Volume 24
Number 4
2003

[Table of Contents]

Lifetime costs of colon and rectal cancer management in Canada

Jean Maroun, Edward Ng, Jean-Marie Berthelot, Christel Le Petit, Simone Dahrouge, William M Flanagan, Hugh Walker and William K Evans

Abstract

Colorectal cancer is the second leading cause of cancer-related mortality among Canadians. We derived the direct health care costs associated with the lifetime management of an estimated 16,856 patients with a diagnosis of colon and rectal cancer in Canada in 2000. Information on diagnostic approaches, treatment algorithms, follow-up and care at disease progression was obtained from various databases and was integrated into Statistics Canada's Population Health Model (POHEM) to estimate lifetime costs. The average lifetime cost (in Canadian dollars) of managing patients with colorectal cancer ranged from $20,319 per case for TNM stage I colon cancer to $39,182 per case for stage III rectal cancer. The total lifetime treatment cost for the cohort of patients in 2000 was estimated to be over $333 million for colon and $187 million for rectal cancer. Hospitalization represented 65% and 61% of the lifetime costs of colon and rectal cancer respectively. Disease costing models can be important policy-relevant tools to assist in resource allocation. Our results highlight the importance of performing preoperative tests and staging in an ambulatory care setting, where possible, to achieve optimal cost efficiencies. Similarly, terminal care might be delivered more efficiently in the home environment or in palliative care units.

Key words: colorectal cancer; costing model; direct care costs; lifetime cost; micro-simulation

Introduction

As we begin the 21^st century, the Canadian health care system is experiencing serious fiscal constraints. Costly technological advances and an aging population are creating daunting challenges for health care. To ensure the best health outcomes, health policy decision-makers need to set priorities for resource allocation. Economic studies of the burden of illness can make a valuable contribution to policy development. Such studies can estimate the present and future cost impact of diseases on society and can also be used as a benchmark against which the cost-effectiveness of new medical treatments or prevention programs can be evaluated.¹

In 2002, colorectal cancer (CRC) was the third most common malignancy affecting Canadian men and women and the second leading cause of cancer-related death. Invasive CRC was diagnosed in an estimated 17,600 Canadians in 2002, and more than 6,600 individuals were expected to die from this disease in the same year.² A comprehensive model of the management of CRC by site (colon and rectum) was developed and incorporated into Statistics Canada's Population Health Model (POHEM)^3,4 to estimate the total lifetime direct cost of treating the cohort of patients with CRC* in 2000.

Methods

To build the economic model incorporating all phases of CRC and their management required data from a number of databases (see Appendices A and B for the main sources and their use). The main source of data on clinical management came from a retrospective review of 700 charts of patients whose CRC was diagnosed in 1991-92 in the Ottawa hospital system (the Ottawa Chart Review). This review provided data on demographic characteristics of patients, staging work-up, treatment (surgery, chemotherapy and radiotherapy) and outcome information.

The management of CRC was categorized into three periods: initial treatment, treatment of local recurrence and management of metastasis. The initial treatment period captured information from diagnosis, the treatment phase and the well-patient follow-up phase (stages I-III). This period ended at five years, and patients were considered cured if they remained disease free. The local recurrence period captured information from that diagnosis, a three-month initial treatment phase and up to five years of active care, unless metastasis was documented. The metastasis period (which includes those whose cancer was initially diagnosed as stage IV) captured information from that diagnosis, a six-month initial treatment phase, and up to five years of active care. Death can occur at any point in these periods, and the three months before any CRC-related death were considered as a period of terminal care.

The stage at the time of primary surgical treatment is the most important determinant of prognosis. Disease progression was modelled from the Ottawa Chart Review for each stage and site at diagnosis, on the basis of the TNM classification.⁵ For patients with stage I, II or III cancer, the three transitions modelled were diagnosis to local recurrence, to distant recurrence (or metastasis) and to death. Once a local recurrence had occurred, the only transitions modelled were to metastasis or to death. For those with metastatic disease at diagnosis (including stage IV), the only transition modelled was to death.

