Volume 24
Number 4
2003

[Table of Contents]

Potential impact of population-based colorectal cancer screening in Canada

William M Flanagan, Christel Le Petit, Jean-Marie Berthelot, Kathleen J White, B Ann Coombs and Elaine Jones-McLean

Abstract

Randomized controlled trials (RCT) have shown the efficacy of screening for colorectal cancer (CRC) using the fecal occult blood test (FOBT) and follow-up with colonoscopy. We evaluated the potential impact of population-based screening with FOBT followed by colonoscopy in Canada: mortality reduction, cost-effectiveness and resource requirements. The microsimulation model POHEM was adapted to simulate CRC screening using Canadian data and RCT results about test sensitivity and specificity, participation, incidence, staging, progression, mortality and direct health care costs. In Canada, biennial screening of 67% of individuals aged 50 to 74 in the year 2000 resulted in an estimated 10-year CRC mortality reduction of 16.7%. The life expectancy of the cohort increased by 15 days on average, and the demand for colonoscopy rose by 15% in the first year. The estimated cost of screening was $112 million per year or $11,907 per life-year gained (discounted at 5%). Potential effectiveness would depend on reaching target participation rates and finding resources to meet the demand for FOBT and colonoscopy. This work was conducted in support of the National Committee on Colorectal Cancer Screening.

Key words: colorectal cancer screening, cost-effectiveness, FOBT, microsimulation, POHEM

Introduction

Colorectal cancer (CRC) is the second leading cause of cancer deaths after lung cancer in Canada in both sexes combined but ranks third after prostate cancer in men and breast cancer in women. It affects men and women almost equally with increasing incidence beginning at age 50. Although incidence and mortality rates have slowly declined over recent years, the absolute numbers have increased as a result of the aging of the population. In 2002, there were an estimated 17,600 new cases of CRC and 6,600 deaths due to CRC.¹ A chart review at the Ottawa Regional Cancer Centre demonstrated that over half of the colorectal cancers detected in Canada were estimated to be stage III or IV cancers with five-year survival rates of approximately 60% and 10% respectively^a. Early detection through screening is expected to lead to better survival outcomes.

Health Canada established the National Committee on Colorectal Cancer Screening (NCCCS) in 1998 with a mandate to make recommendations on population-based CRC screening in Canada. Randomized controlled trials (RCTs) have shown the efficacy of CRC screening with fecal occult blood testing (FOBT) followed by colonoscopy for those with positive test outcomes.^3-5 Because real-life conditions are not necessarily captured in RCTs and the follow-up periods of the CRC trials were relatively short, the Population Health Model (POHEM) was used to evaluate the potential impact of population-based screening with FOBT followed by colonoscopy for colorectal cancer in Canada.

Methods

The specification of the screening protocol was developed in close collaboration with the NCCCS. Screening with FOBT (Hemoccult II, nonrehydrated) followed by colonoscopy for those with positive results was chosen as the screening modality to model, since evidence of its efficacy has been reported in three RCTs: from Funen, Denmark;³ Nottingham, UK;⁴ and Minnesota, USA.⁵

The Funen trial was population-based and had a clearly documented recruitment strategy that could be reproduced in POHEM to generate similar follow-up periods (10-year). We used it as the primary RCT to specify the screening model and used parameter estimates from the other trials where appropriate. The screening model was validated against the outcomes of the Funen trial before being applied to the Canadian setting.

The Population Health Model (POHEM)

POHEM is a microsimulation tool developed by Statistics Canada to model various aspects of the health of Canadians and evaluate possible interventions.^6,7 It creates synthetic, longitudinal population samples, starting with the birth of each individual in the cohort, and dynamically simulates their aging, including exposures to risk factors, disease onset conditional on risks, treatment, case fatality and costs. POHEM estimates the characteristics of a population cohort by synthesizing a large sample of complete individual health and socio-economic biographies based upon a myriad of detailed empirical observations.

POHEM includes detailed models of colorectal cancer, lung cancer, breast cancer, use of hormone replacement therapy, heart disease and fractures. It has been used to evaluate the impact of preventive tamoxifen in Canada,⁸ interventions in lung cancer,⁹ and the lifetime cost of breast cancer¹⁰ and colorectal cancer.

