|
Cost-effectiveness of hepatic metastasectomy in patients with metastatic colorectal carcinoma: a state-transition Monte Carlo decision analysis |
Gazelle G S, Hunink M G, Kuntz K M, McMahon P M, Halpern E F, Beinfeld M, Lester J S, Tanabe K K, Weinstein M C |
|
|
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 use of hepatic metastasectomy in patients with metachronous liver metastases from colorectal carcinoma (CRC) was under evaluation.
Study population The study population comprised hypothetical cohorts of 10,000 patients of 65-year-old men who had undergone resection of a primary CRC and were known to have developed metachronous liver metastases.
Setting The setting was secondary care. The economic study was carried out in Boston (MA), USA.
Dates to which data relate The effectiveness data were gathered from studies published between 1978 and 1998. The resource data were gathered from studies and other sources published between 1991 and 1998. The price year was 1998.
Source of effectiveness data The effectiveness data were derived from the literature where valid data were available, or otherwise from expert opinion.
Modelling A state-transition Monte Carlo model was created to simulate the costs and effectiveness assigned to each strategy. The three health states described were alive_res, alive_nores and dead. Patients in the alive_res state were potential candidates for resection. Patients moved to the alive_res state when they were either found to be unresectable (i.e. to have more metastases than the threshold for metastasectomy in the strategy under consideration) or when they had had the maximum allowable number of resections for the strategy under consideration.
The model included only one generic imaging and treatment strategy. This was defined using specific model parameters (operative threshold, image/treat interval, test sensitivity). The model tracked up to 15 individual hepatic metastases in each patient. The cycle length was 1 month and a 5-year time horizon was used. Several assumptions were made. One assumption was that helical, contrast-enhanced computed tomography (CT) scanning was used in all patients. Another assumptions was that each segment in which at least one metastasis was detected was resected, and that segments in which no metastases were detected were not resected.
Outcomes assessed in the review The outcomes assessed in the review and used as model inputs were:
the patients' age,
the tumour volume doubling time,
the normal liver volume,
the mortality and morbidity associated with hepatic metastasectomy,
the intraoperative ultrasound sensitivity,
the size threshold for tumour detection,
the median survival, and
the mortality hazard rates.
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 Two primary studies were used to assess the average distribution of primary and revision surgeries for middle ear or mastoid procedures. About 26 primary studies were used to assess the different outcomes.
Methods of combining primary studies Investigation of differences between primary studies Results of the review The base-case estimates were reported, with the values used in the sensitivity analysis given in brackets.
The patients' age was 65 years (55).
The tumour volume doubling time was 155 days (78 - 233).
The normal liver volume was 1,225 cm3.
The probability of mortality associated with hepatic metastasectomy was 5% (10%).
The probability of morbidity associated with hepatic metastasectomy was 20%.
The sensitivity of intraoperative ultrasound was 0.95.
The size threshold for tumour detection was 0.5 cm (1 cm) with CT and 0.3 cm (0.5 cm) with intraoperative ultrasound.
The median survival was 11.5 months for patients with less than 25% of liver replacement by tumour (LVRT) and 6.3 months for patients with 25% or more LVRT.
The mortality hazard rate for patients without liver metastases was twice as high as the age- and gender-adjusted rates for the US population.
The mortality hazard rate was 1.3203 for patients with less than 25% of LVRT and 0.7427 for patients with 25% or more LVRT.
Methods used to derive estimates of effectiveness The number of metastases per patient was calculated using a Poisson distribution with a mean of 6 metastases per patient. The size of the metastases was calculated using a gamma distribution with a mean tumour diameter of 2.5 cm. Expert opinion was used to estimate the mortality and morbidity associated with laparotomy. Authors' opinion was used to estimate the sensitivity of CT.
Estimates of effectiveness and key assumptions The probabilities of mortality and morbidity associated with laparotomy were 4% (mortality) and 1% (morbidity), respectively.
The sensitivity of CT was 0.80 (0.75 - 0.85).
Each metastasis was assumed to have an independent, equal probability of being located in each of the eight liver segments.
Growth in tumour volume was assumed to be exponential, at rates determined from the literature.
It was assumed that no new liver metastases developed over time, as all patients were assumed to be of the status post removal of their primary tumours.
