Colon Cancer Screening Models: Lessons and Challenges

Abstract

In choosing among different strategies for colorectal cancer (CRC) screening, the best evidence would come from randomized clinical trials that assess head-to-head comparisons of different tests, test combinations, and time intervals for their ability to reduce CRC incidence and mortality. Such trials are impractical and may be inefficient because of the sample size, effort, and expense required. In this setting, mathematical modeling can provide useful comparisons. Colon cancer screening is particularly well suited to modeling because so much data are available from randomized clinical trials of fecal occult blood testing and sigmoidoscopy and from observational studies. Also, we think we know much about how CRC develops from adenomas and about how well various tests detect cancer and adenomas in different stages and sizes. Limitations of modeling include the need for assumptions about input data and model structure.1Pignone M. Ransohoff D.F. Cross-model comparisons to improve the value of modeling: the case of colorectal cancer screening.Med Decis Making. 2011; 31: 524-526Crossref PubMed Scopus (7) Google ScholarModeling can be difficult to understand and evaluate because of its inherent complexity. However, we should not shoot the messenger—model making—because of the complexity and uncertainty that it requires us to examine. Indeed, we should perhaps thank the messenger for helping us understand the nature and magnitude of uncertainties that we then can examine in research and account for in clinical and policy decisions. Whatever its challenges, modeling is relied on by sophisticated consumers such as the US Preventive Services Task Force, the Centers for Medicare and Medicaid Services, and health maintenance organizations, which have both the motivation and resources to examine the choices and consequences among different screening strategies.For CRC screening, models have told us that, compared with no screening, any of several common strategies (such as yearly high-sensitivity fecal occult blood testing or fecal immunochemical testing, periodic colonoscopy, or sigmoidoscopy combined with fecal occult blood testing) reduce CRC mortality by roughly similar magnitudes if screening is adhered to over time. Several strategies are recommended as acceptable.2Pignone M. Saha S. Hoerger T. et al.Cost-effectiveness analyses of colorectal cancer screening: a systematic review for the U.S. Preventive Services Task Force.Ann Intern Med. 2002; 137: 96-104Crossref PubMed Scopus (506) Google Scholar, 3Zauber A.G. Lansdorp-Vogelaar I. Knudsen A.B. et al.Evaluating test strategies for colorectal cancer screening: a decision analysis for the U.S. Preventive Services Task Force.Ann Intern Med. 2008; 149: 659-669Crossref PubMed Scopus (471) Google ScholarIn this issue of Clinical Gastroenterology and Hepatology, modeling is used to assess a new “hybrid strategy,” periodic fecal immunochemical testing performed between ages 50 and 65, followed by a single colonoscopy at age 66, to determine how it compares with other common strategies.4Dinh T. Ladabaum U. Alperin P. et al.Health benefits and cost-effectiveness of a hybrid screening strategy for colorectal cancer.Clin Gastroenterol Hepatol. 2013; 11: 1158-1166Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar The analysis was commissioned by Kaiser Permanente Northern California to assess whether high-quality screening could be performed in a way that reduces use of colonoscopy, a limited resource, and the potential opportunity cost for a health maintenance organization with the responsibility of caring for a denominator population. The Archimedes model4Dinh T. Ladabaum U. Alperin P. et al.Health benefits and cost-effectiveness of a hybrid screening strategy for colorectal cancer.Clin Gastroenterol Hepatol. 2013; 11: 1158-1166Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar incorporated patient-level detail for the Kaiser Permanente population regarding age, sex, ethnicity, family history, and screening history, and it followed up subjects for 30 years.The hybrid strategy was shown to be comparable in effectiveness with current common strategies. A population of 100,000 persons with no screening would have 6004 CRCs and 1837 CRC deaths. Screening with periodic sigmoidoscopy reduces those numbers to 2822 CRCs and 839 CRC deaths, with fecal immunochemical testing it reduces those numbers to 1882 and 587, with the new hybrid strategy it reduces those numbers to 1664 and 495, and with periodic colonoscopy it reduces those numbers to 1451 and 428. The number of colonoscopies was reduced from 290,400 to 212,100 in the hybrid strategy compared with the colonoscopy strategy, whereas costs per person were slightly lower. Analyses using quality-adjusted life-years show similar relationships.These