A Canadian CRC model of incidence and progression had been completed and incorporated into POHEM in 2000. It included disease incidence by age, sex and site (colon or rectum); disease progression to local recurrence, metastasis and death; and treatment options and cost. Incidence data were obtained from the Canadian Cancer Registry (1995), and stage distribution and survival data were derived from a chart review conducted at the Ottawa Regional Cancer Centre. Treatment options and associated costs were obtained from hospital discharge abstracts, surveys of oncologists, billing data and numerous consultations. The CRC screening module was integrated with this base model.

Simulating a screening program

Screening was simulated in POHEM to include the recruitment period; the target age of the population; screening frequency; participation in first and subsequent screens; FOBT sensitivity, specificity and sojourn time; FOBT outcomes; compliance with follow-up by colonoscopy; complications of colonoscopy; pre-clinical and interval cancer detection; and follow-up after polyp detection.

Simulated individuals within a targeted age range during the period of recruitment were eligible for FOBT screening, provided that they had no history of CRC. The recruitment period was either the year 2000 to generate a fixed cohort or the period 2000 to 2024 to generate a dynamic cohort. Fixed cohorts were used to simulate clinical trial conditions and to estimate the mortality reduction and cost-effectiveness of screening. Dynamic cohorts, which take into account the changing population structure, were used to determine the potential impact on resources, such as the volumes of FOBTs and colonoscopies that would be required in a population-based screening program.

Participation was simulated for first and subsequent invitations to FOBT screening and for follow-up with colonoscopy. Only individuals participating in the first screening round were re-invited. Participation in a subsequent round did not otherwise depend on participation in the previous round. Individuals not complying with follow-up by colonoscopy were no longer screened.

The recruitment strategy plays an important role in the overall participation rate and cost of a screening program. It was proposed that recruitment through media promotion, letters of invitation and visits to family physicians could achieve a 67% participation rate of the target population in Canada. It was estimated that those complying with screening would make an average of 1.5 visits to consult with physicians and to receive the FOBT test kit.

Outcomes of the FOBT, as shown in Figure 1, were simulated using the sensitivity and specificity estimates from the clinical trials. For simulated individuals identified in the microsimulation as having pre-clinical cancer, the sensitivity estimate was applied to assign a true positive or false negative outcome. A false negative outcome meant that the cancer was missed but would be detected clinically before the next screen as an interval cancer. For simulated individuals not having pre-clinical cancer at the time of screening, the specificity was applied to assign a true negative or false positive outcome. The presence of pre-clinical cancer potentially detectable by FOBT was simulated by calculating the probability of incidence of CRC within the next two years for biennial screening or within one year for annual screening.

FIGURE 1
FOBT screening paths

Only cases with positive test results were offered further consultation with a gastroenterologist and follow-up investigation by colonoscopy. The colonoscopy was assumed to be 95% sensitive¹¹ and 100% specific in detecting CRC. Complications associated with colonoscopy were modeled for perforation (0.17%), hemorrhage (0.03%) and death (0.02%).¹² After a negative colonoscopy, participants were exempt from screening for 10 years provided that no polyps were found. When polyps were found (but no cancer), follow-up colonoscopies were performed at 3-, 5- and 10-year intervals according to guidelines^13,14 and expert opinion.

It was assumed that colonoscopy would detect polyps greater than 1 cm in diameter and that removal of polyps would have no impact on the incidence of CRC, consistent with findings in the Funen trial over the 10-year follow-up period. The prevalence of polyps greater than 1 cm was assumed to increase linearly from 3% at age 50 to 5% at age 70 to 5.5% at age 80.¹⁵

Stage was assigned according to how the cancer was detected (Table 1). The stage distribution for the Canadian reference (control) population was derived from a chart review at the Ottawa Regional Cancer Centre. The stage distribution for a hypothetically screened population in Canada was estimated by applying the relative stage shift observed in the RCTs to the Canadian reference stage distribution.

TABLE 1
Estimated Canadian stage distributions according to how cancer was detected

^a Derived from an Ottawa chart review
^b Estimated from Funen trial observations
^c Estimated from Minnesota trial observations

The improved stage distribution accounted for part, but not all, of the improved survival observed in the trials. Relative risks were applied to the survival of the reference group for cancers detected from the first screen (RR = 0.53), subsequent screen (RR = 0.62), in the interval (RR = 0.88) and in the non-participants (RR = 1.04).¹⁶

Costs related to CRC screening were difficult to estimate since a program does not exist in Canada. Table 2 shows the estimated cost of screening by component together with a higher cost option to take into account uncertainty. Treatment costs were included from the base CRC model. No indirect costs were modelled.

