|
Cost-effectiveness analysis of two strategies for mass screening for colorectal cancer in France |
Berchi C, Bouvier V, Reaud J-M, Launoy G |
|
|
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 Two approaches for 20 years of biennial colorectal cancer (CRC) screening were studied. The approaches were an automated immunological test (Magstream) and the guaiac stool tests (Haemoccult). Individuals with positive results underwent a colonoscopic investigation.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised a hypothetical cohort of individuals aged 50 to 74 years.
Setting The setting appears to have been primary care. The economic study was carried out in France.
Dates to which data relate The effectiveness evidence came from studies published between 1982 and 2001. No dates for resource usage were explicitly reported. The price year was not reported.
Source of effectiveness data The effectiveness evidence was derived from a synthesis of completed studies.
Modelling A Markov model was constructed to determine the costs and benefits of CRC screening in a hypothetical cohort of 165,000 individuals aged 50 to 74 years, who were undergoing biennial CRC screening for 20 years. The model included six health states, which were defined according to the status of the individuals in relation to screening (refusal, positive or negative). The health states were no cancer or adenoma, adenoma less than 1 cm, adenoma more than 1 cm, CRC stage A, B, C (according to Dukes classification) and metastasised, follow-up, and death. The cycle length was one year.
Outcomes assessed in the review The outcomes estimated from the literature were:
the prevalence of adenomas in relation to age;
the annual probability of transition of an adenoma of less than 1 cm into one of more than 1 cm;
the annual probability of transition of an adenoma of more than 1 cm into cancer;
the frequency of CRC in screened individuals in relation to age;
the frequency of CRC in patients refusing a test in relation to age;
the occurrence and distribution of CRCs per diagnostic stage;
the rate of specific mortality of CRC at 1 to 10 years per diagnostic stage;
the sensitivity and specificity of the two screening tests; and
the rate of participation to screening.
Study designs and other criteria for inclusion in the review It was not stated whether a systematic review of the literature was undertaken. The design of the primary studies was unclear, although the number of participants was given for some studies. Mortality was derived from French life tables, while the occurrence or distribution of CRC was derived from a French registry (screening programme run in Calvados from 1991 to 1994).
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 Eight primary studies provided the evidence.
Methods of combining primary studies The method used to combine the primary studies was not reported. However, it appears that each estimate has been derived from the most reliable study, while extreme values have been used as ranges in the sensitivity analysis.
Investigation of differences between primary studies Results of the review The prevalence of adenomas in relation to age was 21 to 53% (range: 26.9 - 58.7).
The annual probability of the transition of an adenoma of less than 1 cm into one of more than 1 cm was 0.02 (range: 0.01 - 0.04).
The annual probability of the transition of an adenoma of more than 1 cm into cancer was 0.0085 (range: 0.00425 - 0.017).
The frequency of CRC in screened individuals in relation to age was 42.1 to 288 per 100,000.
The frequency of CRC in patients refusing a test in relation to age was 52 to 590 per 100,000.
The sensitivity of the Haemoccult test was 52% and the specificity was 99.5%.
The sensitivity of Magstream was 82% (range: 70 - 90) and the specificity was 96% (range: 90 - 100).
The rate of participation in screening was 43.7%.
The occurrence and distribution of CRCs per diagnostic stage, and the rate of specific mortality of CRC at 1 to 10 years per diagnostic stage, were not reported.
Measure of benefits used in the economic analysis The summary benefit measure was the number of life-years saved with each screening strategy. This was obtained using modelling. It would appear that no discounting was applied.
Direct costs Discounting was relevant since the costs were incurred during a 20-year timeframe. An annual discount rate of 5% was applied. The unit costs were clearly presented, but the information on resource use was limited. The health services included in the economic evaluation were organisation of mass screening campaign, tests (purchasing, distribution, and revelation), colonoscopy, and cancer treatment (dependent on disease stage). A detailed breakdown of the cost items was provided. The costs of diagnosing cancers in individuals with negative tests were also considered. The costs of follow-up consisted of one colonoscopy performed every 3 years, as recommended by French gastroenterologists. The cost/resource boundary of the Social Security Service was adopted. The total costs were estimated using modelling. The costs were estimated mainly using reimbursement rates derived from the Calvados screening campaign. The source of the resource use data was unclear, although some quantities of services were based on local recommended treatment patterns. The price year was not explicitly reported.
Statistical analysis of costs The costs were treated deterministically.
Indirect Costs The indirect costs were not considered.
Sensitivity analysis Univariate sensitivity analyses were carried out to assess the impact of variations in model inputs on the estimated cost-effectiveness ratios. Variations in the participation rate, costs of tests and colonoscopy, sensitivity and specificity of Magstream, prevalence of disease, and the annual transition rates were explored. In general, the ranges used were derived from the literature.
Estimated benefits used in the economic analysis The estimated number of life-years saved for a 20-year screening programme was 16.7201 with Magstream and 16.7003 with Haemoccult. The corresponding figures for a 10-year programme were 9.7960 (Magstream) and 9.7901 (Haemoccult), respectively.
Cost results For a 20-year screening programme, the estimated discounted (undiscounted) cost of screening per targeted person was Euros 238 (Euro 316) with Magstream and Euro 179 (Euro 234) with Haemoccult. The corresponding figures for a 10-year screening programme were Euro 195 (Euro 230) and Euro 151 (Euro 177), respectively.
