Cancer drug funding decisions in Scotland: impact of new end-of-life, orphan and ultra-orphan processes

Drugs to treat rare conditions (including those formally defined as orphan and ultra-orphan drugs), and drugs to treat patients in their last months of life, have been the subject of contentious funding decisions in many countries, including the United Kingdom (UK). For rare conditions, small patient populations often result in the drugs being expensive, and they may not have the necessary level of evidence when submitting to reimbursement agencies to support a funding decision with certainty. For drugs related to end-of-life care (notably in cancer), there is a belief that extensions to the last months of life are particularly valuable to patients and their families [1] although the empirical data on public preferences for funding such drugs are equivocal [2].

Health Technology Appraisal is a devolved responsibility in the UK, and England and Scotland responded differently over time to the challenges of assessing these drugs. The first adjustment was by National Institute for Health and Care Excellence (NICE) in England in 2009, introducing its ‘End of Life Criteria’, which define situations in which evaluation committees can place greater weighting on health gain for drugs to treat patients with short life expectancy [3]. The Scottish Medicines Consortium (SMC) did not make similar adjustments at that time; however in 2012 ‘modifiers’ were introduced which allow SMC greater flexibility under specific conditions, in their acceptance of uncertainty in the economic case or a high cost per Quality-Adjusted Life Year (QALY); the uncertainty modifier applies to orphan drugs, but end of life is not specifically mentioned [4].

SMC introduced a further set of reforms in 2014, specifically for medicines to treat EOL and rare conditions. The changes take the form of an add-on process that is triggered only if the initial advice of SMC’s New Drug Committee (NDC) is not to recommend funding [5, 6] and include:

The SMC definitions are different from the end-of-life criteria used by NICE, both structurally and numerically (Table 1), and are more permissive than NICE’s current criteria.

The 2014 SMC changes are significant in the context of UK health technology appraisal, as a novel approach to incorporating broader aspects of value into cost-effectiveness analysis. With continued calls for NICE reform from advocacy groups, and similar pressures in other countries, it is important to evaluate the Scottish approach as a possible model for assessing value beyond the QALY, or as a pointer to key factors to include. Further, the changes could have important effects on patient access to new drugs in Scotland, with implications for spending priorities within NHS Scotland and for cross-border parity of patient access between the nations of the UK. Cancer is of particular interest, as a category that is expected to benefit from the end-of-life element, and where there are existing disparities in access arrangements, due to the existence of the Cancer Drugs Fund (CDF) in England and the pre-existence of NICE’s EOL criteria. Hence this paper evaluates the impact of the changes on funding decisions in the first 2 years of their operation, focusing on cancer, with the aim of contributing to ongoing debate on evolution of health technology assessment processes in the UK and elsewhere.

SMC advice on submissions based on the new processes was published from October 2014 [6], so advice published from October 2014 to the date of analysis (September 2016) was selected. Comparison data for the 2-year pre-period were taken from advice published October 2012 to September 2014. Full and Abbreviated Submissions, Resubmissions, Superseded Advice and Independent Review Panels were included; non-submissions and Marketing Authorisation withdrawals were excluded. Resubmissions (ie re-application for a drug that had previously had a Not Recommended final decision from SMC) were analysed as part of the complete data set, and also separately, as we hypothesised they could be informative on any differences in decisions following the reforms. Cancer drugs were identified as British National Formulary (BNF) category 8: Malignant disease and immunosuppression, as classified by SMC, and selecting submissions for cancer indications.

Advice Summaries and Detailed Advice (pdf) documents were accessed from the SMC website, and data extracted and tabulated in Excel for descriptive statistics and thematic analysis. Data included submission type, EOL and rarity status, decision, incremental cost effectiveness ratios (ICERs), restrictions, use of PACE and PAS, and the PACE summaries. Descriptive statistics of the extracted data were analysed in Excel.

Because a simple before/after analysis does not isolate the effect of the changes, we took advantage of the structure of the reform, as an additional process triggered only for submissions that received a preliminary NDC position of a) Not Recommended, or b) Restricted more narrowly than requested by the manufacturer [7]. We therefore made the assumption that all drug candidates using the new process were initially Not Recommended, and that a change in that initial decision was due to use of the new process. Drugs using the new process were identified by the statement in the Advice headline that the submission was assessed “under the end of life/orphan/ultra-orphan process”.

