, national programmes instead of opportunistic screening, because publicly organised, population-based programmes provide an administrative structure for service delivery and facilitate quality assurance and evaluation of the effectiveness and side-effects of the programme. Whether quality assurance is implemented within the framework of a national programme or an opportunistic one, the processes to improve delivery of a healthcare service, like colonoscopy, require multiple steps. Practice needs to be observed and compared with recognised performance standards. When deficiencies are recognised, change needs to be implemented. Finally, follow-up work ('closing the loop') is needed to assess whether improvement has occurred. In the following paragraphs, we highlight the key opportunities for quality assurance with colonoscopy programmes during the preprocedural, intraprocedural and postprocedural period. Assessment of the patient experience is also reviewed.
Assuring a safe patient experience with endoscopy begins with detailed attention to the environment and equipment used to perform the procedure. All practicing endoscopy units need to guarantee that general infection-control policies are in place and adhered to. The prevention of intravenous infection is particularly important in this regard. A recent high-profile case of documented Hepatitis C transmission to six individuals attending an endoscopy clinic related to improper use of multidose medication vials has been described.
Generally, when thinking of infection control in endoscopy, most would immediately consider issues related to endoscope reprocessing. Thankfully, transmission of infection through the endoscope is extraordinarily rare. However, a recent report of infectious transmission through a properly cleaned duodenoscope serves as an important reminder that risk is not zero and apparent outbreaks require prompt investigation. While the details of effective reprocessing fall outside the purview of this brief summary, excellent reviews can be found elsewhere. Guidelines for effective practice in this area have been widely published and need to be meticulously followed.
Beyond these environmental safety considerations are direct patient-level concerns that can impact procedural safety. There is good evidence that some persons undergoing screening may be too sick or frail to benefit from that intervention and would be at high risk for complications. A study based in Veterans Administration hospitals suggested that 40% of those undergoing screening in that system had a severe comorbid disease that would limit life expectancy to less than 5 years. Those hospitals or programmes offering direct access endoscopy are at the most risk in this area and need to have systems in place to be sure that those presenting for screening are generally appropriate for such testing. Also, medication review with particular attention to antiplatelet agents and anticoagulants is warranted. Again, both European and American guidelines exist to facilitate best practice. While guidelines can be used to inform practice, ultimately, decisions regarding continuation or interruption of such therapy need to be individualised. Systems should allow for such decision-making to be made well before the patient has initiated a bowel prep to avoid patient inconvenience and late cancellation of procedures. Likewise, the preprocedural assessment should include evaluation of such factors that influence the safe use of electrocautery (ie, pacemaker, Automated Implantable Cardioverter-Defibrillator, or previous joint replacement). An assessment of American Society of Anesthesiologists class has also been found to be helpful to better assess those most at risk for complications of sedation and, by extension, who would benefit from anaesthesia evaluation and support. Tools to make routine at the provider or practice level this type of evaluation so that it occurs uniformly are necessary. Once in place, audits of documentation are required to assure that guidance is being followed.
The performance and documentation of good quality colonoscopy is essential to any colonoscopy screening programme. The critical components of a quality colonoscopy include adequate preparation, caecal intubation and lesion detection and removal.
Bowel Preparation. As described above, the effectiveness of colonoscopy in reducing CRC incidence and mortality is stronger in the distal colon relative to the proximal. This discrepancy is multifactorial, but colon preparation and caecal intubation are two factors that contribute to this issue. Recently, expert panels from both the UK and USA have released guidance on bowel preparation for colonoscopy. Both endorse split dose preparation (dividing the bowel prep into two parts; one taken the evening before the exam, the other in the morning of the exam) or same day bowel preparation in case of afternoon colonoscopy as a way to improve preparation quality and explicitly encourage documentation of preparation quality in all patients. The US multisociety task force (USMSTF) guideline sets a target for adequate preparation of 85%. The first step, then, for practices not routinely documenting preparation quality is to mandate and audit that item. In a recent study examining 12 Dutch practices and 4800 colonoscopy exams, only 62% of the reports mentioned the quality of the preparation. Once regular documentation is accomplished, assuring prep adequacy to the levels suggested in the USMSTF guideline is the next step, with implementation of a quality improvement exercise if that is not accomplished.