The Ottawa Chart Review was also used to estimate the type and frequency of treatment procedures by stage and disease site for each period. However, in light of the recent emergence of new chemotherapy drugs for the treatment of CRC, a Canada-wide survey of oncologists was conducted in 1998 to re-evaluate the proportion of patients receiving chemotherapy (and/or radiotherapy) and the type of chemotherapy selected.⁶ These survey data replaced the Ottawa Chart Review data for all chemotherapy treatment information at initial diagnosis. The duration of surgery-related hospitalization at initial treatment (as well as at progression) was estimated using Statistics Canada's national person-oriented information (POI) database of hospital discharges⁷ for each type of surgery identified by the Ottawa Chart Review. All hospital admissions up to 30 days before and 60 days after the admission date for surgery were included as a proxy for any hospitalization for preoperative tests and procedures as well as readmissions due to surgical complications.

The Canadian Cancer Registry (CCR)⁸ and population estimates from Statistics Canada⁹ were used to estimate age-sex specific incidence rates of colon and rectal cancer in Canada. At the time of analysis, the only reliable data available were for 1995, and therefore these data were used as a proxy for the Canadian situation in 2000. Survival was modelled with a piece-wise Weibull function⁴ using the Lifetest procedure from SAS.¹⁰

Cost assessment

Unless otherwise noted, all costs were determined in 1998 Canadian dollars. The economic analysis was carried out from the perspective of the government as payer in a universal health care system and did not take into account indirect costs such as lost income. The principal source of cost information on physician fees, procedures, laboratory tests and CRC-related surgeries was the 1998 Ontario Health Insurance Plan (OHIP)^†.¹¹ The unit cost components for the major items are presented in Appendix C.

To cost specific items (e.g., surgical procedures) for which more than one type of procedure could be used, the cost associated with the procedure was derived by summing the products of the cost, from OHIP, of each type of procedure multiplied by the proportion of its use according to the treatment algorithm derived from the Ottawa Chart Review. Similarly, the overall cost of a specific item was derived by multiplying the average weighted cost of that item by the proportion of the population using that resource. For example, the weighted cost of chemotherapy for stage III colon cancer ($4,120) was derived by multiplying the average weighted cost of chemotherapy for that population ($4,709) by the estimated proportion receiving it (87.5%).

The Population Health Model (POHEM)

POHEM is a framework for organizing health information as well as a micro-simulation model to estimate disease progression, treatment consequences and lifetime cost in a large population. While estimates of this lifetime cost can be derived crudely in cell-based macro-simulation models, more detailed macro-modeling tends to explode the number of cells required and render this approach impractical. By contrast, micro-simulation models use relatively simple random processes to build complex synthetic life paths drawn from distributions of empirical data. The model is not inherently limited in the level of detail that can be included - only, rather, by the availability of clinically relevant descriptions of disease processes and the data required to model them.

Practically, POHEM simulates and synthesizes a sample of individual health and socio-economic life histories to re-create the age-sex and tumour stage distribution of the year 2000 cohort of new CRC patients in Canada. Each synthetic individual is aged and subjected probabilistically to demographic events, disease onset and progression, and use of health care resources, based on incidence, treatment and disease progression incorporated in POHEM. Whether, for a simulated individual, colon or rectal cancer develops is determined by a random number process as a function of the estimated probability of incidence. On the basis of the stage distribution by site, each individual with a diagnosis of CRC in the simulated population is randomly assigned a stage at diagnosis. The stage determines the treatment modalities according to the derived initial treatment algorithms for both colon cancer and rectal cancer. Disease progression and outcomes are determined by comparing random numbers with the derived survival function. The distribution of recurrence and treatment are also incorporated to estimate costs over time. The simulation sample size of 32 million individuals ensures that the Monte Carlo error is small relative to the model outputs of interest.

