|
A cost effectiveness analysis of cyclosporine in rheumatoid arthritis |
Anis A H, Tugwell P X, Wells G A, Stewart D G |
|
|
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 second line agents for the treatment of patients with severe rheumatoid arthritis (RA). Specifically the agents examined were the immunosupressants cyclosporine (CyA), D-penicillamine (D-pen) and azathioprine (Aza).
Economic study type Cost-effectiveness analysis.
Study population Patients with severe rheumatoid arthritis eligible for treatment with a second line agent.
Setting Hospital and community. The economic analysis was conducted in Vancouver, British Columbia, Canada.
Dates to which data relate Effectiveness data were taken from a meta-analysis, the results of which were published in 1993. Resource data were collected during 1991 and 1991 prices were used.
Source of effectiveness data The results of a meta-analysis were used to estimate the probability of a reduction in tender joints, an increase in functional ability and the incidence of adverse drug reactions.
Modelling A decision analysis model was used to combine the estimates of probability of patient improvement, adverse effects and costs of treatment during a one year period, using CyA, Aza, D-pen or a placebo. Two probability trees were used, one representing the placebo comparison and the other representing either of the two head-to-head comparisons.
Outcomes assessed in the review An assessment of functional improvement. Specifically a 25% or greater reduction in tender joint counts, or a 25% or more functional improvement using Lee's index were used. Adverse drug episodes were also reported.
Study designs and other criteria for inclusion in the review Double blinded randomised controlled trials comparing CyA with either placebo, Aza or D-pen were included 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 5 primary studies were included in the analysis. Three trials compared CyA with a placebo, and the remaining two used Aza and D-pen respectively as the comparators. One of the trials was of one year duration, three of six months and one of four months.
Methods of combining primary studies Studies were combined by meta-analysis using a fixed effects model. Two approaches were used: the first combined only the 3 placebo-controlled studies; and the second compared all 5 CyA groups to the 3 placebo groups. The significance of the overall treatment effect was assessed using chi-squared techniques.
Investigation of differences between primary studies Tests for homogeneity of treatment effect were conducted using Cochrane's Q statistic. This was not found to be significant.
Results of the review The results for all CyA patients compared to all placebo patients showed that the proportion of patients with 25% reduction in tender joint counts was 0.353 (SD 0.0297) for CyA, compared to 0.174 (SD 0.0302) for the placebo group. This corresponds to 35% of CyA patients who improved compared to 17% of placebo patients. The average reduction in tender joints was 2.54 (SD 0.301) and 0.68 (SD 0.326) respectively.
The proportion of patients reporting a 25% improvement using Lee's functional index was 0.518 (SD 0.0417) for CyA compared to 0.385 (SD 0.0524) for placebo. The respective average improvements in the functional index were 3.21 (SD 0.360) and 1.0 (SD 0.426). There appears to be little difference between the results for CyA using all 5 groups compared to using only the three directly placebo controlled trials, although no statistical analysis of this appears to have been carried out. The results for the 5 CyA groups combined were used in the model.
The head-to-head drug comparisons found no significant difference in improvement rates between CyA and either D-Pen or Aza.
As the studies were of varying durations, the rates of adverse drug reactions (ADRs) in the individual studies were extrapolated to one-year durations by doubling the number of ADRs in the 6 month trials and trebling the number in the 4 month trial. These extrapolated rates are as follows:
CyA vs Placebo (1st 6 months); Overall probabilities of ADR per patient were 0.8123 and 0.6165.
CyA vs Placebo (2nd 6 months); Overall probabilities of ADR per patient were 0.7363 and 0.4955.
CyA vs Aza (annual); Overall probabilities of ADR per patient were 0.7797 vs 0.3103
CyA vs D-Pen (annual); Overall probabilities of ADR per patient were 0.6957 compared with 0.5106.
The majority of ADRs were categorised as mild as opposed to moderate or severe.
Measure of benefits used in the economic analysis The measure of benefits used was the proportion of 'patients improved', defined as those with a 25% or greater reduction in number of tender joints. For the comparison between CyA and Aza or D-Pen no specific measure of benefit was used since the efficacy analysis showed no difference in clinical effectiveness between the three drugs.
Direct costs The costs of drugs, hospital costs, diagnostic and other related costs were estimated. No costs were discounted and 1991 prices were used. Costs were estimated from both the societal perspective and the perspective of a third party payer. The costs of medication were estimated using the Ontario formulary and the best available price and a 10% surcharge were used where data were available. Dispensing fees for medication were based on a weighted average of the market rate and the rate for those patients supported by the Ontario Drug Benefit programme. The costs of treating adverse drug reactions were in part based on estimates of resource use provided by an expert panel. Hospital costs included costs for diagnostic tests, surgical procedures, physician fees and length of stay in different types of hospital ward. The costs of diagnostic tests were estimated by multiplying the unit cost of the test by the number of workload measurement units assigned to the test. These unit costs were taken from a hospital cost model for McMaster University Medical Centre, and the number of measurement units assigned to each test was obtained from different laboratories at the Ottawa Civic Hospital. Surgical costs included costs for the operating room, physician fees and recovery room costs. The length of stay for hospitalisations resulting from adverse drug reactions were determined from a questionnaire sent to the expert panel. Over-the-counter medication costs were priced according to typical pharmacy prices in 1991. For the third party payer perspective a distinction was made between those whose medication expenses were covered by the Ontario Drug Benefit programme.