Caecal Intubation. Like bowel preparation, caecal intubation is another factor that lends itself to benchmarking. First, target rates have been issued by expert consensus panels. The target rates for caecal intubation differ for screening (95%) and surveillance (90%) indications. Second, the factor can be audited. While admittedly, there is not strong data to demonstrate the accuracy of images of the caecum to definitively document colonoscopy completion, that practice should be encouraged. Pictures of the ileum have been shown to be more definitive, however, routine ileal intubation for this purpose is not recommended. Explicit documentation examination extant in the postprocedure report should be required. The use of computerised databases for procedural documentation facilitate the development of numerator/denominator reports for each endoscopist which could be used both within practices and externally to the public if required.
Adenoma Detection. Colonoscopy's effectiveness as a cancer prevention tool is largely accomplished through adenoma detection and resection. Historically, targets for adenoma detection during routine adult screening colonoscopy have been established at ≥25% for men and ≥15% for women. However, most recently, two US-based gastroenterology organisations (American College of Gastroenterology and the American Society of Gastrointestinal Endoscopy) have released revised guidance in this area. Specifically, the targets have been increased to an ADR ≥30% for men and ≥20% for women. The change in recommendation largely rests on improved data directly linking higher ADRs with lower subsequent interval cancer rates at the level of the endoscopist.
A number of specific comments with regards to ADR relative to other potential similar quality measures merit comment. While ADR lends itself to regular audit (ie, numerator/denominator measure) and quality improvement exercises, if suboptimal, there are challenges. In the absence of linkage between colonoscopy and pathology databases, it may be difficult to ascertain ADR. For this reason, polyp detection rate (PDR) has been proposed as a reasonable surrogate measure for ADR. Separating colonoscopy cases by indication may provide another challenge. Finally, ADR does not perfectly capture differences in absolute ADRs. However, there is outstanding evidence that subsequent cancer detection is less likely to occur when a previous colonoscopy was performed by a high adenoma detector relative to a low one. Based upon current evidence, ADR is likely the single most important procedural quality metric, and practices should work to develop mechanisms to reliably measure and follow it. Given the correlation between PDR and ADR, PDR can be employed until systems are in place to reliably measure ADR.
While ADR measurement is the near-term standard for adequate neoplasia detection during colonoscopy, admittedly, it is a surrogate measure. The goal of CRC screening programmes is early detection and cancer prevention. The detection of interval cancer after colonoscopy is well described and often represents missed neoplastic lesions. For obvious reasons, most individual providers or small groups cannot reliably detect the frequency of such occurrences in their practice. However, for larger healthcare system providers (eg, Veterans Administration, Kaiser Permanente) or nations with robust electronic data capture and linkages to cancer registries, routine measurement of interval cancers at the level of the endoscopist may someday be possible.
Polyp Resection. Of course, adenoma detection does not prevent subsequent CRC. Adequate resection is required for that. The contribution of incomplete polyp resection to the problem of interval cancer likely approaches 20%. While historically, endoscopists have taken resection for granted, one recent study using research biopsies around the margins of polyps thought to be entirely resected, estimated incomplete resection at 10% and significantly higher than that for large and serrated lesions. Presently, there is no easy way to measure this factor. Some experts have recommended retrieval rates as one marker of resection technique. Perhaps in the future, the use of endoscopic techniques (eg, chromoendoscopy), pathological techniques (eg, margin assessment) or a limited sampling of biopsy margins (eg, in large polyps) may be recommended. Currently, the minimal requirement should be adequate documentation in the endoscopic record as to the technique of endoscopic resection (piecemeal, en bloc) and whether specimens were submitted for pathological analysis.
Procedural complications are an important postprocedural quality metric. There is good evidence that complications from colonoscopy occurring in large screening programmes are likely more frequent than that described in those participating in well-characterised clinical trials. Early complications are much more easily tracked than those occurring days after the procedure. For this reason, careful tracking of adverse events related to sedation (eg, use of reversal agents) and unexpected admission immediately postprocedure are likely the most easily obtained surrogates for this measure. European guidelines do explicitly recommend a 30-day mortality review of all screened patients and an 8-day unplanned admission review although the guidelines recognise that such evaluation may not be feasible in all settings. An American Society for Gastrointestinal Endoscopy (ASGE) workgroup suggested routine follow-up at 14-days and developed a standard lexicon for describing and grading such occurrences. At minimum, endoscopy instructions should be explicit, encouraging those experiencing complications (such as abdominal pain and perforation) to call the endoscopy unit. Serious complications (ie, bleeding and perforation) identified either immediately or in delayed reporting needs regular tracking and rates of complications should be compared with expected rates for the procedure. Cases of severe complications should be audited and discussed.
