Rheumatology Advance Access originally published online on May 13, 2008
Rheumatology 2008 47(7):1107; doi:10.1093/rheumatology/ken173
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Comment on: Modelling the cost effectiveness of TNF-
antagonists in the management of rheumatoid arthritis: results from the British Society for Rheumatology Biologics Registry: reply
1Centre for Health Evaluation and Outcome Sciences, St Paul's Hospital, Vancouver, Canada, 2Health Economics and Decision Science, School of Health and Related Research (ScHARR), The University of Sheffield, Sheffield, 3Arthritis Research Campaign Epidemiology Unit, The University of Manchester, Manchester and 4MRC Biostatistics Unit, Institute of Public Health, University Forvie Site, Cambridge, UK
Correspondence to: N. Bansback, Centre for Health Evaluation and Outcomes Sciences - Pharmaeconomics group, 620B-1081 Burrard St Vancouver, BC V6Z 1Y6, Canada. E-mail: nbansback{at}cheos.ubc.ca
SIR, We thank Dr Von Vollenhoven for his interest in our paper. The questions [1] he raises about our article that examined the cost–effectiveness of biologics using the British Society for Rheumatology Biologics Registry (BSRBR) [2] provide an excellent opportunity to describe why decision modelling should be used in calculating a majority of cost–effectiveness ratios.
We refer to an article by Sculpher et al. [3], which eloquently explains many of these points in detail. It should be noted that while the title of the paper refers to trials, much of its content is applicable to a registry like the BSRBR. We summarize the key points:
- It is necessary to base cost–effectiveness calculations on the question being addressed—e.g. are biologics in RA a cost-effective use of healthcare resources? rather than what data are available—e.g. the cost-effectiveness of a trial or a registry.
- A trial (or registry) will rarely provide enough information on its own to answer the policy question. The examples given by Sculpher et al. are (i) a failure to compare all relevant options, (ii) a truncated time horizon, (iii) lack of relevance to the decision context, (iv) inadequate quantification of decision uncertainty. The pragmatic nature of the BSRBR means evidence is provided that is superior to nearly all existing clinical trials in RA, which is why a new cost–effectiveness analysis was deemed necessary. However, not all issues are addressed completely. For example, the BSRBR alone could not produce a lifetime estimate as requested by the National Institute for Health and Clinical Excellence (NICE), but rather a 5-yr estimate, which has implications for long-term progression, and treatment switches. Neither is a complete resource utilization history included.
- Sculpher et al. [3] conclude that a combination of evidence synthesis with decision models will be the most appropriate approach for nearly all cost–effectiveness questions. ‘Models have sometimes been characterized as alternatives to trials, but this is to misunderstand their respective roles. The RCT (here the BSRBR) provides estimates of particular parameters in a specific group of patients in a particular healthcare environment. Decision models provide a structure within which evidence from a range of sources can be directed at a specific decision problem for a defined population and context.’
It is for these reasons we used the BSRBR to inform parameters for a specific decision problem. Using the ‘actual’ data whereby only the BSRBR data is utilized, would simply not provide an answer to the question posed by NICE. Unless models are used to extrapolate relatively short-term data to long-term outcomes the potential benefits of many rheumatological interventions may be underestimated. This may mean that, when compared with treatments for other diseases that have more immediate effects (e.g. cancer survival), they do not appear as worthy a use of healthcare resources.
This relates to the first point raised about the type of patients in the DMARD arm, where we believe modelling could actually help. We had too few patients in the control group (at the time of the analysis) to select patients newly starting a DMARD, so we chose to use all patients but adjust for key parameters. It is possible that there would now be sufficient patients to restrict the analysis to those starting a new DMARD or we could use an external data source to provide a new estimate. Only modelling provides a framework for this.
As new evidence emerges from the BSRBR and other registries, it should be included in the modelling to provide new estimates to inform particular policy questions.
Disclosure statement: The authors have declared no conflicts of interest.
 |
References |
|---|
 Top References |
|---|
- van Vollenhoven RF. Comment on:Modelling the cost effectiveness of TNF-
antagonists in the management of rheumatoid arthritis: results from the British Society for Rheumatology Biologics Registry. Rheumatology (2008) 47:1106–7.[Free Full Text] - Brennan A, Bansback N, Nixon R, et al. Modelling the cost effectiveness of TNF-alpha antagonists in the management of rheumatoid arthritis: results from the British Society for Rheumatology Biologics Registry. Rheumatology (2007) 46:1345–54.
[Abstract/Free Full Text] - Sculpher MJ, Claxton K, Drummond M, McCabe C. Whither trial-based economic evaluation for health care decision making? Health Econ (2006) 15:677–87.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||