Sensitivity analysis

"Discounting" takes into consideration the fact that future costs and benefits need to be reduced or discounted, as the present value of today's dollars is greater than the present value of future dollars. Sensitivity analyses using yearly discount rates of 0%, 3% and 5% were carried out.

Results

Disease incidence, stage and survival

Table 1 provides a summary of disease evolution in patients with colon and rectal cancer based on POHEM. The overall expected number of years lived was 7.6 for colon and 7.4 for rectal cancer. The risk of local recurrence did not increase with stage of disease at diagnosis. However, there was a strong positive relation between stage and the risk of metastatic disease as well as the risk of disease-specific death, and an inverse relation between stage and survival duration. Overall, 52% and 57% of colon and rectal cancer patients respectively die of their cancer over a lifetime, and the majority of deaths occur within five years.

TABLE 1
Summary of colon and rectal cancer disease progression based on POHEM

^a Estimates over the lifetime of CRC patients
^b CRC = colorectal cancer

Estimated costs of initial treatment period

All patients with cancer diagnosed at stages I to III were treated with surgery^‡. Adjuvant chemotherapy was given to 19% and 88% of patients with colon cancer stages II and III respectively. The use of radiotherapy was limited to patients with rectal cancer. Combination chemotherapy and radiotherapy was given to 60% and 81% of patients with rectal cancer stages II and III respectively. Active care was given to 70% and 83% of patients with stage IV colon and rectal cancer respectively.

Table 2 shows that the average costs associated with the diagnosis and the treatment phases of the initial treatment period for stages I to IV colon and rectal cancer ranged between $11,598 and $19,742; 53% to 86% of these costs were attributable to hospitalization. For stage IV, 30% to 33% of the cost was attributable to first and second line chemotherapy. The cost of diagnosis and staging was minor in comparison, ranging between $375 and $568.

Well-patient follow-up practice is dependent on stage. The total annual follow-up cost per person decreased over time, from approximately $400 in year one to $150 in year five with an overall lifetime average of $615 to $908 for stages I to III (results partially shown in Table 3).

TABLE 2
Average per patient cost of initial treatment for colon and rectal cancers by stage/state and component (1998 Cdn $)

N.B. Numbers may not add up to the total because of rounding.

TABLE 2 (cont'd)
Average per patient cost of initial treatment for colon and rectal cancers by stage/state and component (1998 Cdn $)

N.B. Numbers may not add up to the total because of rounding.

TABLE 3
Average and total lifetime costs ($000) of disease management of colon and rectal cancer by stage at presentation and components (1998 Cdn $)

^a The POHEM simulated result for total initial treatment cost is slightly different from the total cost shown in Table 2 because of the probabilistic nature of simulation.
^b Terminal care refers to costs in the last 3 months of life
Numbers may not add up because of rounding.

Estimated costs of the local recurrence and metastatic phases

Table 2 shows the estimated cost of disease management for local recurrence and for metastatic disease from POHEM. The average cost of diagnosing and treating local recurrence was $6,611 for colon and $6,708 for rectal cancer patients. Hospitalization accounted for 69% and 56% of these costs for colon and rectal cancer respectively. In fact, a sizeable proportion of the patients with local recurrence (56% for colon and 68% for rectal cancer) did not undergo surgical resection at recurrence, and thus the hospitalization cost was less than that in the initial treatment phase of stages I to IV. For patients with colon cancer, less than 20% received radiotherapy or chemotherapy at the time of recurrence. However, 53% of rectal cancer patients received radiotherapy, and 12% received chemotherapy. The costs of diagnosis and staging for local recurrence were once again minor in comparison, ranging between $407 and $477 per patient.

The corresponding estimated costs associated with metastatic disease were higher (about $8,200), and hospitalization represented approximately 45% of these costs. The majority of patients did not have surgery (65%), and only 9% had radiotherapy. The mainstay of treatment at metastasis was chemotherapy. Based on the Ottawa chart review, 46% received first line chemotherapy, and less than 20% of them went on to receive second line therapy. While the chemotherapy cost for second line therapy was $19,314 per person, only 8% received this treatment, and thus the weighted cost was about $1,550. The chemotherapy practice survey showed that new and/or experimental regimens, such as irinotecan, are often used, and are costly. In addition, with the emerging new effective combinations, a larger proportion of patients are being treated in today's practice.

