The pharmaceutical press were quick to express their disappointment earlier this year at a draft decision by NICE, not to fund pembrolizumab (Keytruda: MSD) for first-line treatment of metastatic non-small cell lung cancer. The Appraisal committee concluded there was such uncertainty in the evidence base that it was impossible to home in on a plausible cost effectiveness ratio; however, estimates from a range of scenarios, stretched only upwards from NICE’s usual threshold of £50 000 per QALY – so despite the high level of evidence uncertainty, the decision itself was less uncertain.
The uncertainties in this appraisal illustrate some recurring themes in the decisions faced by NICE’s Appraisal Committees; they are worth a brief exploration here, ahead of the final decision due in June.
Firstly, the key clinical trial was stopped early, because pembrolizumab showed statistically superior overall survival compared to chemotherapy; the trial’s governance body stopped the trial to allow the patients on chemotherapy to receive pembrolizumab. This is an ethical imperative for those patients, whilst still providing strong benefit/risk data for regulators in deciding whether to allow the drug onto the market – in fact pembrolizumab in this indication was given expedited approval by MHRA, through the UK’s Early Access to Medicines Scheme. However, it creates challenges for assessing comparative effectiveness, because the investigators lose the ability to confirm what happens to those pembrolizumab patients in the long term. This is a recurring bind for promising drugs – the better the survival, the longer it takes for the survival data to mature, and the less ethically viable the control arm becomes. The issue for funding decisions is that the health service is being asked to pay for benefits, of which (in this case) over half accrue in the period beyond the extent of the data. That’s a huge unknown to gamble with taxpayers’ money.
Secondly, and related to the immaturity of the survival data: the optimal duration of treatment is unknown. This is another recurring theme with drugs that are taken to progression, and is not limited to immunotherapies – for example pomalidomide and lenalidomide in haematological cancers presented the same challenge for decision-makers. Options to manage the financial uncertainty include stopping rules (which have been used for immunotherapies in both England and Scotland) or dose capping schemes (eg lenalidomide in multiple myeloma, where the NHS only pays for the first 26 cycles of treatment).
Finally, there were questions about health state utilities: Health-Related Quality of Life (HRQoL) reported for particular disease states in the trial, was better than ‘normal’ HRQoL in the population. That isn’t consistent with patient and clinicians’ description of the physical and psychological distress caused by this condition. We have observed other similar examples in previous appraisals, and they – unsurprisingly – always cause difficulties for the Appraisal Committee. It’s a phenomenon worthy of further research.
In contrast, consider nivolumab (Opdivo: BMS) – another immunotherapy launched within months of pembrolizumab, and targeting the same indications. To some disappointment, nivolumab did not achieve its primary endpoint in this indication, and the manufacturer is not now seeking marketing authorisation from EMA. It is not entirely clear why nivolumab failed where pembrolizumab performed so well, although one difference between the trials was the level of expression of the drugs’ target in the patients’ tumours.
So the current situation is that neither drug is available for first-line use: one because the trial failed, and the other – paradoxically – because it did so well. Managing that paradox needs a pricing mechanism that pays for the benefits that are known about, with a commitment to adjust price as we build evidence of the long-term benefit. That might be through the reformed CDF, with continued monitoring of ongoing trials, and data collection on duration of treatment. However, entry to the CDF requires the manufacturer to agree a price that’s ‘plausibly’ cost-effective, and the manufacturer’s freedom to move may be being constrained by the price already established for earlier, high-value indications. Is this a case that would be facilitated by indication-specific pricing?
Update, July 3rd 2017: The guidance on pembrolizumab in this indication was published by NICE on June 28th 2017, and recommends funding through a managed access agreement in the Cancer Drugs Fund. The ongoing KEYTNOTE-024 trial will provide survival data, with real-world data being collected on duration of use within the NHS.
-Dr Liz Morrell