The other major postprocedural quality benchmark that bears regular audit is colonoscopy surveillance recommendations. Although explicit guidelines are available, there is good evidence that they are often not followed. There are potential harms to both the patient and healthcare system when surveillance colonoscopy is either repeated too soon or too late. Tracking the appropriateness of surveillance recommendations is made challenging by the delayed nature of that determination. Such guidance demonstrates that the pathology is available, and many times the explicit instructions to the patient will fall outside the report itself (eg, into a letter) which becomes difficult to audit. Natural language processing likely will become an important tool to assess this measure. In addition to being sure that appropriate follow-up recommendations for colonoscopy are being made, systems should also be in place to facilitate those subsequent surveillance exams in an appropriate time frame.
Finally, but perhaps most importantly, a robust colonoscopy quality assurance programme needs to directly assess the patient experience. The patient experience begins well before the procedural day, and includes items, such as ease of scheduling, choice of procedural day and waiting time. It also includes the experience and interactions within the unit on the day of procedure (eg, the process of obtaining informed consent) and adequacy of communication in the days that follow the exam. The global rating scale (GRS) was initially developed in the UK and has been studied both within the UK and outside for the purpose of gathering this important information. A GRS questionnaire was recently applied in a group of over 1500 individuals undergoing outpatient endoscopy across 12 endoscopy units. Significant variation across practices was identified. For example, when patients were asked whether they were adequately informed about what to do if problems arose after discharge, those responding affirmatively ranged from 43% to 98%. Practices not routinely gathering information on patient experience should consider such an approach to capture this metric so that quality improvement programmes can be developed should deficiencies be identified.
Quality Assurance in Colonoscopy
Screening programmes are ambitious, attempting to prevent disease in those that appear not to have any. Many will require evaluation to prevent disease in a few, and so most that submit to screening can never benefit.
It is in this light, that quality assurance programmes for colonoscopy are incredibly important. First, small decrements in the safety or quality of colonoscopy programmes could have significant impact on the net gains from any large-scale CRC screening programme. For example, injuring just a small fraction of those undergoing the test will erode any benefits of the programme. Further, from the quality perspective, of the many exams performed, only a few could actually detect or prevent a cancer. Missing that opportunity because the caecum was not reached, or the prep was poor, directly impacts the small margin of success that most screening programmes operate under. Finally, quality assurance programmes need to maximise the patient experience. While a single colonoscopy exam may be quite technically adept, if good patient experience is lacking, important follow-up may be compromised. A positive experience from the patient perspective both enhances the likelihood that the individual will return for needed follow-up and, through word of mouth, that larger numbers of individuals will submit to screening.
CRC screening is organised differently in different parts of the world. The European Union recommends organised, national programmes instead of opportunistic screening, because publicly organised, population-based programmes provide an administrative structure for service delivery and facilitate quality assurance and evaluation of the effectiveness and side-effects of the programme. Whether quality assurance is implemented within the framework of a national programme or an opportunistic one, the processes to improve delivery of a healthcare service, like colonoscopy, require multiple steps. Practice needs to be observed and compared with recognised performance standards. When deficiencies are recognised, change needs to be implemented. Finally, follow-up work ('closing the loop') is needed to assess whether improvement has occurred. In the following paragraphs, we highlight the key opportunities for quality assurance with colonoscopy programmes during the preprocedural, intraprocedural and postprocedural period. Assessment of the patient experience is also reviewed.
Preprocedure
Assuring a safe patient experience with endoscopy begins with detailed attention to the environment and equipment used to perform the procedure. All practicing endoscopy units need to guarantee that general infection-control policies are in place and adhered to. The prevention of intravenous infection is particularly important in this regard. A recent high-profile case of documented Hepatitis C transmission to six individuals attending an endoscopy clinic related to improper use of multidose medication vials has been described.
Generally, when thinking of infection control in endoscopy, most would immediately consider issues related to endoscope reprocessing. Thankfully, transmission of infection through the endoscope is extraordinarily rare. However, a recent report of infectious transmission through a properly cleaned duodenoscope serves as an important reminder that risk is not zero and apparent outbreaks require prompt investigation. While the details of effective reprocessing fall outside the purview of this brief summary, excellent reviews can be found elsewhere. Guidelines for effective practice in this area have been widely published and need to be meticulously followed.