It was assumed that, below a certain size threshold, all metastases were missed by a diagnostic imaging test.
Measure of benefits used in the economic analysis The benefit measure used was the number of quality-adjusted life-years (QALYs) saved. The QALYs were obtained by multiplying median survival by quality of life (QOL) weights, which were derived from the literature and authors' opinion. The benefits were discounted at a rate of 3%.
Direct costs The direct costs of the hospital and patient and family were considered in the analysis. The hospital costs included the costs associated with diagnostic, liver resection, patient care, complications and physician services. The costs of transportation, care by family and friends, or household modifications related to either diagnostic testing or treatment were excluded, as there were no significant differences between the strategies.
The unit costs were reported. The unit costs for the diagnostic tests, liver resection and professional costs were determined on the basis of the Medicare payment schedule. Two assumptions were made. In one, all preoperative diagnostic CT would be performed on an outpatient basis and would therefore use the Medicare reimbursement rate. In the other, intraoperative ultrasound would be used in all patients undergoing metastasectomy. The costs of complications were collected from a multi-institutional study on hepatic resection, while the cost estimates were derived using software (Eclypsis). The costs associated with dying were calculated by summing the cost of 5 intensive care unit days and the physician costs related to care for 5 days. The cost of patient care with local and metastatic colorectal cancer was estimated from the literature. The daily wage rate was derived from the US Bureau of Labor Statistics and was used as a proxy for the time costs of diagnostic tests, therapeutic interventions and complications. All the costs were updated to 1998 US dollars using the Consumer Price Index. The costs were discounted at a rate of 3%.
Statistical analysis of costs No statistical analysis of the costs was performed.
Indirect Costs The indirect costs were not included in the analysis.
Sensitivity analysis One-way sensitivity analyses were performed across a range of assumptions. The assumptions concerned the number of metastases per patient, the patients' age, the tumour volume doubling time, the discount rate, mortality hazard rates, the QOL weights, the costs of surgery and patient care, the discount rate, and the thresholds for tumour detection at imaging and surgery. Additional analyses were performed in which several parameters were varied simultaneously.
Estimated benefits used in the economic analysis The authors reported the estimated benefits only for the non dominated strategies (i.e. 4 strategies and the reference strategy).
The average number of QALYs per patient with the reference strategy was 0.7688.
The average number of QALYs per patient with the strategy involving a 6-month follow-up interval, the resection of no more than 6 metastases, and a maximum of 1 resection allowed was 3.0968.
The average number of QALYs per patient with the strategy involving a 12-month follow-up interval, the resection of no more than 6 metastases, and up to 1 repeat metastasectomy was 3.2550.
The average number of QALYs per patient with the strategy involving a 6-month follow-up interval, the resection of no more than 6 metastases, and up to 1 repeat metastasectomy was 3.3321.
The average number of QALYs per patient with the strategy involving a 4-month follow-up interval, the resection of no more than 6 metastases, and up to 1 repeat metastasectomy was 3.3591.
Cost results The average cost per patient with the reference strategy was $6,000.
The average cost per patient with the strategy involving a 6-month follow-up interval, the resection of no more than 6 metastases, and a maximum of 1 resection allowed was $47,000.
The average cost per patient with the strategy involving a 12-month follow-up interval, the resection of no more than 6 metastases, and up to 1 repeat metastasectomy was $50,700.
The average cost per patient with the strategy involving a 6-month follow-up interval, the resection of no more than 6 metastases, and up to 1 repeat metastasectomy was $53,200.
The average cost per patient with the strategy involving a 4-month follow-up interval, the resection of no more than 6 metastases, and up to 1 repeat metastasectomy was $54,900.
Synthesis of costs and benefits The incremental cost-effectiveness ratio (ICER) of the strategy involving a 6-month follow-up interval, the resection of no more than 6 metastases, and a maximum of 1 resection allowed was $17,600 per QALY saved compared with the reference strategy.
The ICER of the strategy involving a 12-month follow-up interval, the resection of no more than 6 metastases, and up to 1 repeat metastasectomy was $23,600 per QALY saved compared with the previous strategy.
The ICER of the strategy involving a 6-month follow-up interval, the resection of no more than 6 metastases, and up to 1 repeat metastasectomy was $31,700 per QALY saved compared with the previous strategy.
