|Incremental cost-effectiveness analysis comparing rofecoxib with nonselective NSAIDs in osteoarthritis: Ontario Ministry of Health perspective
|Marshall J K, Pellissier J M, Attard C L, Kong S X, Marentette M A
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.
The health intervention examined in the study was rofecoxib (24.7 mg daily) for the treatment of patients with osteoarthritis (OA).
Economic study type
The study population comprised a hypothetical group of patients with OA, aged over 65 years who did not experience adequate pain relief with at least 4 g/day of paracetamol (acetaminophen).
The setting was not explicitly stated. The economic analysis was carried out in Ontario, Canada.
Dates to which data relate
The effectiveness evidence was largely based on a study published in 1999. Data on resource use and costing were derived from studies published over a number of years. The price year was 1999.
Source of effectiveness data
The effectiveness evidence came from a review of the literature and authors' assumptions.
A decision tree model was constructed to compare the costs and benefits of the two alternative treatments reflecting practice in the Canadian setting. The model referred to a hypothetical cohort of patients with OA who were treated with either rofecoxib or NSAIDs over a period of one year. Patients could develop GI adverse events, which were considered to be minor if managed empirically, or major if investigated. The two main branches of the model (rofecoxib or NSAIDs) were similar.
Outcomes assessed in the review
The model parameters estimated in the review were as follows:
the rates of NSAID-related GI adverse events;
perforations, ulcers and bleeds (PUB) given GI adverse events;
investigated PUB given GI adverse events;
treatment given minor GI adverse events; and
cumulative annual incidence of PUB.
The percentage of nonselective NSAIDs prescribed with a gastrointestinal agent (GPA) was also estimated.
Study designs and other criteria for inclusion in the review
All studies included in the review were double-blind, randomised, controlled trials and involved a total of 5,435 patients.
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 were included in the review.
Methods of combining primary studies
Investigation of differences between primary studies
Results of the review
The results of the review were as follow:
The rate of NSAID-related GI adverse events was 0.299 with rofecoxib and 0.295 with non-selective NSAIDs.
The rate of PUB, given GI adverse events, was 0.050 with rofecoxib and 0.091 with non-selective NSAIDs.
The rate of investigated PUB, given GI adverse events, was 0.028 with rofecoxib and 0.072 with non-selective NSAIDs.
The probability of treatment, given minor GI adverse events, was 0.241 with rofecoxib and 0.370 with non-selective NSAIDs.
The cumulative annual incidence of PUB was 0.0149 with rofecoxib and 0.0268 with non-selective NSAIDs.
The mid-point of the literature-based estimates of GPA use was 23%.
Methods used to derive estimates of effectiveness
The authors made some assumptions used in the decision model.
Estimates of effectiveness and key assumptions
The main assumption used in the model was that the two treatments (rofecoxib and NSAIDs) had the same efficacy, meaning that they had equal analgesic and anti-inflammatory effects. This implied that the costs and consequences of treating OA were assumed to occur equally with both treatments and were not considered in the economic analysis.
Measure of benefits used in the economic analysis
The summary benefit measure used in the economic analysis was the number of PUB averted per patient year, which was produced by the model. Benefits were not discounted, as the time horizon of the model was one-year.
The cost/resource boundary adopted in the analysis was that of the Ontario Ministry of Health. The health service costs included in the economic analysis were those for daily study medications, prophylactic gastroprotective agents (GPAs), hospital treatment for PUB (surgical inpatient treatment and physician fees), and outpatient treatment for PUB (outpatient clinic visit, day surgery, blood count, routine chemistry, radiology, endoscope with biopsy, repeat endoscopy, clinic physician, gastroenterologist, and pathologist). Discounting was not applied, as costs were incurred over a period of one year. Unit costs were reported separately but the quantities of resources used were not. The estimation of unit costs was based on actual data derived from the Ontario Schedule of Benefits, Ontario Case Costing Project, Ontario Public Claims Data, and Merck Frosst Canada. Resource use was calculated on the basis of data retrieved from the clinical trials used to derive the effectiveness evidence, from assumptions made by the authors, and by experts who were contacted to estimate treatment patterns in the Canadian setting. Costs were updated to 1999 values using the health and personal care component of the Consumer Price Index.
Statistical analysis of costs
Costs were treated deterministically in the base case.
Indirect costs were not included in the economic evaluation.
One-way and two-way sensitivity analyses were conducted to evaluate the robustness of the estimated cost-effectiveness ratios to variations in the following parameters: the rate of GPA co-prescription in both strategies; the relative baseline risk of PUB in the study population; drug costs; and the effectiveness of GPAs in preventing complications.
Estimated benefits used in the economic analysis
The number of PUB per patient per year was 0.0258 with NSAIDs and 0.0149 with rofecoxib. Thus rofecoxib led to an increment of 0.0109 in the number of PUB avoided.