Beyond these environmental safety considerations are direct patient-level concerns that can impact procedural safety. There is good evidence that some persons undergoing screening may be too sick or frail to benefit from that intervention and would be at high risk for complications. A study based in Veterans Administration hospitals suggested that 40% of those undergoing screening in that system had a severe comorbid disease that would limit life expectancy to less than 5 years. Those hospitals or programmes offering direct access endoscopy are at the most risk in this area and need to have systems in place to be sure that those presenting for screening are generally appropriate for such testing. Also, medication review with particular attention to antiplatelet agents and anticoagulants is warranted. Again, both European and American guidelines exist to facilitate best practice. While guidelines can be used to inform practice, ultimately, decisions regarding continuation or interruption of such therapy need to be individualised. Systems should allow for such decision-making to be made well before the patient has initiated a bowel prep to avoid patient inconvenience and late cancellation of procedures. Likewise, the preprocedural assessment should include evaluation of such factors that influence the safe use of electrocautery (ie, pacemaker, Automated Implantable Cardioverter-Defibrillator, or previous joint replacement). An assessment of American Society of Anesthesiologists class has also been found to be helpful to better assess those most at risk for complications of sedation and, by extension, who would benefit from anaesthesia evaluation and support. Tools to make routine at the provider or practice level this type of evaluation so that it occurs uniformly are necessary. Once in place, audits of documentation are required to assure that guidance is being followed.
Intraprocedure
The performance and documentation of good quality colonoscopy is essential to any colonoscopy screening programme. The critical components of a quality colonoscopy include adequate preparation, caecal intubation and lesion detection and removal.
Bowel Preparation. As described above, the effectiveness of colonoscopy in reducing CRC incidence and mortality is stronger in the distal colon relative to the proximal. This discrepancy is multifactorial, but colon preparation and caecal intubation are two factors that contribute to this issue. Recently, expert panels from both the UK and USA have released guidance on bowel preparation for colonoscopy. Both endorse split dose preparation (dividing the bowel prep into two parts; one taken the evening before the exam, the other in the morning of the exam) or same day bowel preparation in case of afternoon colonoscopy as a way to improve preparation quality and explicitly encourage documentation of preparation quality in all patients. The US multisociety task force (USMSTF) guideline sets a target for adequate preparation of 85%. The first step, then, for practices not routinely documenting preparation quality is to mandate and audit that item. In a recent study examining 12 Dutch practices and 4800 colonoscopy exams, only 62% of the reports mentioned the quality of the preparation. Once regular documentation is accomplished, assuring prep adequacy to the levels suggested in the USMSTF guideline is the next step, with implementation of a quality improvement exercise if that is not accomplished.
Caecal Intubation. Like bowel preparation, caecal intubation is another factor that lends itself to benchmarking. First, target rates have been issued by expert consensus panels. The target rates for caecal intubation differ for screening (95%) and surveillance (90%) indications. Second, the factor can be audited. While admittedly, there is not strong data to demonstrate the accuracy of images of the caecum to definitively document colonoscopy completion, that practice should be encouraged. Pictures of the ileum have been shown to be more definitive, however, routine ileal intubation for this purpose is not recommended. Explicit documentation examination extant in the postprocedure report should be required. The use of computerised databases for procedural documentation facilitate the development of numerator/denominator reports for each endoscopist which could be used both within practices and externally to the public if required.
Adenoma Detection. Colonoscopy's effectiveness as a cancer prevention tool is largely accomplished through adenoma detection and resection. Historically, targets for adenoma detection during routine adult screening colonoscopy have been established at ≥25% for men and ≥15% for women. However, most recently, two US-based gastroenterology organisations (American College of Gastroenterology and the American Society of Gastrointestinal Endoscopy) have released revised guidance in this area. Specifically, the targets have been increased to an ADR ≥30% for men and ≥20% for women. The change in recommendation largely rests on improved data directly linking higher ADRs with lower subsequent interval cancer rates at the level of the endoscopist.
A number of specific comments with regards to ADR relative to other potential similar quality measures merit comment. While ADR lends itself to regular audit (ie, numerator/denominator measure) and quality improvement exercises, if suboptimal, there are challenges. In the absence of linkage between colonoscopy and pathology databases, it may be difficult to ascertain ADR. For this reason, polyp detection rate (PDR) has been proposed as a reasonable surrogate measure for ADR. Separating colonoscopy cases by indication may provide another challenge. Finally, ADR does not perfectly capture differences in absolute ADRs. However, there is outstanding evidence that subsequent cancer detection is less likely to occur when a previous colonoscopy was performed by a high adenoma detector relative to a low one. Based upon current evidence, ADR is likely the single most important procedural quality metric, and practices should work to develop mechanisms to reliably measure and follow it. Given the correlation between PDR and ADR, PDR can be employed until systems are in place to reliably measure ADR.