TABLE 2
Estimated costs of screening program by component 

^a Estimated from Cancer Care Ontario (2000) (unpublished report)
^b Estimated from the Ontario Health Insurance Plan
^c Based on quotes from private laboratories
^d Estimated from Day Procedure Group cost lists (Manitoba Health Services and Alberta Standard Cost List for Health Economics Evaluations) and estimates from Prince Edward Island by NCCCS member Dr. Don Clark

The cost-effectiveness ratio was calculated as the incremental cost incurred divided by the incremental life-years saved due to screening. Cost-effectiveness ratios less than $40,000 per life-year saved are generally considered cost-effective.¹⁷ Discounting for costs and life-years was performed at 0%, 3% and 5%. All costs were in Canadian dollars.

Simulating control and screen groups

We used a sample of approximately 7 million to minimize the random error associated with the simulation. This sample consisted of two identical cohorts, a reference (control) cohort and a screen cohort. The life histories of individuals in the screen cohort were identical to those of the reference cohort until they became eligible for screening. The screen cohort was then subjected to the modelled screening protocol. The impact of screening was evaluated by comparing outcomes from the screened cohort with outcomes from the unscreened reference cohort. This approach was repeated for each screening scenario evaluated. The main outcomes of analysis were the reduction in mortality from CRC, life expectancy gains, cost-effectiveness and volumes of FOBTs and colonoscopies generated by screening.

Validation

To validate the screening component of POHEM we used the characteristics of the Funen trial (Table 3) to reproduce the trial's observed mean mortality reduction. The stage distribution observed in the Funen trial was used to assign stage according to how the cancer was detected. The results were standardized to the age group and sex of the population structure of the Funen trial.

TABLE 3
Simulated scenarios of screening with FOBT Hemoccult II (nonrehydrated) 

Frequency	Validationa	Canada
	Biennial	Biennialb	Annual	Biennial	Biennial	Biennial
Participation rate (%)	67	67	67	50	67d	100
Re-screen rate (%)	93	93	93	93	93	100
Colonoscopy compliance (%)	89	89	89	89	89	100
Age group	45-75	50-74	50-74	50-74	50-74	50-74
Sensitivity (%)	51	51	80.8c	51	51	51
Specificity (%)	98	98	97.7c	98	98	98

^a Based on Funen trial results
^b Canadian core scenario
^c Based on Minnesota trial results
^d Ramp-up scenario in which the target participation rate of 67% was reached over 5 years

Canadian screening scenarios

Table 3 shows the parameters chosen for each of the Canadian scenarios. Parameters for the Canadian core scenario were chosen to be as similar as possible to the conditions of the population-based Funen trial (i.e., the validation parameters), since the relation between trial participation patterns and observed mortality reduction may not extend to other participation rates. Thus, the core scenario was characterized by biennial screening, 67% participation in the first screen round, 93% participation in subsequent screening rounds, 89% compliance with follow-up by colonoscopy, an FOBT sensitivity of 51% and a specificity of 98%. To reflect the Canadian context, we used a target age range of 50 to 74, the estimated Canadian stage distributions and the Canadian population age structure.

The annual screening scenario for Canada used estimates of sensitivity and specificity observed in the Minnesota trial (for the subgroup of nonrehydrated FOBTs). To evaluate the impact of participation on mortality reduction and cost-effectiveness, participation in biennial screening was reduced from 67% to 50%. To study the impact on resources, participation was ramped up gradually to the target level of 67% over five years. Finally, to assess potential life expectancy gains, a cohort of eligible individuals aged 50 was simulated to participate fully in all aspects of biennial screening until age 74.

Results

Validation

The simulated validation scenario generated a mortality reduction of 17.9% (95% CI: 16.9%-18.9%). This was consistent with the mortality reduction in the Funen trial of 18% (95% confidence interval [CI]: 1%-32%) after 10 years of follow-up. The confidence intervals were tighter in our model because the sample was much larger (7 million versus 60,000). The validation increased our confidence in our simulation of the impact of CRC screening in Canada.

Mortality reduction and cost-effectiveness

When biennial screening was simulated under the assumptions of the Canadian core scenario, there was an estimated 16.7% (95% CI: 15.8%-17.6%) reduction in the 10-year CRC mortality. This result was lower than observed in the Funen trial, reflecting a more restricted target age range. Figure 2 shows the projected change in mortality reduction over time. It peaked during the first few years of screening because of the high number of prevalent cases that could be detected and then steadily declined, since improved survival did not necessarily preclude mortality from CRC. In other words, death from CRC was postponed for some individuals. For the remaining individuals, the avoidance of CRC death was replaced by death from another cause at a later time, as illustrated by the lowest curve in Figure 2.

