Overall Study Question
The objective of this study was to determine the cost-effectiveness of treatment with unfractionated heparin compared with enoxaparin (a low molecular weight heparin) in patients who present with unstable angina and non-Q wave myocardial infarction in Canada. The authors use the data from the Canadian patients enrolled in the ESSENCE clinical trial and supplement this with Canadian cost and resource use data. The one-year results of the ESSENCE trial show that enoxaparin is superior to unfractionated heparin in the primary combined endpoint of death/myocardial infarction and recurrent angina.
Are the Results of the Study Valid?
Did the analysis provide a full economic comparison of health care strategies?
Yes. This study provided a complete economic analysis of health care strategies. Both outcomes and costs were considered although not as transparently as in an analysis where additional money results in additional output. At first glance, it may appear to be simply a cost analysis. The research question was well defined and phrased in an answerable format. The analysis used a relevant comparison to unfractionated heparin since use of unfractionated heparin and aspirin is the gold standard comparison for unstable angina treatment. The authors specify the viewpoint as that of the healthcare payer perspective (Ontario Ministry of Health) which is fairly broad and relevant in the context of the question posed for analysis. The outcomes considered were based on the ESSENCE study protocol which was explained in detail in the article. The ESSENCE trial showed enoxaparin to be superior to unfractionated heparin. Since the strategy was dominant (more effective and less costly), an incremental cost-effectiveness (C/E) ratio was not reported.
Were the costs and outcomes properly measured and valued?
Yes. The outcomes were those in the ESSENCE trial which showed enoxaparin to be superior to unfractionated heparin (i.e. efficacy data). Effectiveness data (i.e. application of the intervention into real world practice) is not available as is common to many economic analyses. Costs identified include those from the initial hospital stay (e.g. drugs, cardiac procedures, others) and at 1-year follow-up (e.g. angiograms, admissions for revascularization, angina, MI, others). The patient outcomes for the period of 30 days to 1 year were obtained by telephone follow-up of patients and their caregivers. Not all patients had 1-year follow-ups. The investigators did not measure other outpatient related costs, however these were likely small in magnitude compared to higher cost items such as hospital-based costs. Drug costs were at market value and may be higher than are available with hospital purchasing contracts. Other hospital related costs were based on the Ontario Case Costing Project (which uses microcosting) as well as a review of hospital charts for unstable angina patients with a regression analysis. The study used 1997 Canadian dollars for analysis and appropriately did not adjust for differential timing (i.e. discounting) since outcomes upon which the analysis was based occurred over a time period of less than 1 year. Since the enoxaparin was a dominant strategy, it was not appropriate to integrate costs and effects.
Was appropriate allowance made for uncertainties in the analysis?
Yes. A sensitivity analysis was conducted for costs such as those incurred for revascularization and admission to community hospitals. The ranges tested in the sensitivity analyses were reflective of plausible ranges related to the study question. The consequences were not varied and this is a limitation of the analysis. Statistical significance was calculated since the economic analysis was based on data from a randomized controlled trial. However, the power of the economic analysis to show a statistical difference was limited by the size of the Canadian subgroup.
Are estimates of costs and outcomes related to the baseline risk in the treatment population?
No. The study did not conduct a subgroup analysis related to the patient baseline risk level. This would have been possible with potential analyses including age and presentation diagnosis (myocardial infarction vs. angina). Subgroup analyses such as these are usually conducted when there is an increased cost per effect to define the group with the best cost-effectiveness (C/E) ratio. Since the enoxaparin strategy was dominant in this analysis, this type of subgroup analysis would be less important for choosing patients who would benefit from the therapy the most, and subsequently have the lowest C/E ratio.
What are the Results?
What were the incremental costs and outcomes of each strategy?