Estimated lifetime costs of CRC care in Canada

Table 3 shows the POHEM results for the individual components of the lifetime costs of providing care to 16,856 patients with CRC in 2000. It also shows that the total lifetime cost of treatment for all patients with colon and rectal cancer (stages I to IV) was over $333 million and $187 million respectively. By phase of illness, almost 80% of these costs were for initial treatment or were incurred during the terminal care phase. Initial treatment accounted for 49% of the total cost of colon cancer, followed by terminal care (28%), mainly as a result of the large amount of hospitalization in these two phases. The figures for rectal cancer were similar. The average cost per case for all stages of colon and rectal cancer was $29,110 and $34,475 respectively and ranged from a low of $20,319 for patients with stage I colon cancer to a high of $39,182 for those with stage III rectal cancer.

Figure 1 presents the cost components for all stages of colon and rectal cancer by intervention and shows that, as expected, hospitalization represented a high proportion of the total cost, at 65% and 61% of the lifetime costs of care delivery for colon and rectal cancer respectively. Since most of the cost of treatment was incurred within the first five years from the base year of 2000, discounting had little impact on the total lifetime cost of treating colorectal cancer in Canada. The corresponding total cost of treatment for colon and rectal cancer respectively was $325 million and $182 million at a 3% discount rate, and $319 million and $178 million at a 5% discount rate.

FIGURE 1
Distribution of per patient lifetime costs of colon and rectal cancer by intervention - all stages

Discussion

This article describes the lifetime cost of care for Canadian patients with a diagnosis of CRC. It is the result of collaboration between Statistics Canada, the Ottawa Regional Cancer Centre, Queen's University and Cancer Care Ontario. Estimates of the lifetime costs of lung and breast cancer have been previously reported.^12-15

Appropriate data are crucial in performing a disease burden study. The CRC model has been developed using a number of Canada-wide databases, such as Statistics Canada's person-oriented hospital discharge database and the Canadian Cancer Registry. Where national data were not available, provincial and regional data were used - for example, the Manitoba database¹⁶ for follow-up patterns after CRC treatment and the Ottawa Chart Review for disease progression and treatment algorithms. Efforts were made to obtain the most up-to-date information. For instance, where existing data on chemotherapy became outdated because of the availability of new drugs, results from a survey of oncologists provided information on the most recent chemotherapy treatment. National initiatives to integrate data on cancer patients and to develop data definitions will hopefully facilitate future data collection.

Our results are similar to those obtained with the lung and breast cancer models. Hospitalization represented 65% and 61% of the lifetime cost of care delivery in colon and rectal cancer, as compared with 76% and 63% for lung and breast cancer. Initial treatment at diagnosis represented 49% of the costs for both colon and rectal cancer, as compared with 53% and 34% for lung and breast cancer. Similarly, terminal care represented 29% and 26% for colon and rectal cancer, as compared with 39% and 27% for the other two sites. These results highlight the importance of performing preoperative tests and staging in an ambulatory care setting, where possible, to achieve optimal cost efficiencies. Similarly, terminal care might be delivered more efficiently in the home environment or in palliative care units.

Chemotherapy costs have been refined to include the effects of toxicity and overhead costs. In the future, a comprehensive analysis of the cost of chemotherapy delivery should include out-of-pocket and other costs to patients and care givers. Finally, the chemotherapy costs are most likely underestimates in view of the recent development of effective agents such as irinotecan and oxaliplatin.¹⁷ However, since the chemotherapy cost is only about 12% of the total cost of CRC treatment, this difficulty in capturing current chemotherapy practice patterns will not have a significant impact on lifetime cost.