Deaths due to the complications of colonoscopy had minimal impact on the estimated mortality reduction: for every 178 CRC deaths avoided in the simulated cohort, one death due to complications was incurred. The overall impact on the cohort of the life-years gained and lost was reflected in an estimated life expectancy gain of 0.040 years (95% CI: 0.038-0.042) or 15 days (Table 4).

FIGURE 2
Estimated mortality reduction over time, Canadian core scenario, screened from 2000 to 2025

Note: Based on a simulated cohort of eligible individuals aged 50-74 recruited in the year 2000 (n = 7,001,322)

TABLE 4
Mortality reduction and cost-effectiveness of biennial, annual and reduced participation screening scenarios (relative to no-screening option)

Note: Based on a simulated cohort of eligible individuals aged 50-74 recruited in the year 2000 (n = 7,001,322)

Similar trends in mortality reduction were observed for the other scenarios. After 10 years of annual screening, CRC mortality was reduced by 26% and life expectancy increased by 0.065 years (24 days). When participation in biennial screening was reduced from 67% to 50%, the 10-year CRC mortality reduction dropped to 10.0% and the life expectancy gain dropped to 0.025 years (9 days). The 10-year CRC mortality reduction dropped by an additional 1.4% when the participation rate in subsequent screening rounds was reduced from 93% to 80% and compliance with colonoscopy was reduced from 89% to 80%.

The cost per life-year gained from biennial screening was $11,907, and this increased to $13,497 under annual screening (discounted at 5%). Both biennial and annual screening remained cost-effective under the high-cost sensitivity analysis. Biennial screening was less cost-effective ($15,688) when the participation rate was reduced from 67% to 50%.

Additional analyses were performed to evaluate the age at which to start and end screening. Using five-year increments, we evaluated start ages from 40 to 60 and end ages from 60 to 90. The increased cost of screening before age 50 was not warranted, given the small gain in life expectancy, and screening after age 75 showed no significant gains in life expectancy. Starting to screen at age 50 and ending at age 74 was shown to be more cost-effective than starting later or ending earlier.

Impact on resources

The impact on resources was evaluated by modelling the recruitment of all eligible individuals aged 50 to 74 to the screening program over the years 2000 to 2024. Biennial screening under the core scenario generated an estimated 2.8 million FOBTs and 55,845 colonoscopies per year on average (Table 5). The average cost of the screening program was $112 million per year over 25 years of screening (discounted at 5%). Physician visits accounted for 63% of this cost and overheads for 8% (Figure 3). The cost of screening represented almost one-quarter of the total cost of detecting and treating colorectal cancer. Early detection through screening reduced the cost of treatment by 4.8%.

TABLE 5
Impact on resources per year from biennial, annual and ramped-up screening (averaged over 25 years of screening program)

^a    Ramp-up scenario in which the target participation rate of 67% was reached over 5 years
Note: Based on a simulated cohort of eligible individuals aged 50-74 recruited from year 2000-2024

FIGURE 3
Estimated average annual cost of biennial screening by component over 25 years of a simulated program

Annual screening nearly doubled the demand on resources compared with biennial screening, as illustrated in Figure 4. The average number of FOBTs rose to 4.9 million per year, and the average number of colonoscopies increased to 111,654 per year. Also shown in Figure 4 is the phased-in demand for resources when the target participation rate of 67% was reached gradually over the first five years of the program (ramp-up scenario). The impact of the aging baby boomers was reflected in the increasing volume of colonoscopies projected over the 25 years of screening.

FIGURE 4
Estimated annual volume of colonoscopies required for annual, biennial and biennial ramp-up screening scenarios for selected years

Full participation

When a cohort of simulated individuals aged 50 who participated fully in all aspects of biennial screening until age 74 were followed until death, the cohort life expectancy increased by 0.10 years (37 days). An individual deemed to develop CRC within this cohort gained an estimated 1.75 years of life. The lifetime incidence of CRC rose slightly, by 0.5%, because screening detected cancer in some individuals who otherwise would have died from another cause before clinical detection. The lifetime mortality rate from CRC dropped from 3.0% to 2.3%. The probability of dying as a result of the complications of a colonoscopy was 0.005%; 0.043% suffered a perforation and 0.008% hemorrhaged as a result of the colonoscopy. Over the 25 years of screening, the probability of having a colonoscopy was 25%.

