Multi-arm, multi-drug trials from a reimbursement perspective. In Platform Trial Designs in Drug Development: Umbrella Trials and Basket Trials. (Book Chapter)
- Competence Center for Methodology and Statistics
In Europe, regulatory decisions concerning approval of drugs are made centralized by the European Medicines Agency (EMA), while reimbursement decisions are made decentralised at the national level by all member states. Historically access to new drugs was dependent on the regulatory approval, but in recent years this has shifted towards reimbursement being the restricting factor. Main drivers of this development are the ever increasing costs of new drugs, the decline in quality and robustness of submitted data, and the increasing pressure to accelerate drug approval and reimbursement processes. The traditional drug development focuses on a linear process where obtaining the regulatory approval is considered a ﬁrst hurdle, with reimbursement decision often coming into consideration after important decisions about the development program have been made already. The observed shift in importance of the regulatory versus reimbursement decisions has also led to an understanding that acceptance of for example trial designs should not only be sought with regulators, but also with other downstream stakeholders. Based on this understanding of shifting roles and need for closer interactions, EMA and the European network for health technology assessment (EUnetHTA) have facilitated interactions between drug developers, national health technology agencies (HTA), and EMA to provide consolidated advice on drug development programs. One of the complicating factors in providing a HTA viewpoint is the huge diversity of methodology used to perform assessments at the national or even regional level. Historically, European member states have different preferences in techniques and philosophy based on their political and ﬁnancial differences.1 The ﬁeld of HTA is also relatively young and still developing, something that makes reimbursement decision-making difﬁcult to predict across the spectrum. As a consequence, the HTA landscape for medicines in Europe is currently fragmented; however in a recent legislative proposal by the EC (January 2018; https://ec.europa.eu/health/technology_assessment/eu) cooperation was adopted with the aim to establish a European system for health technology assessment, and harmonise health technology assessment criteria in order to assess the added therapeutic value of medicines. Such a network with mandatory uptake of relative effectiveness assessments can have a major impact on transparency and policy in the future, but will likely not be fully operational before 2024. In December 2017, a workshop was organized by the EMA oncology working party on “Site and Histology – Independent Indications in Oncology.” As a contribution to this workshop, EUnetHTA conducted an informal survey among national members of EUnetHTA as well as international HTA contacts. The results of this survey are not meant to be representative or provide a consolidated opinion, but rather reﬂect the current viewpoint of those HTAs that participated in it. The survey focused on multi-arm, multi-drug trials such as ‘Umbrella type’ or ‘Basket type’ trials. Questions focused mainly on awareness, experience, and acceptability of these new trial designs. Five EUnetHTA members and two international HTAs contributed to the survey. All HTAs were familiar with the concept of Basket/Umbrella trials, but the lack of a consolidated deﬁnition of such trials was pointed out as one of the major problems at the moment. Not all agencies had seen such trial designs in submissions, and those that had received submissions including such trials had seen them mainly in early development phases, not as pivotal source of evidence per se. While none of the agencies a priori disapproved these types of trial designs, a number of potential problems were identiﬁed across all agencies. Concerns were raised about the potential lack of statistical power, all agencies had concerns about the uncontrolled aspects of such trials, and the issue of independence of arms was raised. These concerns mirrored those already identiﬁed by regulators and methodologists (and discussed in earlier sections/chapters) and indicate that the acceptance of HTAs will not necessarily differ too much from that of regulators per se. One aspect that was speciﬁcally problematic from an HTA perspective was the lack of collection of Quality of life (QoL) data seen in some recent examples. Umbrella-type trials, if deﬁned as testing different drugs or drug combinations in identical populations, are less problematic from an HTA perspective. Such trials would potentially allow a direct comparison between arms (directly if populations are truly identical or possibly indirectly if not) and assessment of relative effectiveness is possible. Problems would arise if all drugs/drug combinations would represent non-approved, nonreimbursed substances, since anchoring results to existing treatments might not be possible. Basket-type trials are perceived as more problematic, as they often resemble what can be perceived as a collection of single-arm trials, lacking the comparative context required to establish relative effectiveness. Since each arm/sub-study would be seen as a separate single-arm trial all problems connected to that approach would still apply, regardless whether the trial is conducted in a basket context (from a logistical perspective) or not. The only possible solution in absence of a comparative context is the attempt to establish relative effectiveness through comparison to historical or observational sources. Currently HTAs consider it problematic to use historical evidence to establish relative effectiveness; more progress has to be made in establishing common quality standards and criteria for such sources of real world evidence (RWE). Before such criteria have been established, the acceptance of use of RWE is largely dependent on the national HTAs regulations and is often determined case-by-case. All HTAs acknowledged that multi-arm, multi-drug trial designs have certain advantages. Faster and more efﬁcient drug development was mentioned by several HTAs, yet this is not considered a strong advantage from a reimbursement perspective if it does not result in lower prices and the guarantee that the provided evidence for decision-making has sufﬁcient robustness. Seemingly, requirements for regulatory approval have been lowered, but this needs to be seen in light of the changing landscape of drug development in smaller, more targeted populations, and the accompanying ethical and logistic problems of conducting studies in such indications. Nonetheless, the changes in evidence robustness pose an increasing problem, in addition to the decreases seen in sample sizes and short follow-up times in standard randomized clinical trials. Increasing numbers of submissions based on single-arm trials will only further aggravate the methodological challenges for the HTAs. In general, potential problems with statistical aspects such as lack of power, handling of missing data, randomisation, and independence when multiple assignment possibilities exist, as well as uncertainties around what other issues might be identiﬁed dependent on the chosen statistical analysis plan suggest that instead of a general acceptance of such trial designs a case-by-case approach will be required. Answering multiple research questions in one trial has been mentioned as another advantage, as well as the possibility to conduct trials in relatively small populations in rare and ultra-rare diseases. It might be that acceptance of multi-arm, multi-drug trials in rare diseases is largest, in particular seen in light of guidance available by several HTAs on evidence requirements being different in such disease areas. This is based on a generally higher acceptance of uncertainty when plans to ﬁll evidence gaps post-authorization are in place in rare diseases than in mainstream indications. An important feedback from the HTA side was that reimbursement based on biomarker selection is not considered impossible. As long as both an identiﬁable population and a standard of care for such a population can be established, reimbursement could be granted based on a validated biomarker.