Initial hospital-based costs were: cost for enoxaparin $101.00 per patient, cost for heparin $39.00 per patient. Revascularization costs for patients given enoxaparin $8531.00. Revascularization costs for patients given heparin $9099.00. Treatment with enoxaparin resulted in an initial cost savings in hospital of $743.00 ($10,666 vs. $9,920; p=0.14). 1-Year follow-up costs were: cost for enoxaparin patients $5,092 vs. heparin patients $5,833 (p=0.26). Total costs were: enoxaparin $15,012, heparin: $16,497. Cost savings with enoxaparin were: $1,485 (CI $-93 to $3,167; p=0.06).
Do incremental costs and outcomes differ between sub-groups?
How much does allowance for uncertainty change the results?
The range of cost savings with the use of enoxaparin was from $1,075 to $1,523. One-way sensitivity analysis based on costs of various resources did not change the results and the conclusion that enoxaparin was cost saving. In the statistical analysis, the 95% CI crossed zero indicating a lack of statistically significant difference in costs between the enoxaparin and unfractionated heparin groups. However, the C/E study was not powered to show a difference since it was not the primary outcome of the ESSENCE study. Bootstrapping was used to estimate confidence intervals.
Will the results help me in caring for my patients?
Are the treatment benefits worth the harms and costs?
Yes. The strong dominance of treatment with enoxaparin as compared with unfractionated heparin for patients with unstable angina/non-Q-wave MI indicates that enoxaparin use can save money both during the hospital stay and for 1-year of follow-up in addition to providing improved clinical efficacy. However, the present study did not consider indirect costs, quality of life, or adverse effects such as bleeding unless this event resulting in a hospital admission. Therefore, the effects and costs of these outcomes are unknown, but are not likely significant.
Could my patients expect similar health outcomes?
The patients included in the ESSENCE trial were considered high risk with a history of coronary artery disease and presentation within 24 hours of having angina. Referring to the inclusion/exclusion criteria of the ESSENCE study to make comparisons to our own unstable angina patient populations, we note that the ESSENCE patients were quite typical unstable angina patients. In order to interpret the economic analysis, the reader must examine their local clinical practice patterns and compare them to those in the ESSENCE study. The efficacy data used by the investigators included 1-year outcomes, representing a sufficient length of patient follow-up.
Could I expect similar costs?
Yes, in general. Costs are outlined in Table IV of the original paper. The authors used 1-year costs from Ontario. Costs may be different in other hospitals and in other areas of Canada, however these costs these values generally represent a more appropriate and accurate estimate of Canadian resource use and cost per unit than economic analyses conducted in other countries. Resource consumption may be different than in general practice since patients were enrolled in ESSENCE and may have had protocol-specific costs. However, these costs should balance out between the unfractionated heparin and the enoxaparin groups. A previously conducted Canadian economic analysis of enoxaparin in this setting has been published by Balen et al. These investigators considered 30-day outcomes only, and the study involved fewer patient numbers. The study yielded some similar costs (i.e. heparin $31 vs. $39, enoxaparin $82 vs. $101), while others differed significantly. Despite these disparities, the previous investigators found similar results (i.e. enoxaparin was the dominant strategy), demonstrating consistency and robustness of the results.
How can the results be put into the context of previous knowledge and clinical practice?
What are any important methodological issues with the study?
How should the study findings be applied to the field of clinical pharmacotherapy?
The results of this economic analysis are consistent with those of Balen and colleagues in Canada in addition to economic analyses conducted in France, Britain and the United States, demonstrating a consistent conclusion that the use of enoxaparin in the treatment of unstable angina/non-Q wave myocardial infarction is more effective and less costly than unfractionated heparin. (2-5) Despite the fact that the acquisition cost of enoxaparin is greater than that of unfractionated heparin, when all the costs and consequences are considered in appropriately conducted economic analyses, enoxaparin emerges as a strongly dominant option. Based on the results of this analysis, enoxaparin should become the gold standard therapy for antithrombotic therapy for patients presenting with unstable angina or non-Q wave myocardial infarction with similar characteristics as those in the ESSENCE trial.