findings should not be surprising in light of other models’ results because the hybrid strategy is a kind of amalgam of acceptable strategies. One feature of the hybrid strategy deserving more attention, however, is what happens if subjects are followed up past age 80. If the last screening test is a colonoscopy at age 66, then people may not be up to date at the 75-year-old “stopping age” suggested by the US Preventive Services Task Force.3Zauber A.G. Lansdorp-Vogelaar I. Knudsen A.B. et al.Evaluating test strategies for colorectal cancer screening: a decision analysis for the U.S. Preventive Services Task Force.Ann Intern Med. 2008; 149: 659-669Crossref PubMed Scopus (471) Google Scholar If all lesions that will become clinically significant CRCs already are present by age 66 and can be detected and removed, then protection may last to age 80 or longer. However, to the extent that new lesions may later appear and grow rapidly, that is, have a short dwell time (the amount of time required for adenomas to develop into cancers and for cancers to become symptomatic and ultimately lethal), then older persons may incur some CRC and CRC mortality. These possibilities, and the magnitude of trade-offs involved in additional screening, should be considered further. One other practical consideration is that implementing a hybrid strategy may involve logistics that incur new costs, infrastructure, and other challenges in adherence; these also would need to be considered further.This modeling exercise illustrates the importance of identifying and examining key assumptions, and it highlights big-picture issues that, in 2013, need further consideration by researchers and policy makers.Assumptions about dwell times can have important implications for the outcome of a highly sensitive test performed infrequently (such as colonoscopy) compared with a less-sensitive test performed more frequently (such as fecal occult blood testing). If cancers that kill people grow rapidly from a small size, then a less-sensitive test applied more frequently may detect more cancers compared with a more sensitive test applied less frequently. The problem is that we do not understand, biologically, what the distribution of dwell times is. CRC models use different dwell times (from 10.6 to 25.8 y) as has been considered in a collaborative effort of modelers to understand the impact of “deep assumptions” on outcomes of CRC models.5Kuntz K.M. Lansdorp-Vogelaar I. Rutter C.M. et al.A systematic comparison of microsimulation models of colorectal cancer: the role of assumptions about adenoma progression.Med Decis Making. 2011; 31: 530-539Crossref PubMed Scopus (80) Google Scholar (The Archimedes model dwell time is roughly 17.5 years).Even if the average dwell time is long, that average may comprise a distribution that includes a substantial portion of lesions with very short dwell times. If so, then screening intervals would need to be tailored to detect that subgroup. The concept of distribution of dwell times is not just theoretical. Such a distribution, from indolent to aggressive, is well understood for cancers of the prostate, breast, or lung. We know less about the distribution of CRC growth rates because CRC and its detectable precursors are removed as soon as they are identified. Short dwell times for CRC might be one explanation for interval cancers found after colonoscopy; the other explanation is that lesions were missed.6Martinez M.E. Baron J.A. Lieberman D.A. et al.A pooled analysis of advanced colorectal neoplasia diagnoses after colonoscopic polypectomy.Gastroenterology. 2009; 136: 832-841Abstract Full Text Full Text PDF PubMed Scopus (436) Google Scholar When dwell times are short, strategies of less-sensitive but frequently applied tests may do a better job of preventing CRC mortality. In one analysis, an assumption of short dwell time suggested that a strategy of sigmoidoscopy combined with fecal occult blood testing could be as effective or even more effective than a strategy of colonoscopy.2Pignone M. Saha S. Hoerger T. et al.Cost-effectiveness analyses of colorectal cancer screening: a systematic review for the U.S. Preventive Services Task Force.Ann Intern Med. 2002; 137: 96-104Crossref PubMed Scopus (506) Google Scholar, 7Frazier A.L. Colditz G.A. Fuchs C.S. et al.Cost-effectiveness of screening for colorectal cancer in the general population.JAMA. 2000; 284: 1954-1961Crossref PubMed Scopus (549) Google ScholarA separate assumption about test independence may compromise a less-sensitive test’s ability to reduce CRC when applied repeatedly, such as fecal immunochemical testing. A test applied repeatedly can have multiple chances to detect a lesion if test results are independent at each application. For example, if fecal immunochemical testing has a sensitivity of 60% at one application, then lesions missed would