Thematic analysis of the PACE meeting summaries was carried out by an inductive process: that is, by constructing themes suggested by review of the text rather than an imposed framework. Themes were developed by determining the core idea of each phrase, and linking similar ideas together to form themes. A given phrase was only linked to one theme. Related themes were grouped together for ease of presentation and understanding. The analysis was carried out by two researchers. One researcher developed an initial thematic framework from a sample of PACE summaries; a second researcher analysed the majority of the summaries, making adjustments to the framework as needed and in discussion with the first researcher, who completed the analysis. Thematic analysis was carried out using MaxQDA software (v12).

Data on NICE appraisals for the same drugs were accessed on the NICE website, and data extracted from the published Technology Appraisal Guidance documents (Final Appraisal Documents only; Appraisal Consultation Documents were excluded as these could change as a result of consultation). Data included approval status, whether EOL criteria were met, and reason for not meeting EOL criteria. For comparison of access between countries, restricted or optimised approvals were considered as ‘access’ along with full recommendation, and the most recent decision on a given drug-indication (that is, the one active in September 2016) was included to provide a point-in-time snapshot of drug availability.

Coverage by the Cancer Drugs Fund in England was determined from the CDF list of September 2015, and the relevant CDF Decision Summaries, accessed from the NHS England website.

This analysis of funding decisions indicates that the SMC changes appear to have resulted in increased access to drugs for end of life and rare conditions, with up to 38 cancer drugs gaining access that they might not have had pre-reforms. The new, additional process was used most commonly for cancer drugs, and the proportion of cancer drugs triggering this process (due to a preliminary Not Recommended decision) is higher than would be predicted by the Not Recommended rate in the pre-period. The relative contribution of the PACE meeting, additional PAS opportunity, and increased ICER flexibility is unclear from this work. Differences between Scotland and England in classification of drugs as EOL lead to cross-border differences in EOL status, but do not fully explain differences in access; other factors include the agencies’ timeframe for appraisals, and which drugs are within their remit.

Our findings are consistent with early comments [8, 9] and a report [10] from SMC, and the findings of the recent Montgomery review [11], all of which suggest the changes are enabling SMC to approve drugs they could not have accepted before. They are also consistent with findings of Garau et al. [12] and Barham [13] in similar analyses. Both of those studies take a longer pre-period, creating a larger dataset, whilst this study chose 2 years to compare with the most recent submissions pre-reforms. Neither those studies, nor this one, have attempted a complete ICER analysis, because of the challenges of extracting comparable numbers from the appraisal documents, due to confidentiality and the presentation of ranges or multiple ICERs. To the best of our knowledge, our study is unique in providing the within-submission comparison, thematic analysis, and comparison with England.

The higher than expected proportion of the cancer submissions triggering the new process could be because this was an inherently challenging cohort of drugs, or because the reformed process – particularly the ICER flexibility – attracted submissions at higher ICERs than in the past. Alternatively, it may be that NDC’s readiness to use the new process is higher than the historical rate of Not Recommended decisions; for example, by being cautious in giving an Accept recommendation, which denies the opportunity for a new or revised PAS and a PACE meeting, thus losing an opportunity to gather additional information and potentially creating difficulties if SMC subsequently makes a Not Recommended decision. Irrespective of the reason, SMC appears to be handling more PACE meetings than might have been anticipated from previous decisions, and this may have implications for resources within SMC.

The nature of the impact of the PACE meetings on funding decisions remains unclear. Similarities between the PACE summaries may be created in part by the structured nature of the meeting and summary template, and the need to provide a concise summary for the appraisal committee, although a study published by Cancer52 found that PACE participants thought the summaries were an accurate reflection of the discussion [14]. The process could perhaps be improved by increased transparency in how the PACE input has been used. A subset of the themes are additional information that is not reflected directly in the quantitative cost-effectiveness analysis; of note, several of these are consistent with ideas for process reform considered by NICE under Value-Based Pricing (eg severity of the condition, innovation, wider social benefits) [15], whilst others such as the value of hope, and impact on the family, suggest further societal values that could be considered for future reform.

