Human Reproduction, Vol. 17, No. 6, 1673,
June 2002
© 2002 European Society of Human Reproduction and Embryology
Models of cost-effectiveness of recombinant FSH versus urinary FSH
Department of Obstetrics and Gynecology McMaster University 1200 Main Street West Hamilton Ontario, Canada L8N 3Z5
Correspondence: E-mail: dayas{at}mcmaster.ca
 |
Introduction |
|---|
 Top Introduction References |
|---|
Dear Sir,
Thank you for the opportunity to respond to the comments made by Drs Al-Inany and Afnan regarding our paper on cost-effectiveness modelling of recombinant (r)FSH versus urinary (u)FSH in assisted reproduction (Daya et al., 2001
). The two main points in their letter pertain to having up-to-date evidence to make inferences and the use of modelling in cost-effectiveness research.
In addressing the first point, it should be emphasized that the control arm of the meta-analysis (Daya and Gunby, 1999) focused on uFSH and not urinary gonadotrophins (such as HMG). Consequently, the reference made by Drs Al-Inany and Afnan to comparisons with HMGs is not germane to the discussion of the meta-analysis. Their comment that the additional studies they identified evaluating the efficacy of rFSH were testing `the exact question upon which [our] cost effectiveness was based' is incorrect because, in all five studies, the control group received HMGs. The sole study that included a uFSH arm (Gordon et al., 2001) was included in the Cochrane meta-analysis (Daya and Gunby, 2001). From a methodological point of view, their approach of using the crude aggregate data from the meta-analysis (as if these results were observed from one study) and adding the data from each of their newly found studies is a simplification of the technique of meta-analysis and is statistically incorrect. In fact, the odds ratio (OR) they used for this purpose was 1.17 [95% confidence interval (CI) 1.01 –1.37), which is different from the OR of 1.21 (95% CI 1.04 –1.42) observed in the Cochrane meta-analysis (Daya and Gunby, 2001). Interestingly, Drs Al-Inany and Afnan missed five other new studies that did compare rFSH and uFSH; these studies will be included in the Cochrane meta-analysis when it is updated.
The meta-analysis (Daya and Gunby, 1999) to which we referred in our cost-effectiveness study provided clear evidence of an improvement in pregnancy rate with rFSH compared with uFSH (overall OR 1.20, 95% CI 1.02–1.42). The probabilities for the various outcomes used in the cost-effectiveness analysis were estimated from the best evidence available at the time, including data from the Cochrane meta-analysis, which had been submitted but not yet published at the time the manuscript was being prepared. The Cochrane meta-analysis included six more studies than the Human Reproduction publication, but the overall ORs for clinical pregnancy were virtually identical. Thus, empirical evidence and not assumption was used to provide the probability estimates for our model.
The second point pertains to the need for modelling in cost-effectiveness analysis. The view of Drs Al-Inany and Afnan that uFSH is more cost-effective than rFSH because the latter is more expensive is simplistic and one-sided because cost-effectiveness analysis involves a comparison of both costs and outcomes. Typically, treatment with assisted reproductive technologies is a multi-step, multi-cycle approach in which many different outcomes are possible (e.g. successful ovarian stimulation, successful oocyte retrieval, fertilization, cleavage, implantation and so on) and these outcomes may vary from cycle to cycle. Randomized trials designed to evaluate the cost-effectiveness of interventions of such complex treatments require considerable resources, a very large sample size and a lengthy follow-up period to provide any meaningful information from all the outcome scenarios that can be encountered. Although such trials are possible to conduct, they are unlikely to be undertaken largely because the costs involved are prohibitive.
An effective method of overcoming these limitations is to employ modelling techniques with computer simulation, an approach that is well established in cost-effectiveness research. This method also allows confidence intervals to be constructed around the estimates of the various outcomes so that the precision of the overall results can be ascertained. In our analysis, rFSH, despite its higher costs, was found to be more cost-effective than uFSH because the total amount of gonadotrophin required was lower and the clinical pregnancy rate was higher with rFSH. Further, results observed in the UK study have now been replicated in studies we have completed in USA, Spain and Germany. These studies all confirm that significantly fewer cycles are required to achieve a pregnancy when rFSH is used instead of uFSH.
In conclusion, the results from valid randomized trials confirm the clinical superiority of rFSH in treatment with assisted reproductive techniques and its greater cost-effectiveness confers benefit that will be appreciated by policy makers, health care providers, insurance companies and consumers.
|
References |
|---|
|
TopIntroductionReferences |
|---|
Daya, S. and Gunby, J. (1999) Recombinant versus urinary follicle stimulating hormone for ovarian stimulation in assisted reproduction. Hum. Reprod., 14, 2207–2215.
Daya, S. and Gunby, J. (2001). Recombinant versus urinary follicle stimulating hormone for ovarian stimulation in assisted reproduction cycles (Cochrane Review). In The Cochrane Library, 2001, Issue 4. Update Software, Oxford, UK.
Daya, S., Ledger, W., Auray, J.P. et al. (2001) Cost-effectiveness modeling of recombinant FSH versus urinary FSH in assisted reproduction techniques in the UK. Hum. Reprod., 16, 2563–2569.
Gordon, U., Harrison, R., Fawzy, M. et al. (2001) A randomized prospective assessor-blind evaluation of luteinizing hormone dosage and in vitro fertilization outcome. Fertil. Steril., 75, 324–331.[ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||