Indirect Costs Costs for lost productivity were estimated in order to provide a societal perspective. The total amount of time off work was estimated by the expert panel and this was valued based on an estimate of average life time earnings. 1991 prices were used.
Sensitivity analysis One and two way sensitivity analyses were performed on the price of CyA (varied between 75% and 125% of base cost), costs of hospitalisation (75% - 125% of baseline) and the proportion of patients covered by the Ontario Drugs Benefit Programme (90% - 110% of baseline). In comparison to the placebo group only the price of CyA was varied as the authors assumed that less error due to pooling would be anticipated.
Estimated benefits used in the economic analysis The proportion of patients who improved (with at least a 25% reduction in tender joint counts) was 0.179 (17.9%) higher in the CyA group compared to the placebo group (0.353 vs 0.174).
Cost results The total annual costs per patient improved for each of the comparisons and from each perspective were as follows:
Societal Perspective:
CyA versus placebo: Can$4,230 compared with Can$525;
CyA versus Aza: Can$4,205 compared with Can$1,319;
CyA versus D-Pen: Can$4,598 compared with Can$867.
Third Party Payer perspective:
CyA versus Placebo: Can$2,612 compared with Can$545;
CyA versus Aza: Can$2,559 compared with Can$1,086;
CyA versus D-Pen: Can$2,663 compared with Can$1,044.
Synthesis of costs and benefits For the comparison between CyA and a placebo control group the incremental costs per patient improved (25% reduction in tender joint counts) were Can$20,698 from the societal perspective and Can$11,547 from the third party payer perspective. For the cost minimisation between CyA and Aza the incremental costs per patient improved were Can$2,886 and Can$1,437 from the societal and third party payer perspectives respectively. Similarly for the comparison between CyA and D-pen these incremental costs were Can$3,731 and Can$1,618 respectively.
Sensitivity analysis found that the results for CyA versus placebo were relatively insensitive to the price of CyA. The price of CyA was of greater significance in relation to Aza and D-Pen. A 25% decrease in the price of CyA led to a 23% and a 28% fall in the incremental cost of Aza and D-Pen respectively (from Can$1,473 to Can$1,138 for Aza, and from Can$1,618 to Can$1,167 for D-Pen).
Authors' conclusions The incremental cost of CyA compared to Aza and D-pen is roughly Can$1500 higher for the same clinical benefit. Although this may be reasonable from an individual patient perspective, given budgetary constraints the use of CyA should only be justified after other less expensive and more effective therapies have been used.
CRD COMMENTARY - Selection of comparators A justification was given for the comparators used. The comparators chosen (Aza and D-Pen) have been the subject of clinical trials which have demonstrated that they have similar efficacy to CyA. However the authors noted that the comparators chosen were limited to these two due to the lack of information on other immunosuppressants which could otherwise have been used.
Validity of estimate of measure of benefit The measure of benefit used seems appropriate, however since the methods used to identify studies for the meta analysis were not explicit it is not clear whether all of the relevant literature was considered. For the head-to-head comparisons, the efficacy of the 3 drugs were assumed to be similar on the basis of only two randomised controlled trials, neither of which directly compared the three drugs. As no individual study details were provided it is difficult to assess the validity of this assumption, however it is unlikely that the evidence was sufficiently strong to support his assumption.
Validity of estimate of costs Insufficient details were provided in this paper of the calculation of some costs, such as those for hospitalisation and the exact methods use for calculating indirect costs.
Other issues The cost data may not be generalisable outside Canada, as much data were derived from national publications and a Canadian hospital cost model. Much of the information on the cost-effectiveness analysis has been reported in a more detailed technical publication. The sensitivity analysis on the results for CyA versus placebo was conducted only on a single variable (the price of CyA) as it was assumed that less error due to pooling would be anticipated in this comparison. This is not a valid assumption to make, and in any case the purpose of sensitivity analysis is not to correct for errors in the analysis, but to examine the robustness of the results of the model and its sensitivity to changes in component variables.
Implications of the study More information is required on the use of immuno-suppressants for the treatment of rheumatoid arthritis. In particular,information is required on drugs other than those mentioned in this study. Guidelines for the use of immuno-suppressants should seek to identify that specific sub group of patients who may benefit to a greater extent from the use of CyA than from other treatment options.
Source of funding Funded by Sandoz Canada Inc.
Bibliographic details Anis A H, Tugwell P X, Wells G A, Stewart D G. A cost effectiveness analysis of cyclosporine in rheumatoid arthritis. Journal of Rheumatology 1996; 23(4): 609-616 Other publications of related interest Wells G, Tugwell P. Cyclosporin A in rheumatoid arthritis: Overview of efficacy. British Journal of Rheumatology 1993;32(Supplement 1):51-6.
Anis A, Tugwell P X, Wells G et al. A cost effectiveness and meta analysis of cyclosporine in rheumatoid arthritis. Technical report for Sandoz Canada Inc. Ottawa: Ottawa Civic Hospital, 1994.
Indexing Status Subject indexing assigned by NLM MeSH Antirheumatic Agents /economics /therapeutic use; Arthritis, Rheumatoid /drug therapy /economics; Azathioprine /economics /therapeutic use; Cost-Benefit Analysis; Cyclosporine /economics /therapeutic use; Humans; Models, Economic; Ontario; Penicillamine /economics /therapeutic use; Randomized Controlled Trials as Topic AccessionNumber 21996000485 Date bibliographic record published 30/11/1998 Date abstract record published 30/11/1998 |
|
|
|