have, on a next application, a 60% chance of being detected, resulting in a cumulative sensitivity for 2 tests of 84%. If, however, some lesions simply do not bleed, then test results are not independent and cumulative sensitivity would not increase. Although most models assume test independence, different assumptions would lead to different outcomes. Colonoscopy can have the problem of nonindependence, too; a lesion missed on one examination because it was hidden behind a fold might tend to be missed on the next examination.A larger issue touched on in this analysis and worth further consideration is postpolypectomy surveillance. Surveillance is becoming an increasingly important component of gastroenterology practice (and was responsible for 63% of all colonoscopies in the colonoscopy strategy in this model).4Dinh T. Ladabaum U. Alperin P. et al.Health benefits and cost-effectiveness of a hybrid screening strategy for colorectal cancer.Clin Gastroenterol Hepatol. 2013; 11: 1158-1166Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar Regardless of which tests are used in primary screening, colonoscopy eventually may be performed because of positive results (true positive or false positive) of fecal immunochemical testing, sigmoidoscopy, or computerized tomography colonography. Colonoscopy may detect adenomas in nearly 50% of subjects8Kahi C.J. Hewett D.G. Norton D.L. et al.Prevalence and variable detection of proximal colon serrated polyps during screening colonoscopy.Clin Gastroenterol Hepatology. 2011; 9: 42-46Abstract Full Text Full Text PDF PubMed Scopus (335) Google Scholar because of high-definition colonoscopes and increased attention to adenoma detection rates. Although the vast majority of adenomas are small or diminutive, postpolypectomy surveillance recommendations may be aggressive even for lesions that signify no increased future risk of CRC (ie, average future risk). In one study of gastroenterologists’ recommendations as indicated in patients’ office charts, 35% of persons with a single tubular adenoma less than 0.5 cm (and excellent preparation) were recommended to return within 3 years and nearly 80% were recommended to return within 6 years.9Ransohoff D.F. Yankaskas B. Gizlice Z. et al.Recommendations for post-polypectomy surveillance in community practice.Dig Dis Sci. 2011; 56: 2623-2630Crossref PubMed Scopus (42) Google Scholar In the Archimedes model, patients with even a small adenoma were enrolled in surveillance colonoscopy at 5-year intervals over a lifetime.4Dinh T. Ladabaum U. Alperin P. et al.Health benefits and cost-effectiveness of a hybrid screening strategy for colorectal cancer.Clin Gastroenterol Hepatol. 2013; 11: 1158-1166Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar We need to consider further how to handle overly aggressive surveillance and to learn how much CRC reduction comes from different strategies and intensities of surveillance.The field of model making has been evolving in healthy ways through transparency and collaboration facilitated by the Cancer Intervention and Surveillance Modeling Network, sponsored by the National Cancer Institute, to assess models’ differences in assumptions about natural history and other critical inputs. The goal is to derive insights about screening choices and consequences and to understand the implications of uncertainty illustrated by various models.5Kuntz K.M. Lansdorp-Vogelaar I. Rutter C.M. et al.A systematic comparison of microsimulation models of colorectal cancer: the role of assumptions about adenoma progression.Med Decis Making. 2011; 31: 530-539Crossref PubMed Scopus (80) Google Scholar, 10van Ballegooijen M. Rutter C.M. Knudsen A.B. et al.Clarifying differences in natural history between models of screening: the case of colorectal cancer.Med Decis Making. 2011; 31: 540-549Crossref PubMed Scopus (36) Google Scholar By publishing its calibration studies, Archimedes has made a useful contribution to this larger field.4Dinh T. Ladabaum U. Alperin P. et al.Health benefits and cost-effectiveness of a hybrid screening strategy for colorectal cancer.Clin Gastroenterol Hepatol. 2013; 11: 1158-1166Abstract Full Text Full Text PDF PubMed Scopus (34) Google ScholarDespite uncertainties and challenges, models have helped establish a big picture conclusion that any of several CRC screening strategies is acceptable in reducing CRC mortality by about 50% or more. The best choice for patients really is the one that they will follow. In the meantime, and as we debate fine details of screening and surveillance, we should not lose sight of this main and welcome larger conclusion about how well CRC screening works. In choosing among different strategies for colorectal cancer (CRC) screening, the best evidence would come from randomized clinical trials that assess head-to-head comparisons of different tests, test