The ICER of the strategy involving a 4-month follow-up interval, the resection of no more than 6 metastases, and up to 1 repeat metastasectomy was $62,700 per QALY saved compared with the previous strategy.
Across a range of model assumptions, more aggressive treatment strategies (i.e. the resection of more metastases and the resection of recurrent metastases) were superior to less aggressive strategies and had ICERs of below $35,000 per QALY saved.
The findings were insensitive to changes in most model parameters, but were somewhat sensitive to changes in surgery and treatment costs.
Authors' conclusions Hepatic metastasectomy is a cost-effective option for selected patients with metachronous colorectal carcinoma (CRC) metastases limited to the liver. More aggressive treatment strategies (i.e. the resection of more metastases and recurrent metastases) were superior to less aggressive strategies.
CRD COMMENTARY - Selection of comparators The reason for the choice of the comparator (no surgery) appears to have been clear. Different intervals for testing and treatment were investigated. You should decide whether the comparator and alternative strategies are relevant to your own setting.
Validity of estimate of measure of effectiveness The authors did not report that a systematic review of the literature was conducted for any of the parameters. The sources searched to identify the primary study, the study designs, other criteria for inclusion in the review, and the validity of the studies were not reported. The authors acknowledged that there was certainly heterogeneity across patients for all model parameters. However, they did not attempt to assess the potential impact of population heterogeneity on the results because of a lack of data on which to base the range estimates. A good feature of the effectiveness analysis was that most of the estimates were varied in the sensitivity analyses. Hence, the robustness of the results to parameter uncertainties was examined. In addition, details of the sensitivity analyses were precisely reported. This enhances understanding for the reader and allows the results to be transferred to other settings.
Validity of estimate of measure of benefit The estimation of benefits was modelled. The decision tree appears to have been relevant for the study question. The model was verified, showing consistent results with the literature. The authors used published studies assessing QOL weights (when available) and authors' opinion to complete the data. All the estimates of QOL and life expectancy were tested via sensitivity analyses. This was appropriate for this study and made it more broadly comparable to other treatments.
Validity of estimate of costs The authors reported that they adopted a societal perspective, although they did not include the indirect costs. The exclusion of such costs might have biased the results in favour of the treatment strategies (in underestimating the wider social costs of treating patients with metachronous liver metastases). Some relevant costs were omitted from the analysis because they were common to both treatment strategies (e.g. cost of transportation, care by family and friends, and household modification). The unit costs were reported and a sensitivity analysis was performed on the costs. This enhances the transferability of the economic analysis to other settings. The ranges of variation used in the sensitivity analyses and the subsequent results were reported in detail. Since all the costs were incurred during 5 years, discounting was relevant and was performed. The date to which the prices related was reported, which will allow possible reflation exercises.
Other issues The generalisability of the results was partially addressed through the population heterogeneity. Relevant sensitivity analyses on the outcome and cost estimates were carried out. Therefore, the external validity of the analysis is probably high. The authors did not compare their findings with those from other studies. The authors reported a number of further limitations to their study. Such limitations pertained to parameter uncertainty, population heterogeneity and the simplification of reality through the model.
Implications of the study The authors recommended that surgeons should be encouraged to consider resection in all patients whose metastases can technically be removed. The authors suggested that additional data collection and analysis in population heterogeneity might be of use in the future.
Source of funding Sponsored in part by the National Cancer Institute (grant R01-CA/HS83960) and the US Department of the Army (grant DAMD 17-99-2-9001).
Bibliographic details Gazelle G S, Hunink M G, Kuntz K M, McMahon P M, Halpern E F, Beinfeld M, Lester J S, Tanabe K K, Weinstein M C. Cost-effectiveness of hepatic metastasectomy in patients with metastatic colorectal carcinoma: a state-transition Monte Carlo decision analysis. Annals of Surgery 2003; 237(4): 544-555 Indexing Status Subject indexing assigned by NLM MeSH Aged; Colorectal Neoplasms /pathology; Cost-Benefit Analysis; Decision Support Techniques; Hepatectomy /economics; Humans; Liver Neoplasms /secondary /surgery; Male; Models, Statistical; Sensitivity and Specificity AccessionNumber 22003001367 Date bibliographic record published 28/02/2005 Date abstract record published 28/02/2005 |
|
|
|