Problems with risk-sharing study must not deter future collaborations

“A costly failure” and “a fiasco” are just two of the epithets that have been used to describe the UK’s first-ever risk-sharing scheme to evaluate the long-term cost-effectiveness of two multiple sclerosis therapies, β-interferon and glatiramer acetate. Or as one commentator put it: “This may well be the most expensive publicly-funded ongoing health-related study in the UK, and probably anywhere, ever.”

The study was initiated in May 2002 after the National Institute for Clinical Excellence (as it then was) had recommended that the two treatments should not be used in the National Health Service because of concerns about their effectiveness and high price. NICE suggested instead that the Department of Health should collaborate with the products’ manufacturers to make them available in a more cost-effective manner; in other words, at a lower price.

The result was a scheme whereby the NHS agreed to make the products available while their long-term cost-effectiveness was being assessed. Four commercial products were included in the scheme: Biogen’s Avonex, Chiron’s Betaseron (marketed by Bayer), Serono’s (now Merck KGaA’s) Rebif (all of which contain β-interferon) and Yeda’s Copaxone (glatiramer acetate, licensed to Teva). The manufacturers, along with the Department of Health, would fund the monitoring study. Recruitment of patients who met the Association of British Neurologists’ criteria for relapsing remitting multiple sclerosis began in May 2002 and was scheduled to be completed by November 2003 (although in fact this did not happen until 2005).

The effectiveness of the various treatments was to be assessed by comparing the progression of the disease in the study patients with the progression observed in a historical Canadian cohort over a period of 25 years, with disability measured using the extended disability status scale.

The scheme was planned to run for ten years, with price-setting reviews every two years. If the observed benefit was less than that predicted by the model that NICE used in its analysis – the Sheffield School of Health and Related Research (ScHARR) model – the drug price would be reduced to achieve a target cost-effectiveness ratio of £36,000 per quality adjusted life year.

It was not long before the trial ran into controversy. In February 2003 two Scottish researchers drew attention to what they perceived as scientific flaws in the study, while two multiple sclerosis charities criticised it for taking too long to recruit patients (scripnews.com, 19 February 2003). Among the criticisms levelled against the trial were that it did not include a placebo group, and that it did not attempt to draw any comparisons with alternative, less expensive treatments.

Although the data were supposed to be reviewed every two years, the analysis of the first two years of data (covering 2005-2007) was not published until last year. This showed that the outcomes were much worse than predicted: not only was disease progression worse than predicted by the model used by NICE, but it was worse than that in the untreated control group. However, the study’s scientific advisory group deemed that it would be premature to reduce the prices of the products without further analysis.

This raises a number of issues.

• Why was the first data analysis carried out seven years after the study began, rather than two years as originally planned?
• What was the composition of the panel that decided against an immediate price review?
• On what grounds did the panel decide that the prices of the products should remain unchanged?
• Should this model be applied to other expensive new treatments, such as certain anticancer drugs?

Unfortunately, the answers to these questions are not all forthcoming. What we do know is that the scientific advisory group included representatives of the Department of Health, clinicians, patient groups and, significantly, the manufacturers. It did not include any representative of budget holders, who had the responsibility to pay for the drugs. This has inevitably led to claims of lack of impartiality. All the groups represented on the panel had a vested interest in maintaining the status quo, according to Christopher McCabe, professor in the Health Economics Unit at Leeds University, writing in the British Medical Journal. To be truly independent, the panel should have resembled the data monitoring committee of a clinical trial, Prof McCabe stated.

As for the reasons the panel gave for rejecting an immediate price cut, these were apparently threefold. First, an assumption made at the beginning of the study that patients’ disabilities could improve may have been invalid. Second, the study used historical controls, and the underlying epidemiology of multiple sclerosis may have changed over time. And third, the cost and utility data used in the ScHARR model may have underestimated the impact of the disability avoided through treatment, and thus underestimated the value of treatment.

As Prof McCabe points out, the organisers of the study were aware of all these factors at the outset. If they were not important enough to stop the scheme going ahead in the first place, why were they given so much weight seven years down the line?

In addition, it is questionable whether these arguments really do support the case for maintaining prices at their current levels. For example, although the incidence of multiple sclerosis seems to be increasing, the published evidence suggests that the disease is becoming less, rather than more, aggressive. The use of historical controls is therefore valid.

But the delay in publishing the first results remains a valid cause for concern. Depending on when individual patients entered the study, the analysis is anything up to five years late. The annual cost of the drugs used in the study is reportedly around £50 million: if the analysis had been performed when originally planned, the NHS could have taken immediate action and would have already saved up to £250 million by now. This is equivalent to 50,000 hip replacements or 15,000 coronary bypass procedures over the five-year period.

Some people have argued that, whatever the rights and wrongs of the study, patients have benefited, first through access to drugs that might not otherwise have been available, and second through the advance in knowledge that the study affords.

Others have been quick to suggest that the main beneficiaries have been the manufacturers of the products concerned, which have been able to sell their products to the NHS at close to full price, with the implication that it was inappropriate for manufacturers to be represented on the scientific advisory group, or that they had undue influence. This is unfortunate at a time when there have been suggestions that it was improper for the WHO to use experts with links to industry when deciding to declare an H1N1 influenza pandemic last year (scripnews.com, 8 June 2010).

When all is said and done, the multiple sclerosis risk-sharing study was an ingenious way of trying to address the recurring problem of how to pay for innovative new treatments in a way that benefited everyone. About a dozen similar schemes are currently underway within the NHS. It would be a shame if the difficulties encountered with this study were to scupper other efforts to find a solution to the cost-benefit conundrum.

If you are a subscriber and have a question arising from our news and analysis coverage, or would like us to do some research for you, use the free Ask the Analyst service by emailing us at [email protected]