combinations, and time intervals for their ability to reduce CRC incidence and mortality. Such trials are impractical and may be inefficient because of the sample size, effort, and expense required. In this setting, mathematical modeling can provide useful comparisons. Colon cancer screening is particularly well suited to modeling because so much data are available from randomized clinical trials of fecal occult blood testing and sigmoidoscopy and from observational studies. Also, we think we know much about how CRC develops from adenomas and about how well various tests detect cancer and adenomas in different stages and sizes. Limitations of modeling include the need for assumptions about input data and model structure.1Pignone M. Ransohoff D.F. Cross-model comparisons to improve the value of modeling: the case of colorectal cancer screening.Med Decis Making. 2011; 31: 524-526Crossref PubMed Scopus (7) Google Scholar Modeling can be difficult to understand and evaluate because of its inherent complexity. However, we should not shoot the messenger—model making—because of the complexity and uncertainty that it requires us to examine. Indeed, we should perhaps thank the messenger for helping us understand the nature and magnitude of uncertainties that we then can examine in research and account for in clinical and policy decisions. Whatever its challenges, modeling is relied on by sophisticated consumers such as the US Preventive Services Task Force, the Centers for Medicare and Medicaid Services, and health maintenance organizations, which have both the motivation and resources to examine the choices and consequences among different screening strategies. For CRC screening, models have told us that, compared with no screening, any of several common strategies (such as yearly high-sensitivity fecal occult blood testing or fecal immunochemical testing, periodic colonoscopy, or sigmoidoscopy combined with fecal occult blood testing) reduce CRC mortality by roughly similar magnitudes if screening is adhered to over time. Several strategies are recommended as acceptable.2Pignone M. Saha S. Hoerger T. et al.Cost-effectiveness analyses of colorectal cancer screening: a systematic review for the U.S. Preventive Services Task Force.Ann Intern Med. 2002; 137: 96-104Crossref PubMed Scopus (506) Google Scholar, 3Zauber A.G. Lansdorp-Vogelaar I. Knudsen A.B. et al.Evaluating test strategies for colorectal cancer screening: a decision analysis for the U.S. Preventive Services Task Force.Ann Intern Med. 2008; 149: 659-669Crossref PubMed Scopus (471) Google Scholar In this issue of Clinical Gastroenterology and Hepatology, modeling is used to assess a new “hybrid strategy,” periodic fecal immunochemical testing performed between ages 50 and 65, followed by a single colonoscopy at age 66, to determine how it compares with other common strategies.4Dinh T. Ladabaum U. Alperin P. et al.Health benefits and cost-effectiveness of a hybrid screening strategy for colorectal cancer.Clin Gastroenterol Hepatol. 2013; 11: 1158-1166Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar The analysis was commissioned by Kaiser Permanente Northern California to assess whether high-quality screening could be performed in a way that reduces use of colonoscopy, a limited resource, and the potential opportunity cost for a health maintenance organization with the responsibility of caring for a denominator population. The Archimedes model4Dinh T. Ladabaum U. Alperin P. et al.Health benefits and cost-effectiveness of a hybrid screening strategy for colorectal cancer.Clin Gastroenterol Hepatol. 2013; 11: 1158-1166Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar incorporated patient-level detail for the Kaiser Permanente population regarding age, sex, ethnicity, family history, and screening history, and it followed up subjects for 30 years. The hybrid strategy was shown to be comparable in effectiveness with current common strategies. A population of 100,000 persons with no screening would have 6004 CRCs and 1837 CRC deaths. Screening with periodic sigmoidoscopy reduces those numbers to 2822 CRCs and 839 CRC deaths, with fecal immunochemical testing it reduces those numbers to 1882 and 587, with the new hybrid strategy it reduces those numbers to 1664 and 495, and with periodic colonoscopy it reduces those numbers to 1451 and 428. The number of colonoscopies was reduced from 290,400 to 212,100 in the hybrid strategy compared with the colonoscopy strategy, whereas costs per person were slightly lower. Analyses using quality-adjusted life-years show similar relationships. These findings should not be surprising in light of other models’ results because the hybrid strategy is a kind of amalgam of acceptable strategies. One feature of the hybrid strategy deserving more attention, however, is what happens if subjects are followed up past age 