|
The long-term cost-effectiveness of clopidogrel plus aspirin in patients undergoing percutaneous coronary intervention in Sweden |
Lindgren P, Stenestrand U, Malmberg K, Jonsson B |
|
|
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 study examined percutaneous coronary intervention (PCI) with clopidogrel plus aspirin versus aspirin alone in patients with unstable coronary artery disease (CAD).
Economic study type Cost-effectiveness analysis.
Study population The study population comprised adult patients undergoing PCI. Patients who had undergone a revascularisation procedure within 3 months or who had received treatment with glycoprotein IIb/IIIa inhibitors 3 days before randomisation were excluded, as were patients contraindicated for antithrombotic therapy and those suffering from Class IV heart failure. Further details can be found in another publication (Mehta et al. 2001, see 'Other Publications Of Related Interest' below for bibliographic details).
Setting The setting was secondary care. The economic evaluation was conducted in Sweden.
Dates to which data relate The main data for this model-based economic evaluation were retrieved from a Swedish national registry from 1 January 1995 to 1 August 2001 and the PCI CURE study (Mehta et al. 2001). The price year was 2004.
Source of effectiveness data This study mainly used two sources of data, namely a review of the literature and estimates.
Modelling A four-state Markov Model was constructed to estimate the long-term costs and effectiveness. The states within the model were after PCI (the starting state for all patients), first year after a myocardial infarction (MI), the second and subsequent years after an MI, and death. The model used yearly cycles.
Outcomes assessed in the review The outcomes assessed in the non-systematic review were the risk reduction from treatment and the utility reduction from MI.
Study designs and other criteria for inclusion in the review Sources searched to identify primary studies The authors stated that MEDLINE was searched when looking for utility data. The search was conducted from 1980 to 2003 using the search terms "utility", "quality of life" and "myocardial infarction".
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 Only one study was used as source of effectiveness data (Mehta et al. 2001).
Methods of combining primary studies Not applicable (only one study was used).
Investigation of differences between primary studies Not applicable (only one study was used).
Results of the review The risk reduction from treatment was 0.72 (95% confidence interval, CI: 0.53 to 0.96).
The utility reduction due to MI was 0.1.
Methods used to derive estimates of effectiveness The authors derived estimates of effectiveness by means of regression analyses, using data from a national registry that included patients undergoing PCI.
Estimates of effectiveness and key assumptions Coefficients for the risk functions used to calculate transition probabilities in the model were as follows.
For cardiovascular death or nonfatal MI for first year after MI: constant -6.761, age 0.057 (95% CI: 0.042 to 0.0738), diabetic 0.552 (95% CI: 0.188 to 0.915) and previous MI 0.220 (95% CI: -0.136 to 0.577).
For cardiovascular death or nonfatal MI in subsequent year after MI: constant -8.185, age 0.054 (95% CI: 0.018 to 0.090, diabetic 0.667 (95% CI: -0.026 to 1.308) and previous MI 0.621 (95% CI: 0.058 to 1.184).
For cardiovascular death: constant -6.417, age 0.081 (95% CI: 0.051 to 0.111), diabetic 0.1656 (95% CI: -0.498 to 0.830) and previous MI 0.289 (95% CI: -0.338 to 0.915).
For death due to other causes: constant -8.442, age 0.0646 (95% CI: 0.046 to 0.083), diabetic 0.462 (95% CI: 0.029 to 0.895) and previous MI 0.365 (95% CI: -0.012 to 0.741).
Measure of benefits used in the economic analysis The measure of effectiveness used was the life-years gained (LYG).
Direct costs The authors used diagnosis-related groups as proxies for resource use. These were mainly health care costs. The authors used these to calculate the initial hospitalisation cost-difference between study groups in their model. They also included the cost of MI obtained from the literature (e.g. inpatient and outpatient care, pharmaceuticals and loss of productivity (work absence), but not the patient's own expenditures). Subsequent costs were also obtained from the literature. All costs were adjusted for inflation using a consumer price index and were presented in 2004 euros. Discounting was carried out using a 3% rate, as recommended by Swedish guidelines for economic evaluations.
Statistical analysis of costs The costs were treated stochastically. For each arm of the study the authors presented the mean direct costs, as well as the indirect and total costs, with their standard deviations (SDs). Uncertainty around the costs was addressed using 1,000 bootstrap replications.
Indirect Costs The study included the indirect costs due to MI, which were obtained from the literature. For their model, the authors used an indirect cost for the first year after MI and for subsequent years. These costs were expressed in 2004 euros and were discounted at a rate of 3%, as recommended by Swedish guidelines for economic evaluations.
Currency Swedish kroner (SEK). These were converted to euros (EUR) at an exchange rate of EUR 1.00 = SEK 9.13.
Sensitivity analysis Sensitivity analyses were performed. Variability in the data was investigated. The authors conducted one-way, sub-group and probabilistic sensitivity analyses (cost-effectiveness acceptability curves). Assumptions were made, based on the literature, (e.g. costs due to increased survival), in order to define ranges for the sensitivity analysis. A deterministic sensitivity analysis was conducted on costs for the first year after an MI, costs of the second year, costs associated with added years of life, discounting, and MI occurring within 7 days of hospital admission.
Estimated benefits used in the economic analysis For the base-case analysis (whole population), the LYG were 0.04 (SD=0.05).
For the sub-group analyses, the LYG were:
with diabetes mellitus, 0.03 (SD=0.05) at age 50 years, 0.04 (SD= 0.06) at age 60 years, 0.05 (SD=0.08) at age 70 years, and 0.09 (SD=0.11) at age 80 years; and
with no diabetes mellitus, 0.03 (SD=0.03) at age 50 years, 0.04 (SD=0.04) at aged 60 years, 0.05 (SD=0.06) at age 70 years, and 0.09 (SD=0.09) at age 80 years.