The data from our model cannot be compared directly with data from other countries. Variation in costs may be due to differences in survivorship, treatment approaches, the nature of the health care systems and the patient populations included in the analyses^§.

We compared our results with two US studies by Taplin¹⁸ and by Brown,¹⁹ and our hospital costs with a study in Nova Scotia.²⁰ While there is agreement that the management of CRC can be divided into initial therapy, continuing care and terminal care, comparison is difficult because of the differences in the definitions of these phases. For example, in our model, initial treatment included the first three months after diagnosis, whereas both Taplin and Brown used the first six months instead. Similarly, we defined terminal care as the last three months of life, whereas Taplin and Brown defined it as six and 12 months.

Brown, using the claims payment information from the SEER-Medicare data in the US, estimated the cost during the initial, continuing and terminal phases of CRC (in $US) to be $18,100 (52.0%), $1,500 (4.3%) and $15,200 (43.7%) respectively. On the basis of enrolment during 1990 and 1991 in the Group Health Cooperative in Washington State, Taplin's estimates for colon cancer only were 52.7%, 4.6% and 42.6% of the total cost for initial, continuing and terminal care respectively. All three studies showed that initial care represented almost half the costs. However, our proportions for continuing care (24%) and terminal care (27%) varied considerably from those reported by Brown and Taplin, possibly because of the difference in definition.

The Nova Scotia study used administrative data to estimate the hospital costs incurred by a population-based cohort of CRC cases up to three years after diagnosis. The length of stay and the hospital-specific per diem rates were used as the measures of resource use. Our study used the Statistics Canada POI data, the resource intensity weight of CIHI²¹ and an intensive provincial costing project²² to estimate the hospital length of stay and per diem respectively. Similar to our findings, the cost of managing CRC in the Nova Scotia study was "significantly less for cases with local spread, highest in the six months around the time of diagnosis and in the final six months of life". However, we consider our estimates to be more realistic, as the Nova Scotia cost did not take into consideration the intensity of care provided during the hospital stay.

Overall, our study shows that the total lifetime cost of treatment for patients with colon and rectal cancer in Canada was over $333 million and $187 million respectively. Sensitivity analysis could have been conducted to examine whether the results were influenced by the estimates or assumptions used. However, given that the total cost of hospitalization, the major cost contributor to the management of CRC, was derived from a national database (POI) as well as from intensive costing projects as mentioned earlier, we felt that the data were sufficiently valid and robust.

In our study, we used the Ottawa Chart Review stage distribution of colon and rectal cancer at diagnosis, assuming that this reflected the situation across Canada. In verifying this assumption, we found that the Ottawa Chart Review distribution was comparable to that of the Manitoba database for stages I and II. However, as the Manitoba database contained more cases with missing information, the decision was made to use the Ottawa Chart Review distribution.

Regarding the survival analysis, some medical experts may consider our five-year survival of 65% for stage III colon cancer to be too high. This result may be due to the small sample and the specific patient characteristics of the Ottawa Chart Review. We do not think that it has a significant impact on the overall lifetime cost estimation. In a sensitivity analysis, we used rectal stage III as a substitute for colon stage III to arrive at a more acceptable survivorship. The overall cost implication of such a change is a mere 3% in the total lifetime cost of treating CRC.

The weighted average costs associated with the diagnosis, staging and initial treatment of CRC showed a gradient of increasing costs for stages I, II and III as well as higher costs for rectal cancer. This was due to an increased use of new and expensive chemotherapy for the more advanced stages as well as the use of radiotherapy for stages II and III rectal cancer. Hospitalization costs, while a major contributor to the total cost, were similar across most stages. For rectal cancer, costs associated with stage IV cancer were lower than those associated with stage III, because of the lower surgical rate and the lower use of radiotherapy.

Finally, the study was carried out from the perspective of the health care system and therefore incorporated only the direct costs associated with CRC management. It did not consider the costs of other comorbidities and did not include additional costs, such as lost productivity or wages, the costs of home care, prostheses, travel and accommodation, or the costs of caregivers in the home.

