|
Potential impact of population-based colorectal cancer screening in Canada |
Flanagan W M, Le Petit C, Berthelot J-M, White K J, Coombs B A, Jones-McLean E |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology The potential impact of population-based screening for colorectal cancer (CRC) in Canada, using a faecal occult blood test (FOBT; Hemoccult II. Nonrehydrated) and follow-up with colonoscopy for those with positive FOBT results, was evaluated. The study investigated the impact of both biennial and annual screening. A screened cohort was compared to a control cohort with no screening.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised a hypothetical sample of 7,001,322 people aged between 50 and 74 years with no history of CRC. Only individuals participating in the first round of screening were re-invited. Individuals not complying with follow-up colonoscopy were no longer screened. The sample comprised two cohorts, a screen cohort and a control cohort with no screening. The life histories of the two cohorts were identical.
Setting The setting appears to have been primary and secondary care. The economic study was conducted in Canada.
Dates to which data relate Much of the effectiveness evidence and resource use data were derived from literature published between 1989 and 1999. The price year was unclear. Note: since this abstract was published the authors, whilst acknowledging that the previous statement is factually accurate, have informed us that most cost components were expressed in 1998 Canadian dollars.
Source of effectiveness data The effectiveness data were derived from a review of published studies and estimates based on experts' opinion and/or authors' assumptions.
Modelling The microsimulation Population Health Model (POHEM), constructed by Statistics Canada, was used to estimate the potential impact of population-based screening with FOBT followed by colonoscopy in Canada. POHEM is a model that endeavours to evaluate the health of Canadians and evaluate possible interventions. It includes a detailed model of CRC. The model was adapted to analyse the impact of screening on cancer mortality and resource use in a cohort of approximately 7 million people. The screening programme was implemented for 25 years. The model simulated a fixed cohort, to simulate clinical trial conditions and to estimate the mortality reduction and cost-effectiveness of the strategy. A dynamic cohort was used to determine the potential impact on resources. This had a recruitment period of 2000 to 2024.
Outcomes assessed in the review The outcomes assessed were:
the sensitivity and specificity of the FOBT;
the complications associated with colonoscopy;
compliance with follow-up by colonoscopy;
stage distribution; and
pre-clinical and interval cancer detection.
Sources searched to identify primary studies Criteria used to ensure the validity of primary studies Methods used to judge relevance and validity, and for extracting data Number of primary studies included Although not explicitly reported, at least 6 primary studies were included in the review. No other information on the types of studies used was provided.
Methods of combining primary studies This was not relevant as the primary studies included in the analysis reported different types of outcomes.
Investigation of differences between primary studies The authors do not appear to have investigated differences between the primary studies.
Results of the review The sensitivity of the FOBT was 51% and the specificity was 98%.
Complications associated with colonoscopy were modelled for perforation (0.17%), haemorrhage (0.03%) and death (0.02%).
Compliance with follow-up colonoscopy was 89%.
The estimated Canadian stage distributions according to how the cancer was detected were as follows.
Stage 1: 13% with no screening, 38% with biennial screening, 22% with interval cancer; 14% for non-participants and 22% with annual screening.
Stage 2: 33% with no screening, 38% with biennial screening, 29% with interval cancer. 32% for non-participants and 32% with annual screening.
Stage 3: 27% with no screening, 17% with biennial screening, 28% with interval cancer, 20% for non-participants and 31% with annual screening.
Stage 4: 27% with no screening, 7% with biennial screening, 21% with interval cancer, 34% for non-participants and 15% with annual screening.
The relative risks (RR) of survival of the reference group were calculated 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).
Methods used to derive estimates of effectiveness The authors made assumptions and used experts' opinions to derive other parameters used in the model.
Estimates of effectiveness and key assumptions The sensitivity of colonoscopy was assumed to be 95% and specificity 100%.
Participation was 67% in the first screening round was and 93% in subsequent screening rounds (based on authors' assumptions). Note: since this abstract was published, the authors have informed us that their justification for this assumption was that "it was achieved in the population based Funen trial" (Kronborg et al 1996, see "Other Publications of Related Interest" below for bibliographic details), and therefore "could potentially be achieved in a population-based screening program in Canada".