While ADR measurement is the near-term standard for adequate neoplasia detection during colonoscopy, admittedly, it is a surrogate measure. The goal of CRC screening programmes is early detection and cancer prevention. The detection of interval cancer after colonoscopy is well described and often represents missed neoplastic lesions. For obvious reasons, most individual providers or small groups cannot reliably detect the frequency of such occurrences in their practice. However, for larger healthcare system providers (eg, Veterans Administration, Kaiser Permanente) or nations with robust electronic data capture and linkages to cancer registries, routine measurement of interval cancers at the level of the endoscopist may someday be possible.
Polyp Resection. Of course, adenoma detection does not prevent subsequent CRC. Adequate resection is required for that. The contribution of incomplete polyp resection to the problem of interval cancer likely approaches 20%. While historically, endoscopists have taken resection for granted, one recent study using research biopsies around the margins of polyps thought to be entirely resected, estimated incomplete resection at 10% and significantly higher than that for large and serrated lesions. Presently, there is no easy way to measure this factor. Some experts have recommended retrieval rates as one marker of resection technique. Perhaps in the future, the use of endoscopic techniques (eg, chromoendoscopy), pathological techniques (eg, margin assessment) or a limited sampling of biopsy margins (eg, in large polyps) may be recommended. Currently, the minimal requirement should be adequate documentation in the endoscopic record as to the technique of endoscopic resection (piecemeal, en bloc) and whether specimens were submitted for pathological analysis.
Postprocedure
Procedural complications are an important postprocedural quality metric. There is good evidence that complications from colonoscopy occurring in large screening programmes are likely more frequent than that described in those participating in well-characterised clinical trials. Early complications are much more easily tracked than those occurring days after the procedure. For this reason, careful tracking of adverse events related to sedation (eg, use of reversal agents) and unexpected admission immediately postprocedure are likely the most easily obtained surrogates for this measure. European guidelines do explicitly recommend a 30-day mortality review of all screened patients and an 8-day unplanned admission review although the guidelines recognise that such evaluation may not be feasible in all settings. An American Society for Gastrointestinal Endoscopy (ASGE) workgroup suggested routine follow-up at 14-days and developed a standard lexicon for describing and grading such occurrences. At minimum, endoscopy instructions should be explicit, encouraging those experiencing complications (such as abdominal pain and perforation) to call the endoscopy unit. Serious complications (ie, bleeding and perforation) identified either immediately or in delayed reporting needs regular tracking and rates of complications should be compared with expected rates for the procedure. Cases of severe complications should be audited and discussed.
The other major postprocedural quality benchmark that bears regular audit is colonoscopy surveillance recommendations. Although explicit guidelines are available, there is good evidence that they are often not followed. There are potential harms to both the patient and healthcare system when surveillance colonoscopy is either repeated too soon or too late. Tracking the appropriateness of surveillance recommendations is made challenging by the delayed nature of that determination. Such guidance demonstrates that the pathology is available, and many times the explicit instructions to the patient will fall outside the report itself (eg, into a letter) which becomes difficult to audit. Natural language processing likely will become an important tool to assess this measure. In addition to being sure that appropriate follow-up recommendations for colonoscopy are being made, systems should also be in place to facilitate those subsequent surveillance exams in an appropriate time frame.
Patient Experience
Finally, but perhaps most importantly, a robust colonoscopy quality assurance programme needs to directly assess the patient experience. The patient experience begins well before the procedural day, and includes items, such as ease of scheduling, choice of procedural day and waiting time. It also includes the experience and interactions within the unit on the day of procedure (eg, the process of obtaining informed consent) and adequacy of communication in the days that follow the exam. The global rating scale (GRS) was initially developed in the UK and has been studied both within the UK and outside for the purpose of gathering this important information. A GRS questionnaire was recently applied in a group of over 1500 individuals undergoing outpatient endoscopy across 12 endoscopy units. Significant variation across practices was identified. For example, when patients were asked whether they were adequately informed about what to do if problems arose after discharge, those responding affirmatively ranged from 43% to 98%. Practices not routinely gathering information on patient experience should consider such an approach to capture this metric so that quality improvement programmes can be developed should deficiencies be identified.