In conclusion, disease costing models such as the POHEM CRC model are important policy-relevant tools to assist in resource allocation. This model can guide the analysis of initiatives to optimize the costs of caring for patients with CRC and the evaluation of new management strategies for colon and rectal cancer. It has recently been used to assess the cost-effectiveness of a potential population-based screening program in Canada (see the article by Flanagan, Le Petit, Berthelot et al. in this issue).

Acknowledgements

The long process of developing a disease model requires cooperation from numerous individuals and organizations. We wish to thank Dr. Michael Wolfson, Christian Houle, Phyllis Will, the Health Statistics Division of Statistics Canada and the Manitoba Cancer Treatment and Research Foundation (now Cancer Care Manitoba). We acknowledge the time and effort of the surgical, radiation and medical oncologists across Canada who responded (without financial remuneration) to a questionnaire prepared under the auspices of the Ottawa Regional Cancer Centre (ORCC). We also acknowledge the cost analyses prepared by the Canadian Institute for Health Information, the ORCC, the Ottawa Hospital, General Campus, and the Ontario Case Cost Project. Special thanks go to Dr. Robin Fairfull-Smith, Dr. Laval Grimard, and Dr. Eva Grunfeld, who provided much appreciated medical expertise, and to Louise Fawcett, Louise Pigeon and Rolande Belanger for their research or administrative assistance. We would like to dedicate this research project to Louise Fawcett who died of cancer during the course of this project.

References

1. Gunning-Schepers LJ, et al. Public health models: tools for health policy making at national and European level. Amsterdam: University of Amsterdam, 1999.

2. National Cancer Institute of Canada. Canadian cancer statistics 2002. Toronto, Canada: NCIC, 2002.

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

4. 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.

5. Sobin LH, Wittekind CH, eds. UICC International Union Against Cancer. TNM classification of malignant tumours, 5th ed. New York: Wiley-Liss, Inc, 1997.

6. Ng E, Maroun J, Berthelot J-M, Dahrouge S. Chemotherapy practice patterns for colorectal cancer patients in 1998: a survey of Canadian oncologists. Current Oncol 2001;8:150-58.

7. Statistics Canada. National person-oriented database of hospital discharges, 1995-97. Ottawa, Canada: Statistics Canada, 2001.

8. Statistics Canada. Canadian Cancer Registry. Unpublished data. Ottawa, Canada: Statistics Canada, 1995.

9. Statistics Canada. Annual demographic statistics. Ottawa, Canada: Statistics Canada, 1998 (Cat. No. 91-213-XPB).

10. SAS Institute. SAS/STAT user's guide, version 6, 4^th ed, volume 2. Cary, North Carolina: SAS Institute Inc., 1990.

11. Ontario Ministry of Health. Schedule of benefits, physicians services under the Health Insurance Act. Toronto: Ontario Ministry of Health, 1998.

12. Evans WK, Will BP, Berthelot J-M, Wolfson MC. Diagnostic and therapeutic approaches to lung cancer in Canada and their costs. Br J Cancer 1995;72:1270-77.

13. Earle C, Will BP, Evans WK. The economics of lung cancer treatment and prevention. Primary Care & Cancer 1998; supplement No. 1:33-38.

14. Will BP, Le Petit C, Berthelot J-M, Tomiak EM, Verma S, Evans WK. Diagnostic and therapeutic approaches for nonmetastatic breast cancer in Canada, and their associated costs. Br J Cancer 1999;79:1428-36.

15. Will BP, Berthelot J-M, 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.

16. Sloan J, Nemecek A. Experiences in constructing cancer patient trajectories through the Manitoba health care system. Final report to the Manitoba Medical Services Foundation. Winnipeg, Manitoba: Manitoba Medical Services Foundation, 1995.

17. Berlin J. New directions in the treatment of advanced colorectal cancer. Oncology (Huntington) 2001;15(3 Suppl 5):27-30.