Discussion

Screening for colorectal cancer with FOBT followed by colonoscopy for those with positive test results was cost-effective for the Canadian scenarios simulated relative to commonly accepted thresholds for health interventions.⁷ However, the potential effectiveness of screening would greatly depend on reaching the targeted participation rate of 67%. To put this into perspective, participation rates in organized breast cancer screening programs in Canada in 1997-98 were well below the target of 70%, with estimates ranging from 12% to 55% across provinces after as much as 10 years of program implementation.¹⁸ A pilot program is currently under way in Ontario that may help indicate attainable participation rates for CRC.¹⁹

As with participation rates, other model parameters, such as the sensitivity and specificity of the FOBT, have uncertainty associated with their estimates that we have not evaluated. Similarly, the model was constructed to reproduce the mean estimate of the mortality reduction but does not take into account the full uncertainty reported in the RCTs. Further analyses could be done to estimate the potential impact of this uncertainty.

A physician-based recruitment strategy will place additional burden on family physicians and may require additional doctors or other trained health care providers to meet the demand. Finding the resources to perform the increased number of colonoscopies may also be a challenge. According to simulation results, biennial screening would increase the demand for colonoscopies by 15% over current-use estimates (year 2000 estimates projected from 1995-96 Canadian Institute for Health Information estimates). Annual screening could double this demand. Given the potentially large number of colonoscopies and FOBTs that would be required in a fully operational screening program in Canada, issues of quality assurance become paramount,²⁰ especially since the potential for death from the complications of colonoscopy among otherwise healthy people raises ethical concerns.

There may also be ethical issues related to the impact of screening on quality of life. False positive FOBT results may increase anxiety in otherwise healthy individuals. Screening may adversely affect the quality of life, given that cancers are detected earlier. Patients live longer with knowledge of their disease and, further, the life-years gained may not be lived in perfect health. On the other hand, the life-years gained may be lived in less severe states of the disease, as suggested indirectly by the reduced cost of treatment in the simulated screening cohort.

Polyp removal was assumed to have no impact on the incidence of CRC, since none was observed over the 10 years of follow-up in the Funen trial. However, more recent results from 18 years of follow-up of the Minnesota trial showed lower than expected incidence rates of CRC,²¹ as did an earlier analysis in the National Polyp Study,²² suggesting a possible link with polyp removal. Consequently, our analysis may have underestimated the benefits of screening in this regard.

The estimates from this study were intended to be representative of Canada but may not reflect provincial variations. For instance, resource limitations may lead some jurisdictions to follow up a positive FOBT result with barium enema instead of colonoscopy. Further analysis would be required to fully explore the potential impact of barium enema as an alternative follow-up procedure.

The current model is easily adapted to using various primary screening modalities but would require strong sources of evidence, such as RCTs, to obtain estimates of test efficacy. Regardless of the modality, acceptance by affected communities would remain critical for the feasibility and effectiveness of a program, as this analysis showed for FOBT.

This analysis provided evidence-based responses to address gaps in information identified by the National Committee on Colorectal Cancer Screening and was valuable in supporting the development of recommendations by the National Committee. It demonstrated the usefulness of modelling in the decision-making process.

Acknowledgements

The authors would like to acknowledge the contribution of the Modelling Subcommittee set up by the NCCCS, namely Dr. Andrew Coldman, Dr. Jean-François Boivin, Dr. Dan Sadowski, Dr. Heather Bryant, Dr. Paul Villeneuve and Dr. François-Pierre Dussault. They provided invaluable guidance in determining the screening protocol to be modelled as well as the adoption of parameters appropriate for the Canadian setting. We would also like to thank Rolande Bélanger for her administrative assistance.

References

1. National Cancer Institute of Canada. Canadian Cancer Statistics 2002. Toronto, Canada, 2002.

2. National Cancer Institute. Surveillance, Epidemiology, and End Results (SEER) Program Public-Use Data (1973-1998). Bethesda, MD: DCCPS, National Cancer Institute, Surveillance Research Program, Cancer Statistics Branch, U.S. Department of Health and Human Services. Released April 2001, based on the August 2000 submission.

3. Kronborg O, Fenger C, Olsen J, Jorgensen OD, Sondergaard O. Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet 1996;348:1467-71.

4. Hardcastle JD, Chamberlain JO, Robinson MH et al. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet 1996;348:1472-77.

5. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. Minnesota Colon Cancer Control Study. N Engl J Med 1993;328:1365-71.

6. Wolfson MC. POHEM - a framework for understanding and modelling the health of human populations. World Health Stat Q 1994;47:157-76.

7. Berthelot J-M, Le Petit C, Flanagan W. Use of longitudinal data in health policy simulation models. Proceedings of the Section on Government Statistics and Section on Social Statistics. American Statistical Association, 1997:120-29.

8. Will BP, Nobrega KM, Berthelot JM, et al. First do no harm: extending the debate on the provision of preventive tamoxifen. Br J Cancer 2001;85:1280-88.

9. Berthelot JM, Will BP, Evans WK, Coyle D, Earle CC, Bordeleau L. Decision framework for chemotherapeutic interventions for metastatic non-small-cell lung cancer. J Natl Cancer Inst 2000;92:1321-29.

10. Will BP, Berthelot JM, Le Petit C, Tomiak EM, Verma S, Evans WK Estimates of the lifetime costs of breast cancer treatment in Canada. Eur J Cancer 2000;36:724-35.

11. Rex DK, Rahmani EY, Haseman JH, Lemmel GT, Kaster S, Buckley JS. Relative sensitivity of colonoscopy and barium enema for detection of colorectal cancer in clinical practice. Gastroenterology 1997;112:17-23.

12. Habr-Gama A, Waye JD. Complications and hazards of gastrointestinal endoscopy. World J Surg 1989;13:193-201.

13. Desch CE, Benson AB, III, Smith TJ, et al. Recommended colorectal cancer surveillance guidelines by the American Society of Clinical Oncology. J Clin Oncol 1999;17:1312.

14. Winawer SJ, Fletcher RH, Miller L. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology 1997;112:594-642.

15. Ransohoff DF, Lang CA. Screening for colorectal cancer. N Engl J Med 1991;325:37-41.

16. Mapp TJ, Hardcastle JD, Moss SM, Robinson MH. Survival of patients with colorectal cancer diagnosed in a randomized controlled trial of faecal occult blood screening. Br J Surg 1999;86:1286-91.

17. Laupacis A, Feeny D, Detsky AS, Tugwell PX. How attractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations. Can Med Assoc J 1992;146:473-81.

18. Health Canada. Organized Breast Cancer Screening Programs in Canada, 1997 and 1998 Report. Ottawa: Minister of Public Works and Government Services Canada, 2001.

19. Cancer Care Ontario. Pilot Project Comparing Primary Care Physician Recruitment to Public Health Unit Program Recruitment for Colorectal Cancer Screening by Fecal Occult Blood Testing in Ontario, March 3, 2003. URL: <http://www.cancercare.on.ca/pdf/colorectalpilot.pdf> , accessed August 27, 2003.

20. Haseman JH, Lemmel GT, Rahmani EY, Rex DK. Failure of colonoscopy to detect colorectal cancer: evaluation of 47 cases in 20 hospitals. Gastrointest Endosc 1997;45: 451-5.

21. Mandel JS, Church TR, Bond JH, et al. The effect of fecal occult-blood screening on the incidence of colorectal cancer. N Engl J Med 2000;343:1603-7.

22. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993;329: 1977-81

Author References

William M Flanagan, Christel Le Petit, Jean-Marie Berthelot, Kathleen J White, Health Analysis and Measurement Group, Statistics Canada, Ottawa, Ontario, Canada

B Ann Coombs, Elaine Jones-McLean, Public Health Agency of Canada, Health Canada, Ottawa, Ontario, Canada

Correspondence: Kathy White, Health Analysis and Measurement Group, Statistics Canada, 24-A RH Coats, Ottawa, Canada, K1A 0T6; Fax: (613) 951-3969; E-mail: kathleen.white@statcan.ca

For readers interested in more information on colorectal cancer screening, Health Canada's Technical Report for the National Committee on Colorectal Cancer Screening and the National Committee on Colorectal Cancer Screening's Recommendations for Population-based Colorectal Cancer Screening can both be found on Health Canada's website at: <http://www.hc-sc.gc.ca/pphb-dgspsp/publicat/ncccs-cndcc/index.html>

1. Canadian stage and survival estimates were derived from a chart review of 700 patients with a diagnosis of CRC in the hospital system in Ottawa, Canada, in 1991-1992 (national data unavailable). Survival rates are comparable to those obtained from SEER*Stat4.0.²

[Previous] [Table of Contents] [Next]