80. If the last screening test is a colonoscopy at age 66, then people may not be up to date at the 75-year-old “stopping age” suggested by the US Preventive Services Task Force.3Zauber A.G. Lansdorp-Vogelaar I. Knudsen A.B. et al.Evaluating test strategies for colorectal cancer screening: a decision analysis for the U.S. Preventive Services Task Force.Ann Intern Med. 2008; 149: 659-669Crossref PubMed Scopus (471) Google Scholar If all lesions that will become clinically significant CRCs already are present by age 66 and can be detected and removed, then protection may last to age 80 or longer. However, to the extent that new lesions may later appear and grow rapidly, that is, have a short dwell time (the amount of time required for adenomas to develop into cancers and for cancers to become symptomatic and ultimately lethal), then older persons may incur some CRC and CRC mortality. These possibilities, and the magnitude of trade-offs involved in additional screening, should be considered further. One other practical consideration is that implementing a hybrid strategy may involve logistics that incur new costs, infrastructure, and other challenges in adherence; these also would need to be considered further. This modeling exercise illustrates the importance of identifying and examining key assumptions, and it highlights big-picture issues that, in 2013, need further consideration by researchers and policy makers. Assumptions about dwell times can have important implications for the outcome of a highly sensitive test performed infrequently (such as colonoscopy) compared with a less-sensitive test performed more frequently (such as fecal occult blood testing). If cancers that kill people grow rapidly from a small size, then a less-sensitive test applied more frequently may detect more cancers compared with a more sensitive test applied less frequently. The problem is that we do not understand, biologically, what the distribution of dwell times is. CRC models use different dwell times (from 10.6 to 25.8 y) as has been considered in a collaborative effort of modelers to understand the impact of “deep assumptions” on outcomes of CRC models.5Kuntz K.M. Lansdorp-Vogelaar I. Rutter C.M. et al.A systematic comparison of microsimulation models of colorectal cancer: the role of assumptions about adenoma progression.Med Decis Making. 2011; 31: 530-539Crossref PubMed Scopus (80) Google Scholar (The Archimedes model dwell time is roughly 17.5 years). Even if the average dwell time is long, that average may comprise a distribution that includes a substantial portion of lesions with very short dwell times. If so, then screening intervals would need to be tailored to detect that subgroup. The concept of distribution of dwell times is not just theoretical. Such a distribution, from indolent to aggressive, is well understood for cancers of the prostate, breast, or lung. We know less about the distribution of CRC growth rates because CRC and its detectable precursors are removed as soon as they are identified. Short dwell times for CRC might be one explanation for interval cancers found after colonoscopy; the other explanation is that lesions were missed.6Martinez M.E. Baron J.A. Lieberman D.A. et al.A pooled analysis of advanced colorectal neoplasia diagnoses after colonoscopic polypectomy.Gastroenterology. 2009; 136: 832-841Abstract Full Text Full Text PDF PubMed Scopus (436) Google Scholar When dwell times are short, strategies of less-sensitive but frequently applied tests may do a better job of preventing CRC mortality. In one analysis, an assumption of short dwell time suggested that a strategy of sigmoidoscopy combined with fecal occult blood testing could be as effective or even more effective than a strategy of colonoscopy.2Pignone M. Saha S. Hoerger T. et al.Cost-effectiveness analyses of colorectal cancer screening: a systematic review for the U.S. Preventive Services Task Force.Ann Intern Med. 2002; 137: 96-104Crossref PubMed Scopus (506) Google Scholar, 7Frazier A.L. Colditz G.A. Fuchs C.S. et al.Cost-effectiveness of screening for colorectal cancer in the general population.JAMA. 2000; 284: 1954-1961Crossref PubMed Scopus (549) Google Scholar A separate assumption about test independence may compromise a less-sensitive test’s ability to reduce CRC when applied repeatedly, such as fecal immunochemical testing. A test applied repeatedly can have multiple chances to detect a lesion if test results are independent at each application. For example, if fecal immunochemical testing has a sensitivity of 60% at one application, then lesions missed would have, on a next application, a 60% chance of being detected, resulting in a cumulative sensitivity for 2 tests of 84%. If, however, some lesions simply do not bleed, then test results are not independent and cumulative sensitivity would not increase. Although most models assume test independence, different assumptions would lead to different outcomes. Colonoscopy can have the problem of nonindependence, too; a lesion missed on one examination because it was hidden behind a fold might tend to be missed on the next examination. A larger issue touched on in this analysis and worth further consideration is postpolypectomy surveillance. Surveillance is becoming an increasingly important component of gastroenterology practice (and was responsible for 63% of all colonoscopies in the colonoscopy strategy in this model).4Dinh T. Ladabaum U. Alperin P. et al.Health benefits and cost-effectiveness of a hybrid screening strategy for colorectal cancer.Clin Gastroenterol Hepatol. 