The assumption that follow-up after polyp detection would be performed at 3-, 5- and 10-year intervals was based on guidelines and experts' opinion.
It was assumed that colonoscopy would detect polyps greater than 1 cm in diameter, and that the removal of polyps would have no impact on the incidence of CRC.
The prevalence of polyps greater than 1 cm was assumed to increase linearly from 3% at age 50, to 5% at age 70 years, to 5.5% at 80 years.
Measure of benefits used in the economic analysis The measures of health benefit used were the reductions in CRC mortality and the gains in life expectancy. These were derived from the microsimulation POHEM using data from published studies and expert opinion.
Direct costs The resource quantities and the costs were reported separately. The quantity/cost boundary adopted was that of the health service. Discounting was performed at rates of 0, 3 and 5%. The costs included in the analysis were for administration and promotion, and were estimated from an unpublished reference. The costs of extra physician visits (per FOBT), consultation (per positive FOBT) and FOBT kits were estimated from the Ontario Health Insurance Plan. The processing costs (per FOBT) were estimated from information provided by private laboratories. The cost of colonoscopy (per positive FOBT or follow up to polyps) was based on Day Procedure Group cost lists and professional opinion. The cost of treatment was obtained from the Canadian CRC base model that was included in the POHEM. The data for this base model were derived from hospital discharge abstracts, billing data and expert opinion. Most cost components (treatment, cost of extra physician visit, FOBT kit, consultation after positive FOBT) were expressed in 1998 Canadian dollars.
Statistical analysis of costs The costs were treated deterministically.
Indirect Costs The indirect costs were not included in the analysis.
Sensitivity analysis A one-way sensitivity analysis was undertaken to investigate uncertainty around the costs and estimates of effectiveness (changes in participation rates). There was little information on how the alternative parameters were derived. The authors acknowledged that other model parameters (e.g. the sensitivity and specificity of the FOBT and the impact of polyp removal on the incidence of CRC) have uncertainty associated with them, although these were not investigated.
Estimated benefits used in the economic analysis With biennial screening of 67% of individuals aged between 50 and 74 years, there was an estimated 16.7% (95% confidence interval, CI: 15.8 - 17.6) reduction in the 10-year CRC mortality. Over 25 years there was a mortality reduction of 14.2%.
With annual screening of 67% of individuals aged between 50 and 74 years, there was an estimated 26.0% reduction in the 10-year CRC mortality. Over 25 years there was a mortality reduction of 22.5%.
Overall, with biennial screening there was an estimated life expectancy gain (undiscounted) of 0.040 years (95% CI: 0.038 - 0.042) or 15 days. With annual screening there was an estimated life expectancy gain of 0.065 years or 24 days.
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 with biennial screening, there was one death due to complications.
Cost results The average cost of the biennial screening programme was $112 million per year over 25 years of screening (discounted at 5%). Early detection through screening reduced the cost of treatment by 4.8%.
The average cost of annual screening programme per year over 25 years was $194 million. Early detection through screening reduced the cost of treatment by 5.8%.
Biennial screening of 67% of individuals aged between 50 and 74 years generated an estimated 2.8 million FOBTs and 55,845 colonoscopies per year, on average. With annual screening there were an estimated 4.9 million FOBTs and 111,654 colonoscopies per year, on average.
Synthesis of costs and benefits The costs and benefits were combined using an incremental cost-effectiveness ratio (i.e. the additional cost to obtain an extra life-year). Compared with no screening, the incremental cost-effectiveness ratio of biennial screening was $11,907 (discounted at 5%). The incremental cost-effectiveness ratio of annual screening was $13,497 (discounted at 5%).
When the costs were increased, the incremental cost-effectiveness ratio was $18,445 (discounted at 5%) with biennial screening and $19,893 (discounted at 5%) with annual screening.
The participation rate was an important determinant of the cost-effectiveness of the screening programme. When the participation rate was reduced from 67% to 50%, the biennial screening became less cost effective ($15,688 discounted at 5%).
