|
Comparison of selection strategies for genetic testing of patients with hereditary nonpolyposis colorectal carcinoma: effectiveness and cost-effectiveness |
Reyes C M, Allen B A, Terdiman J P, Wilson L S |
|
|
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 Four strategies for the genetic testing of patients (gene carriers) with hereditary nonpolyposis colorectal carcinoma (HNPCRC) were examined.
Amsterdam strategy: germline MSH2 and MLH1 mutational analysis by direct sequencing of only very high-risk individuals who meet Amsterdam clinical criteria.
Modified strategy: tumour microsatellite instability (MSI) analysis of individuals who meet less stringent modified clinical criteria and germline mutational analysis of individuals with a high-frequency MSI (MSI-H) tumour.
Test-all strategy: tumour MSI analysis of all CRC individuals and germline mutational analysis of individuals with an MSI-H tumour.
Mixed strategy: germline mutational analysis of individuals who satisfy the Amsterdam criteria and tumour MSI analysis of the remainder of individuals who meet the less stringent modified criteria, with subsequent germline MSH2/MLH1 testing of those with an MSI-H tumour.
The Amsterdam criteria were defined as three family members with CRC, one of whom should be the first-degree relative of the other two, two successive generations with cancer, one cancer diagnosed before 50 years of age, and exclusion of polyposis syndromes. The modified testing guidelines were any individual with CRC with a personal or family history of another CRC or endometrial carcinoma, or who was diagnosed before 50 years of age, regardless of personal or family history.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised hypothetical individuals diagnosed with CRC who were at risk of HNPCRC. The definition of eligibility depended on the testing strategy adopted.
Setting The setting was primary care. Data for the model parameters were derived from US sources.
Dates to which data relate The effectiveness data were derived from studies published between 1997 and 2001. No dates relating to resource use were reported. The price year was 2001.
Source of effectiveness data The effectiveness evidence came from published studies, augmented by the authors' assumptions.
Modelling A decision tree model was constructed to evaluate the cost-effectiveness of the four testing strategies. The tree started with a hypothetical population of 1,000 patients diagnosed with CRC, who agreed to genetic testing, and their 6,000 first-degree relatives. The scheme of the tree was reported in the paper.
Outcomes assessed in the review The model parameters derived in the review were the specificity and sensitivity of the sequencing test, and the probabilities of the following:
satisfying the Amsterdam criteria,
sequencing the mutation given the Amsterdam criteria,
satisfying the modified guidelines,
MSI given satisfaction of the modified guidelines,
sequencing the mutation given MSI given modified guidelines,
MSI in all CRC patients,
sequencing the mutation given MSI in all CRC patients,
the first-degree relatives undergoing the test, and
a first-degree relative being a gene carrier.
Study designs and other criteria for inclusion in the review 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 The effectiveness evidence was obtained from 10 primary studies.
Methods of combining primary studies The review was not systematic and the model estimates were selective.
Investigation of differences between primary studies Results of the review The average probability values were:
0.01 (range: 0.003 - 0.15) for satisfying the Amsterdam criteria,
0.44 (range: 0.25 - 0.63) for sequencing the mutation given the Amsterdam criteria,
0.218 for satisfying the modified guidelines,
0.29 (range: 0.07 - 0.38) for MSI given satisfaction of the modified guidelines,
0.5 (range: 0.30 - 0.63) for sequencing the mutation given MSI given satisfaction of the modified guidelines,
0.12 (range: 0.02 - 0.42) for MSI in all CRC patients,
0.27 (range: 0.07 - 0.47) for sequencing the mutation given MSI in all CRC patients,
0.43 (range: 0 - 0.8) for first-degree relatives undergoing the test, and
0.42 for a first-degree relative being a gene carrier.
The specificity the sequencing test was 0.725 (range: 0.50 - 0.95) and the specificity was 0.90 (alternative value: 0.99).
Methods used to derive estimates of effectiveness The authors made some assumptions used in the decision tree.
Estimates of effectiveness and key assumptions The authors assumed an average of 6 first-degree relatives for a given proband, a 2% MSH2 and MLH1 prevalence rate, a 100% proband and first-degree relative testing, and a 50% MSH2 and MLH1 prevalence rate in first-degree relatives.
Measure of benefits used in the economic analysis The benefit measure used in the economic analysis was the number of gene carriers detected per 1,000 CRC patients. It was derived from the decision model. No discount rate was used because the time between the test and gene detection was very short.
Direct costs Discounting was not performed since the time horizon of the study was less than one year. The unit costs were reported, but the quantities of resources used were not. The direct costs included were for genetic counselling, initial screen, MSI testing, initial proband sequencing (MSH2, MLH1) and family member sequencing. The unit costs were estimated from both local and national sources, such as the resource-based relative value system for physicians and the Medicare fee schedule rate. MSI testing costs were derived from a published study. In some cases, charges were used instead of the actual costs. Data on resource consumption was mainly derived from published studies and the authors' assumptions (on the family size). All of the costs were inflated to 2001 prices using the medical care component of the Consumer Price Index-Urban.