2013; 11: 1158-1166Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar Regardless of which tests are used in primary screening, colonoscopy eventually may be performed because of positive results (true positive or false positive) of fecal immunochemical testing, sigmoidoscopy, or computerized tomography colonography. Colonoscopy may detect adenomas in nearly 50% of subjects8Kahi C.J. Hewett D.G. Norton D.L. et al.Prevalence and variable detection of proximal colon serrated polyps during screening colonoscopy.Clin Gastroenterol Hepatology. 2011; 9: 42-46Abstract Full Text Full Text PDF PubMed Scopus (335) Google Scholar because of high-definition colonoscopes and increased attention to adenoma detection rates. Although the vast majority of adenomas are small or diminutive, postpolypectomy surveillance recommendations may be aggressive even for lesions that signify no increased future risk of CRC (ie, average future risk). In one study of gastroenterologists’ recommendations as indicated in patients’ office charts, 35% of persons with a single tubular adenoma less than 0.5 cm (and excellent preparation) were recommended to return within 3 years and nearly 80% were recommended to return within 6 years.9Ransohoff D.F. Yankaskas B. Gizlice Z. et al.Recommendations for post-polypectomy surveillance in community practice.Dig Dis Sci. 2011; 56: 2623-2630Crossref PubMed Scopus (42) Google Scholar In the Archimedes model, patients with even a small adenoma were enrolled in surveillance colonoscopy at 5-year intervals over a lifetime.4Dinh T. Ladabaum U. Alperin P. et al.Health benefits and cost-effectiveness of a hybrid screening strategy for colorectal cancer.Clin Gastroenterol Hepatol. 2013; 11: 1158-1166Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar We need to consider further how to handle overly aggressive surveillance and to learn how much CRC reduction comes from different strategies and intensities of surveillance. The field of model making has been evolving in healthy ways through transparency and collaboration facilitated by the Cancer Intervention and Surveillance Modeling Network, sponsored by the National Cancer Institute, to assess models’ differences in assumptions about natural history and other critical inputs. The goal is to derive insights about screening choices and consequences and to understand the implications of uncertainty illustrated by various models.5Kuntz K.M. Lansdorp-Vogelaar I. Rutter C.M. et al.A systematic comparison of microsimulation models of colorectal cancer: the role of assumptions about adenoma progression.Med Decis Making. 2011; 31: 530-539Crossref PubMed Scopus (80) Google Scholar, 10van Ballegooijen M. Rutter C.M. Knudsen A.B. et al.Clarifying differences in natural history between models of screening: the case of colorectal cancer.Med Decis Making. 2011; 31: 540-549Crossref PubMed Scopus (36) Google Scholar By publishing its calibration studies, Archimedes has made a useful contribution to this larger field.4Dinh T. Ladabaum U. Alperin P. et al.Health benefits and cost-effectiveness of a hybrid screening strategy for colorectal cancer.Clin Gastroenterol Hepatol. 2013; 11: 1158-1166Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar Despite uncertainties and challenges, models have helped establish a big picture conclusion that any of several CRC screening strategies is acceptable in reducing CRC mortality by about 50% or more. The best choice for patients really is the one that they will follow. In the meantime, and as we debate fine details of screening and surveillance, we should not lose sight of this main and welcome larger conclusion about how well CRC screening works. Health Benefits and Cost-effectiveness of a Hybrid Screening Strategy for Colorectal CancerClinical Gastroenterology and HepatologyVol. 11Issue 9PreviewColorectal cancer (CRC) screening guidelines recommend screening schedules for each single type of test except for concurrent sigmoidoscopy and fecal occult blood test (FOBT). We investigated the cost-effectiveness of a hybrid screening strategy that was based on a fecal immunological test (FIT) and colonoscopy. Full-Text PDF