Our analysis identifies disparities in funding decisions between Scotland and England reflecting both process and decision differences between SMC and NICE. From a patient perspective, this disparity was reduced in reality by the existence of the CDF in England, which funded several of the drugs rejected or pending by NICE. The future funding of these CDF drugs is still being determined. The CDF reforms remedied some of the process differences between NICE and SMC, notably: a) removal of the rarity requirement in NICE’s end-of-life criteria, bringing the definitions into closer alignment; b) expectation that NICE will publish a decision more quickly than previously, normally within 90 days of the marketing authorisation for cancer drugs; and c) requirement that all new cancer drugs will be referred to NICE, which was previously not the case. However, as we saw in this cohort of drugs, even where both agencies have completed an evaluation under EOL conditions they can still come to different decisions, so some differences in access are likely to remain, reflecting other process differences, differences in clinical practice, implicit preferences, and varying commercial arrangements with the pharmaceutical industry.

A limitation of our study is our assumption that all the submissions that triggered the new process, would have been rejected had the reforms not been in place. Some of them might in fact have been granted some level of access following further deliberation and negotiation. Importantly, NDC do not use the ‘modifiers’ available to SMC to allow additional flexibility in specified circumstances; application of these modifiers may have supported a Recommended decision by SMC even without the reforms, despite an initial Not Recommended position from NDC. Comments from stakeholders on the NDC decision, or new analysis suggested by the NDC review, could also lead to SMC reversing an initial Not Recommended NDC position. In either case, our analysis would overestimate the number of additional approvals permitted by the reforms. Further, as outlined in the Methods section, the new process can also be triggered if the NDC are minded to impose a tighter restriction on use than proposed in the submission; however, if we take account of the 11 cases identified in the PACE analysis - assuming they were at risk of involuntary restriction rather than rejection - our observations on distribution of decisions still hold.

A second limitation is our use of information available in the public domain, which includes summaries by SMC staff and excludes commercially confidential information, leading to risk of bias due to missing and secondary information - although our analysis does reflect the summary information available to SMC committee at the point of decision. The summarised nature of the public documents in particular limits our ability to determine which elements of post-reform resubmissions – the new process, or other improvements to the case for funding – led to the reversal of the previous decision. Further, we have not considered whether the additional spending represents value for money for NHS Scotland, which would require a comparison of the health gain achieved with expenditure; estimating the actual costs is particularly challenging given the confidentiality of any commercial arrangements. If acceptance of these drugs for funding is at higher ICERs than would be accepted for other drugs, opportunity costs are imposed elsewhere in the health system. It is not clear from the empirical literature that such prioritisation of medicines for end of life and rare conditions is in line with the preferences of the Scottish public [16].

Strengthening the PACE analysis by using verbatim transcripts of the meetings rather than summaries, and supplementing with qualitative work exploring the experiences of key players in the process, could provide additional insight into the impact of PACE on decisions. In addition, the reforms occurred in parallel with SMC moving to holding appraisal committee meetings in public, and increase of the New Medicines Fund to £90million to help cover the cost of these medicines. In combination, these changes may have had additional effects beyond numbers of approvals, such as affecting buy-in to decisions, or encouraging pricing flexibility, and the contribution of these diverse elements deserves additional exploration. Further work is also needed to determine whether the choices made under the new SMC process provided value for money, ideally by examining actual outcomes from use of the drugs in practice. The CDF in England has been criticised for failing to provide such an evaluation of its expenditure [17], and data are currently not available consistently across Scotland [11]; findings from usage in Scotland may provide useful insight for future evolution of technology evaluation processes and potentially for increased unification of decisions across the UK.

We identified up to 38 cancer drugs that were funded post-reform that might previously have been rejected, but the contribution of each element of the reforms to the funding decision remains unclear. Differences in decisions between England and Scotland are not fully explained by differences in definitions of EOL/rare conditions, and disparities in access are likely to remain despite recent changes in England. Future improvements to the process could include enhanced transparency in use of the PACE input, and evaluation of patient outcomes from funding access to these drugs.