Authors' conclusions Screening for colorectal cancer (CRC) in Canada, using the faecal occult blood test (FOBT) followed by colonoscopy for those with positive test results, was cost-effective for the different scenarios simulated. However, the authors stated that the potential effectiveness would be dependent on achieving the targeted participation rate of 67% and finding the resources to meet the demand for FOBT and colonoscopy.
CRD COMMENTARY - Selection of comparators The rationale for the choice of the comparator (no screening) was clear. You should decide if this represents a valid comparator in your own setting.
Validity of estimate of measure of effectiveness The authors did not state that a systematic review of the literature had been undertaken to identify relevant research and minimise biases. They also did not report the sources searched to identify relevant studies, or the inclusion and exclusion criteria used. The authors used data from the available studies selectively. They primarily used only one RCT to gather data on effectiveness, supplementing this with parameter estimates from other trials where necessary. The participation rate was derived from the authors' assumptions. This was not explicitly justified, although it was investigated in a sensitivity analysis (but please see note under "Estimates of effectiveness and key assumptions" above). Other estimates were derived from the medical literature.
Validity of estimate of measure of benefit The estimation of benefits was modelled. The measure of life years gained reflected the impact on the patients' health, although the use of a measure that incorporated quality of life issues might have been more appropriate. The benefits were discounted at a rate of 5%.
Validity of estimate of costs Although the perspective adopted in the economic analysis was not explicitly stated, it appears that all the categories of costs relevant to a health service perspective have been included in the analysis. The unit costs were presented separately from the quantities of resources used, which will allow the study to be replicated in other contexts. The authors limited their analysis to direct costs. The resource quantities were derived from published source and expert opinion. A sensitivity analysis was not undertaken, and this may limit the interpretation of the study's findings. The unit costs were derived from published sources and opinion, and a sensitivity analysis was conducted. The authors discounted the costs at rates of 3 and 5%, although they only presented costs discounted at 5%. The date to which the prices related was not stated clearly for all cost components, which will present difficulties in terms of any future reflation exercise.
Other issues The authors made some comparisons with the findings from other studies, although they primarily focused on one other trial, which was the basis of many of the estimates of effectiveness. The authors acknowledged that the study findings may be limited to certain regions of Canada, owing to provincial variations in resource availability. The authors acknowledged the uncertainty around some of the parameters of effectiveness derived from the primary studies, such as the sensitivity and specificity of the FOBT. The authors reported a further limitation in that polyp removal was assumed to have no impact on the incidence of CRC, although recent studies showed a possible relationship between lower incidence of CRC and polyp removal. Therefore, the analysis might have underestimated the benefits of screening in this instance. The external validity of the study was limited by the lack of a sensitivity analysis on these parameters. The authors' conclusions reflected the scope of the analysis.
The authors have informed us that a broader technical report, prepared for the National Committee on Colorectal Cancer Screening, and which contains a wider literature review and additional sensitivity analyses, is available at the following url:
http://www.phac-aspc.gc.ca/publicat/ncccs-cndcc/index.html
Implications of the study The authors acknowledged that further studies may be necessary to investigate other areas of uncertainty and the potential impact of barium enema as an alternative to colonoscopy as a follow-up procedure.
Bibliographic details Flanagan W M, Le Petit C, Berthelot J-M, White K J, Coombs B A, Jones-McLean E. Potential impact of population-based colorectal cancer screening in Canada. Chronic Diseases in Canada 2003; 24(4): 81-88 Other publications of related interest 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.
Hardcastle JD, Chamberlain JO, Robinson MH, et al. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet 1996;348:1472-77.
Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for faecal occult blood. New England Journal of Medicine 1993;328:1365-71.
Indexing Status Subject indexing assigned by NLM MeSH Aged; Canada /epidemiology; Colonoscopy; Colorectal Neoplasms /diagnosis /epidemiology /mortality; Cost-Benefit Analysis; Humans; Mass Screening /economics /methods /statistics & Middle Aged; Occult Blood; numerical data AccessionNumber 22004008773 Date bibliographic record published 31/10/2005 Date abstract record published 31/10/2005 |
|
|
|