Cost results The direct costs were EUR 2,277 (SD=1,478) for patients receiving aspirin and EUR 2,726 (SD=1,220) for patients receiving clopidogrel plus aspirin.
The indirect costs were EUR 523 (SD=174) for patients receiving aspirin and EUR 282 (SD=179) for patients receiving clopidogrel plus aspirin.
The total costs were EUR 2799 (SD=1,494) for patients receiving aspirin and EUR 3,132 (SD=1,253) for patients receiving clopidogrel plus aspirin.
Synthesis of costs and benefits The authors calculated the incremental cost-effectiveness ratios (ICERs). For the base-case analysis, the ICER was EUR 10,993/LYG when only the direct costs were included and EUR 8,127/LYG when both the direct and indirect costs were included.
For the sub-group analyses, the ICERs were:
with diabetes mellitus, dominance at age 50 years, EUR 1,969/LYG at age 60 years, EUR 7,213/LYG at age 70 years, and EUR 3,961/LYG at age 80 years; and
with no diabetes mellitus, EUR 7,243/LYG at age 50 years, EUR 6,929/LYG at age 60 years, EUR 7,937/LYG at age 70 years, and EUR 4,609/LYG at age 80 years.
The authors reported that the inclusion of MIs that occurred within 7 days of admission had a substantial effect on the results.
Authors' conclusions Treatment with clopidogrel plus aspirin, as in the PCI CURE study, appears to have been cost-effective in this model analysis of patients with unstable coronary artery disease (CAD) undergoing percutaneous coronary intervention (PCI) in Sweden.
CRD COMMENTARY - Selection of comparators The authors chose the comparator for their analysis on the basis of those compared in a previous study (Mehta et al. 2001). You should decide if this represents a relevant technology in your own setting.
Validity of estimate of measure of effectiveness The effectiveness was calculated by means of a Markov model, using data from a single study and estimates derived using data from a national registry. The authors did not report a systematic review of the literature. Although this is common practice with models, it does not always ensure that the best available data are used in the model. The authors could have explored other sources for the treatment risk reduction variable.
Other risks of events were derived using regression analyses, i.e. logistic regression models. The authors stated clearly how they developed these and reported the results of the analyses. However, they did not report statistics for the regression models and, as such, it is not possible to know, for instance, how well these models behaved. However, the authors appropriately investigated the input estimates using one-way and probabilistic sensitivity analyses.
Validity of estimate of measure of benefit The authors used the LYG as their effectiveness measure, but used quality-adjusted life-years (QALYs) within their sensitivity analyses. Although the authors searched MEDLINE, they found no useable estimates on the reduction in quality of life after an MI for the study population. They therefore used an assumed reduction in quality of life of 0.1, as used in a previous study, and acknowledged this as a limitation of their study.
Validity of estimate of costs The authors conducted their analysis from a societal perspective. All the cost categories (e.g. direct and indirect) relevant to this approach seem to have been included in the analysis, as were all relevant costs within each category. The costs were treated stochastically within the Markov model. The price year was reported. The authors did not report the resource quantities and the unit costs separately, which could make it difficult to rework the analysis in other settings and therefore limit the transferability of the results.
Other issues The authors compared their findings with those from other studies which, in general, showed similar results. They did not explicitly address the issue of the generalisability of their results, but they did discuss the effects of some of the data used in the model that were particularly different from those used in other studies. The authors do not appear to have presented their results selectively and their conclusions reflected the scope of the analysis.
A number of limitations were reported. For example, the cost data were old and based on a small sample. However, the authors conducted sensitivity analyses on these variables and found their results to be robust. Another limitation was the use of inclusion/exclusion criteria from a prior study that had not enrolled patients with previous PCI or coronary artery bypass graft. The authors pointed out that this could have led to an underestimate of the risk of an event in the model. A further limitation was the lack of data on QALYs since good QALY data would have enhanced the analysis.
Implications of the study The predicted cost-effectiveness ratios are well below the willingness-to-pay thresholds generally considered cost-effective.
Source of funding Funded in part by a grant from Sanofi-Synthelabo, Paris, France.
Bibliographic details Lindgren P, Stenestrand U, Malmberg K, Jonsson B. The long-term cost-effectiveness of clopidogrel plus aspirin in patients undergoing percutaneous coronary intervention in Sweden. Clinical Therapeutics 2005; 27(1): 100-110 Other publications of related interest Because readers are likely to encounter and assess individual publications, NHS EED abstracts reflect the original publication as it is written, as a stand-alone paper. Where NHS EED abstractors are able to identify positively that a publication is significantly linked to or informed by other publications, these will be referenced in the text of the abstract and their bibliographic details recorded here for information.
Mehta SR, Yusuf S, Peters RJ, et al. Effects of pretreatment with clopidogrel and aspirin followed by long term therapy in patients undergoing percutaneous coronary intervention: the PCI CURE study. Lancet 2001;358:527-33.
Indexing Status Subject indexing assigned by NLM MeSH Aged; Aged, 80 and over; Angioplasty, Balloon, Coronary; Aspirin /administration & Computer Simulation; Coronary Disease /economics /therapy; Cost-Benefit Analysis; Drug Administration Schedule; Drug Costs; Drug Therapy, Combination; Female; Humans; Logistic Models; Male; Markov Chains; Middle Aged; Models, Economic; Monte Carlo Method; Platelet Aggregation Inhibitors /administration & Ticlopidine /administration & Time Factors; derivatives /economics /therapeutic use; dosage /analogs & dosage /economics /therapeutic use; dosage /economics /therapeutic use AccessionNumber 22005008121 Date bibliographic record published 30/04/2007 Date abstract record published 30/04/2007 |
|
|
|