Statistical analysis of costs The costs were treated deterministically.
Indirect Costs The indirect costs were not included.
Sensitivity analysis One-way sensitivity analyses were conducted on all model inputs to assess the robustness of the estimated cost-effectiveness ratios. The ranges used were derived from the literature for effectiveness values and from actual data for the costs.
Estimated benefits used in the economic analysis The number of gene carriers detected per 1,000 CRC patients was 7.8 with the Amsterdam strategy, 54.3 with the modified strategy, 59.6 with the mixed strategy, and 67.2 with the test-all strategy.
Cost results The total costs per 1,000 CRC patients were $277,000 with the Amsterdam strategy, $595,000 with the modified strategy, $611,000 with the mixed strategy, and $998,000 with the test-all strategy.
Synthesis of costs and benefits An incremental cost-effectiveness analysis was conducted to combine the costs and the benefits of the four strategies under study. The incremental cost per additional gene carrier detected was $6,832 with the modified strategy over the Amsterdam strategy, $3,007 with the mixed strategy over the modified strategy, and $51,151 with the test-all strategy over the modified strategy.
The results of the analysis were generally robust to the variations investigated in the sensitivity analyses. The probability of fulfilling the Amsterdam criteria was the variable with the greatest impact on the results, due to the wide range used in the sensitivity analysis. When this probability was higher than 2.5% (quite close to the base-case value), then the mixed strategy dominated the test-all strategy. Increasing the probability that family members would be tested if the proband had a germline mutation improved the cost-effectiveness of all testing strategies.
Authors' conclusions A very stringent strategy (Amsterdam strategy) was not effective since many gene carriers were missed, while a test-all strategy was too expensive. The mixed strategy showed the lowest incremental cost-effectiveness ratio. Also, it dominated the remaining strategies across a range of plausible variations in key factors.
CRD COMMENTARY - Selection of comparators The authors stated that the four strategies compared in the analysis represented those generally employed in the USA for genetic testing of HNPCRC. You should decide whether they are valid comparators in your own setting.
Validity of estimate of measure of effectiveness The effectiveness analysis used data derived from published studies, but a formal review of the literature was not undertaken. Consequently, the authors did not report the search methods or the study designs of the primary studies. The model parameter estimates were selective. The consequence of this was a high degree of uncertainty in the model, which was investigated through sensitivity analyses.
Validity of estimate of measure of benefit The number of gene carriers detected per 1,000 CRC patients was used as the benefit measure in the economic analysis. Such a measure appears to be the natural outcome measure of the screening test.
Validity of estimate of costs The authors stated that they adopted a societal perspective, but they did not include any measure of productivity loss. The sources of the unit costs and quantities were numerous. Published average unit cost and cost estimates were used. In some cases, charges rather than the actual costs were used, which may bias the results. The price year was stated and the unit costs were reported, but the quantities were not provided. Hence, the study results could not be reproduced in the future. The costs and the quantities were treated deterministically in the base-case, but the uncertainty around the cost results was investigated through a sensitivity analysis.
Other issues The authors compared their findings with those of a published study that provided information on the Bethesda guidelines for genetic testing. The authors did not address the issue of the generalisability of the study results, but performed several sensitivity analyses on key variables. The results of the base-case and sensitivity analyses were presented in detail. An incremental cost-effectiveness analysis was appropriately conducted. The results were reported in full and the conclusions reflected the scope of the analysis. The time horizon (1 year) may be considered short.
Implications of the study The study suggests that the modified or mixed strategies should be used for genetic testing of patients with HNPCRC. The authors highlighted the importance of patient compliance with testing, and the need to examine whether the "test-all" strategy would become more viable if the downstream costs and survival were considered.
Source of funding Supported by a grant from the Theodora Betz Foundation.
Bibliographic details Reyes C M, Allen B A, Terdiman J P, Wilson L S. Comparison of selection strategies for genetic testing of patients with hereditary nonpolyposis colorectal carcinoma: effectiveness and cost-effectiveness. Cancer 2002; 95(9): 1848-1856 Indexing Status Subject indexing assigned by NLM MeSH Adaptor Proteins, Signal Transducing; Base Pair Mismatch; Carrier Proteins; Colorectal Neoplasms, Hereditary Nonpolyposis /diagnosis /economics /genetics; Cost-Benefit Analysis; Costs and Cost Analysis; DNA Repair /genetics; DNA, Neoplasm /genetics; DNA-Binding Proteins; Decision Support Techniques; Genetic Testing /economics /methods; Germ-Line Mutation; Heterozygote Detection; Humans; Microsatellite Repeats /genetics; MutS Homolog 2 Protein; Neoplasm Proteins /genetics; Nuclear Proteins; Proto-Oncogene Proteins /genetics AccessionNumber 22002001891 Date bibliographic record published 31/10/2003 Date abstract record published 31/10/2003 |
|
|
|