Total estimated costs per patient per year were Can$584.91 with NSAIDs and Can$609.36 with rofecoxib. Therefore, the incremental cost per patient per year with rofecoxib was Can$24.45.
Synthesis of costs and benefits
An incremental cost-effectiveness ratio was calculated to combine costs and benefits of the two alternative treatments.
The incremental cost per PUB averted was Can$2,246.56.
The results of the sensitivity analysis showed that:
above the threshold rate of 27.5% for GPA co-prescription, rofecoxib became dominant;
if no GPAs were prescribed, rofecoxib became dominant when the rate of GPA prescription with NSAIDs exceeded a threshold value of 25%;
an increase in PUB risk of 215% (high-risk patients) caused rofecoxib to be cost-saving; and
the model was sensitive to variations in drug prices.
The authors concluded that, despite higher drug acquisition costs, the use of rofecoxib in place of NSAIDs for the treatment of OA reduced the incidence of serious GI adverse events at a modest incremental cost to a Canadian provincial government. Rofecoxib may become a dominant option in patients at increased baseline risk of severe GI events.
CRD COMMENTARY - Selection of comparators
The rationale for the choice of the comparator was clear. The authors stated that NSAIDs represent a standard and effective therapy for OA pain management in patients not responding to paracetamol. It may have been useful to add a paracetamol alternative to the analysis if possible. You, as a user of this database, should decide whether they represent a widely used treatment for patients with OA in your own setting.
Validity of estimate of measure of effectiveness
The effectiveness analysis was based on a review of the literature. The details and methodology of the review were not reported, but all 8 primary studies were double-blind, randomised, controlled trials, thus the source of evidence was appropriate. It would have been interesting to know what methods were used to combine the primary study data. The authors also made some assumptions, which were tested in the sensitivity analysis. The main assumption (equal efficacy of the two treatments) was based on published evidence.
Validity of estimate of measure of benefit
The number of avoided PUB per patient per year was used as the summary benefit measure in the economic analysis. It represented the natural outcome of the two treatments, as the focus of the analysis was not on the efficacy (which was assumed equal) but on the reduction of adverse events associated with the treatment. The benefit was modelled, and details of the decision model were reported. The use of avoided PUB does not permit a comparison of the benefits of the present study with those observed with other treatments funded by the health service provider.
Validity of estimate of costs
The perspective adopted in the study was stated and it appears that all relevant categories of costs were included in the analysis. The price year was reported, thus facilitating reflation exercises in other settings. However, although unit costs were reported, quantities of resources used were not, thus the reproducibility of the economic analysis in other settings may be difficult. The source of cost data was stated for each item and a detailed breakdown of the costs evaluated was provided. Costs were treated deterministically and sensitivity analyses were conducted only on drug costs and some assumptions concerning resource use. The authors stated that the potential inclusion of indirect costs (which were not relevant from the perspective adopted) would have favoured rofecoxib, as previous studies showed that the study drug reduced pain and disability in comparison with NSAIDs.
The authors made some comparisons of their findings with those from other studies, but did not address the issue of the generalisability of the study results to other settings. Some sensitivity analyses were conducted, but these were aimed at evaluating the robustness of the conclusions. Most of the estimates used in the analysis were specific to the Canadian setting and the authors acknowledged that data might vary even within Canada. The study referred to average-risk patients with OA who were not responding to paracetamol and this was reflected in the conclusions of the analysis.
Implications of the study
The main implication of the study is that rofecoxib may represent a feasible, safe, and cost-effective alternative to NSAIDs in patients with OA not responding to paracetamol treatment. The authors suggest that further studies should evaluate the impact of rofecoxib on compliance and related benefits.
Source of funding
Funded by Merck Frosst Canada Ltd.
Marshall J K, Pellissier J M, Attard C L, Kong S X, Marentette M A. Incremental cost-effectiveness analysis comparing rofecoxib with nonselective NSAIDs in osteoarthritis: Ontario Ministry of Health perspective. PharmacoEconomics 2001; 19(10): 1039-1049
Other publications of related interest
Langman M J, Jensen D M, Watson D J, et al. Adverse upper gastrointestinal effects of rofecoxib compared with NSAIDs. JAMA 1999;282(20):1929-1933.
Subject indexing assigned by NLM
Aged; Anti-Inflammatory Agents, Non-Steroidal /adverse effects /economics /therapeutic use; Cost-Benefit Analysis; Cyclooxygenase 2; Cyclooxygenase 2 Inhibitors; Cyclooxygenase Inhibitors /adverse effects /economics /therapeutic use; Decision Support Techniques; Gastrointestinal Hemorrhage /chemically induced /economics; Humans; Isoenzymes /metabolism; Lactones /adverse effects /economics /therapeutic use; Membrane Proteins; Ontario; Osteoarthritis /drug therapy /economics; Prostaglandin-Endoperoxide Synthases /metabolism; Stomach Ulcer /chemically induced /economics; Sulfones
Date bibliographic record published
Date abstract record published