Cancer drug access has been in the headlines again this week, following publication of articles by journalist A A Gill, in which he commented on (un)availability of the drug nivolumab for lung cancer on the NHS in England. In particular, his writing drew attention to the fact that the drug is currently available in Scotland for this indication, but not in England. Given the interest that his comments generated, we looked into the details of these funding decisions, as a case study in how similar countries can make different choices for a given drug in a particular indication.
By way of background: nivolumab (brand name Opdivo) is produced by Bristol-Myers Squibb, and is a member of a class of immunotherapy drugs known as “checkpoint inhibitors”; it blocks the interaction of a receptor known as PD-1 on immune T-cells, with PD-L1 on the tumour cells, thereby maintaining the activity of the T-cell to attack the tumour. It was one of the first drugs to receive Promising Innovative Medicine (PIM) status in the UK, and was given early access arrangements ahead of full marketing authorisation, through the UK’s Early Access to Medicines Scheme. It is licenced and funded for use in advanced melanoma where it is both effective and cost-effective; several additional indications including lung cancer have also received marketing authorisation.
Why isn’t it funded in England?
NICE hasn’t actually said no. The funding decision is still in progress. It’s been an extended process, with 2 consultations, and negotiation of a revised Patient Access Scheme (PAS) at a late stage. A decision is expected in January. (Update January 2017: this has now been delayed to allow BMS to submit additional data.)
Despite the PAS, the incremental cost effectiveness ratio (ICER) remained above that considered good use of NHS resources, even allowing for end-of-life considerations. (For those interested in the numbers: the ICER was £66-74K/QALY.) This contrasts with nivolumab in melanoma, which was considered cost effective with readily acceptable ICERs. Three main difficulties with the appraisal were identified:
- choice of curve used to extrapolate the overall survival data
- Quality of Life (QoL) data used, and
- duration of treatment
NICE has invited the company to submit a proposal for funding in the reformed Cancer Drugs Fund (CDF). Specifically, NICE has suggested BMS stratify the population by expression level of PD-L1 on the tumour cells, allowing cost-effective treatment for a subgroup where nivolumab is more effective. CASMI’s recent analysis of uncertainty in NICE’s appraisals of cancer drugs (Morrell et al, unpublished data) suggest that the CDF data collection plan will focus on QoL and duration, and might include exploration of clinical stopping rules based on some ongoing studies.
So why is it funded in Scotland?
The numbers are different. The cost-per-QALY estimates discussed by the Scottish Medicines Consortium (SMC) are lower than the ones seen by NICE, and low enough to be accepted as an end-of-life medicine. They are different despite using the same clinical data, the same comparator, and the same decision model as seen by NICE. In numbers: the company’s base case ICER is £47K/QALY, with scenarios exploring the same issues as NICE taking it up to £65K/QALY.
This is not the only example of different numbers between England and Scotland, but one of the few where they fall cleanly on either side of acceptability (CASMI’s unpublished data). Because of commercial sensitivity, it is not obvious from the public domain information what is driving this difference in this case, although there is an important contribution from the choice of extrapolation curve; it is also possible there are different commercial arrangements in the two nations.
What does this case contribute to our thinking on drug access? Firstly, this may be an example where indication-specific pricing would have facilitated access in lung cancer. Nivolumab’s price was accepted in the high value melanoma indication; because drugs have only a single price and Patient Access Scheme, the manufacturer can’t reduce the price to make the lung cancer indication cost-effective, without losing revenue from use in melanoma. Allowing indication-specific pricing would enable a ‘value-based price’ to be set for each indication – although of course, such an approach is not without its executional challenges.
Secondly, does this illustrate Scotland being more generous in its approvals than England? We can’t conclude that from this case, as the numbers fully explain the different decisions. Further, in other decisions for this class of drugs, there are examples where England’s approval has been more permissive than Scotland’s. It may be that manufacturers are prepared to reduce the overall price in a small market like Scotland, but this becomes untenable for the tenfold-larger population of England – an interesting inversion of the more usual ‘economies of scale’ logic.
Update January 2017: pembrolizumab – a similar drug, also approved in both countries for melanoma – was recommended for funding in lung cancer in January 2017, in both England and Scotland. So patients in A A Gill’s situation do now have the option of this type of treatment in both countries. This highlights another inequity observed in areas of rapid development and where patients have short life expectancy – inequity of timing. For some patients, had they been diagnosed six months later, their treatment options would have been different, which might have affected their outcomes. In this context, the goal of completing Technology Appraisal of cancer drugs within 90 days of marketing authorisation – introduced as part of the CDF reforms – is most welcome.
– Dr Liz Morrell