The greatest cost component was colonoscopic investigation (63% for Magstream and 37% with Haemoccult), followed by screening tests (20% for Magstream and 33% with Haemoccult).
Synthesis of costs and benefits An incremental cost-effectiveness ratio (ICER) was calculated to combine the costs and benefits of the screening strategies. The incremental cost per life-year saved with Magstream over Haemoccult was Euro 2,980 for a 20-year programme after the costs were discounted, and Euro 7,458 when the costs were not discounted. The corresponding figures for a 10-year programme were Euro 4,141 (discounted) and Euro 8,983 (undiscounted).
The sensitivity analysis showed that the ICER was positively correlated with the participation rate (e.g. a decrease in participation from 43.7 to 20% led to a 50% decrease in the ICER). Similarly, the ICER was positively correlated with the cost of colonoscopy. On the other hand, the ICER was negatively correlated with the cost of cancer treatment. When the cost of the Haemoccult test equalled the cost of Magstream, the ICER was Euro 4,898 for 20 years of screening (18% increase compared with the basic scenario).
For a given specificity, the ICER was negatively correlated with the sensitivity of Magstream, and for a given sensitivity the ICER was negatively correlated with the specificity of Magstream. In particular, with 70% sensitivity, the ICER was Euro 26,107 for 90% specificity and Euro -3,607 with a 100% specificity. Therefore, under particular scenarios, Magstream dominated Haemoccult.
Finally, the cost-effectiveness ratio was very sensitive to the parameters associated with the natural history of CRC.
Authors' conclusions The substitution of the Haemoccult test with Magstream in mass screening for colorectal cancer (CRC) proved to be a cost-effective strategy from the perspective of the third-party payer in France. However, the results were sensitive to the hypotheses underlying the model used in the analysis.
CRD COMMENTARY - Selection of comparators The authors provided a justification for the choice of the comparators. Haemoccult (with or without rehydration) represented the most widely used screening strategy, while Magstream was a newer immunological approach. You should decide whether they are valid comparators in your own setting.
Validity of estimate of measure of effectiveness The analysis of effectiveness used data obtained from completed studies. It was unclear whether a systematic review of the literature was carried out, and limited information on the primary studies was provided. Therefore, it is not possible to assess the validity of the sources used. Some of the evidence came from local registries. The most reliable estimate, among those available in the literature, was selected in the base-case, while estimates from other studies provided the basis for the ranges of values tested in the sensitivity analysis.
Validity of estimate of measure of benefit The summary benefit measure was appropriate to determine the impact of the interventions on the patients' health. In addition, it represents a measure widely used in studies evaluating cancer screening programmes and it is comparable with the benefits of other health care interventions. However, the impact of the screening on quality of life, which would have been interesting, was not assessed. No discounting was applied. The use of a discount rate on survival in the sensitivity analysis would have been helpful.
Validity of estimate of costs The authors explicitly stated the perspective adopted in the study. As such, it appears that all the relevant categories of costs have been included in the analysis. The unit costs were provided, but the information on resource use was less clear and appears to have been based on local treatment patterns. The price year was not reported, which makes reflation exercises in other settings difficult. The costs were treated deterministically in the base-case, but were then varied in the sensitivity analysis. Discounting was applied but undiscounted results were also reported. The authors noted that charges rather than true costs were used in the analysis, therefore the real costs could have been underestimated.
Other issues The authors did not compare their findings with those from other studies. It was stated that most of the data were derived from French sources, which reduces the possibility of transferring the conclusions of the analysis to other settings. However, extensive sensitivity analyses were conducted within reasonable ranges, thus increasing the external validity of the analysis.
Implications of the study The authors stated that forthcoming results of French population-based experiments with immunological tests would confirm (or not) their hypothesis concerning the quality of immunological testing. If further supporting evidence on Magstream becomes available, French health authorities would have a satisfactorily alternative to guaiac tests.
Source of funding Supported by 'la Ligue Contre le Cancer' of 'la Manche'.
Bibliographic details Berchi C, Bouvier V, Reaud J-M, Launoy G. Cost-effectiveness analysis of two strategies for mass screening for colorectal cancer in France. Health Economics 2004; 13: 227-238 Other publications of related interest Khandker RK, Dulski JD, Kilpatrick JB, Ellis RP, Mitchell JB, Baine WB. A decision model and cost-effectiveness analysis of colorectal cancer screening and surveillance guidelines for average-risk adults. International Journal of Technology Assessment in Health Care 2000;16:799-810.
Bouvier V, Herbert C, Launoy G. Stage of extension and treatement for colorectal cancer after a negative test and among non-responders in mass screening with guaiac faecal occult blood test: a French experience. European Journal of Cancer Prevention 2001;10:323-6.
Frazier AL, Colditz GA, Fuchs CS, Kuntz KM. Cost-effectiveness of screening for colorectal cancer in the general population. JAMA 2000;284:1954-61.
Indexing Status Subject indexing assigned by NLM MeSH Aged; Colorectal Neoplasms /diagnosis; Cost-Benefit Analysis; Female; France; Health Services Research; Humans; Male; Mass Screening /economics /methods; Middle Aged AccessionNumber 22004008106 Date bibliographic record published 28/02/2005 Date abstract record published 28/02/2005 |
|
|
|