18. Taplin SH, Barlow W, Urban N, et al. Stage, age, comorbidity, and direct costs of colon, prostate, and breast cancer care. J Natl Cancer Inst 1995;87:417-26.

19. Brown ML, Riley GF, Potosky AL, Etzioni RD. Obtaining long-term disease specific costs of care: application to Medicare enrollees diagnosed with colorectal cancer. Med Care 1999;37:1249-59.

20. O'Brien BD, Brown MG, Kephart G. Estimation of hospital costs for colorectal cancer care in Nova Scotia. Can J Gastroenterol 2001;15:43-7.

21. Canadian Institute of Health information. Resource intensity weights, summary of methodology 1996/97. Ottawa: CIHI, 1996.

22. Ontario Case Cost Project. Special retrieval from April 1994 to March 1996 database of 13 Ontario hospitals. Toronto, Ontario, 1998.

23. Earle C, Coyle D, Smith A, et al. (1997) The cost of radiotherapy at an Ontario regional cancer centre. Eur J Cancer 1998;33S4; (abstr. OP9) (suppl. 9).

APPENDIX A
Summary of major data sources^a

^a Most data sources referred to in the article are already included in the references.
^b Earle C, Grunfeld E, Coyle D, et al. Empirically derived standards of care for the follow-up of colorectal cancer patients after potentially curative treatment: practices attitudes and costs. Submitted to Cancer Prevention and Control.

APPENDIX B
List of main data requirements and sources

^a For the frequency and type of utilization of diagnostic procedures for each period, the average of the Ottawa Chart Review and Manitoba database results were used.
^b The cost of the fecal occult blood test was obtained from MDS Nordion, a biomedical company based in Ottawa, Ontario.
^c Well-patient follow-up pattern was not captured in the Ottawa Chart Review. It was derived from a Manitoba database that contains three years of follow-up data. This was supplemented by a survey of oncologists on their management of stage III colorectal cancer patients, which contained information on five-year patterns.

APPENDIX C
Main components of disease management for cost evaluation

Author References

Jean Maroun, Ottawa Regional Cancer Centre, Ottawa, Ontario, Canada

Edward Ng, Jean-Marie Berthelot and Christel Le Petit, Health Analysis and Measurement Group, Statistics Canada, Ottawa, Ontario, Canada

Simone Dahrouge, Ottawa Regional Cancer Centre, Ottawa, Ontario, Canada

William M Flanagan, Health Analysis and Measurement Group, Statistics Canada, Ottawa, Ontario. Canada

Hugh Walker, Queen's University, Kingston, Ontario, Canada

William K. Evans, Cancer Care Ontario, Toronto, Ontario, Canada

Correspondence: Dr. Jean Maroun, c/o Ottawa Regional Cancer Centre, 503 Smyth Road, Ottawa, Ontario, Canada K1H 1C4; Fax: (613) 247-3511; E-mail: jean.maroun@orcc.on.ca

Disclaimer: The views presented in this paper are those of the authors and may not reflect the views of the agencies they represent.

* Since cancer of the anus is managed differently from colon and rectal cancer, anal cancer data were excluded from POHEM.

^† A commissioned study conducted by the Canadian Institute of Health Information concluded that OHIP reimbursements closely approximated average provincial costs in Canada. While this study was on the costs of breast cancer treatment, it is reasonable to assume that Ontario costs would also be close to the provincial average for colorectal cancer.

^‡ Treatment algorithms are available upon request.

^§ Notwithstanding the cautious remark on international comparison, the Canadian costs were found to be lower than the U.S. costs. For example, although the average cost of treating patients with colon cancer as estimated by Taplin¹⁸ was US$28,396, our Canadian estimate was CDN$29,110. We have compared colorectal cancer mortality by stage using the Ottawa Chart Review and the SEER data set from the United States to validate our survival data externally, and the result will be reported in another article in progress.

[Previous] [Table of Contents] [Next]