Dr Virginia Acha, The ABPI's Executive Director of Research, Medical & Innovation, addresses BMJ research into the ‘Availability of evidence of benefits of overall survival and quality of life of cancer drugs approved by European Medicines Agency: retrospective cohort study of drug approvals 2009-13’.
The article by Davis et al[i] on the efficacy of cancer medicines raises some important issues about how we establish and evaluate the evidence of a medicine. The authors pose questions about the approaches undertaken by leading international regulatory agencies, focusing in this case on the European Medicines Agency (EMA) in response to a similar study undertaken in the US, which considered assessments by the US Food and Drug Administration (FDA)[ii]. Central to their approach is a view that the correct means of evaluating the value of a medicine is the results in terms of overall survival and quality of life benefits. Intuitively that makes sense; we can imagine that cancer patients would want to know that the treatments they take would help them to live their lives with the best quality possible and for as long as possible.
However, research design is bounded by the nature of what we are studying and the means to assess the evidence. Some of the novel cancer medicines under scrutiny in this study have intrinsic (its efficacy, its selection to certain subgroups of patients) and extrinsic (trial design where patients are switched following progression, ethical requirements) features that mean that overall survival is not an effective measure by which to determine benefits over risks. Instead, we look to establish more proximate measures that either demonstrate clinical benefit in their own right (e.g. progression free survival[iii]) or are demonstrably predictive of clinical outcomes and benefit.
To illustrate this further, we can draw an analogy to the introduction of an educational intervention. Innovators in education introduce a new intervention in children's education to improve their life chances and their quality of life. It is difficult to assess the value to the individual on these measures in a practical timeframe to determine if the intervention is beneficial and/or cost-effective. We would be waiting decades to know whether the intervention was valuable by those measures. Instead, research design for educational intervention includes more proximate measures to assess value, and researchers seek to confirm the outcomes over time and through real data on these individuals.
Industry responded to the article by Davis et al[iv] soon after it appeared in press[v], raising a question about why, in seeking to assess what gains to overall survival and quality of life are derived from approved cancer medicines, the authors look not to real world data on actual patient outcomes, but to randomised clinical trials. Given that the sponsors had determined that it was not feasible or appropriate to undertake the clinical research to capture overall survival for these medicines in the first place, perhaps it is not surprising that a meta-analysis of these randomised clinical trials yields little difference in result. Instead, the means to assess the longer term effects on overall survival and other measures (including quality of life) might be better addressed through real world evidence, including registries.
The real question is: do our current approaches to regulation and health technology assessment suffice to answer the question to the best of our scientific abilities whether these cancer medicines bring value to patients. Regulators, like the EMA, hold patient benefit and safety central to their purpose, and the rigour by which they undertake these assessments involves considerable peer review across 28 Member States and EEA Observer countries as well as publication of results. Internationally, there is clearly cross-referencing of decisions and discussions over scientific disagreements in regulatory science and assessments by the leading regulatory authorities. In the UK, NICE applies world-leading rigour to its assessments of cost-effectiveness of medicines, and indeed, this was applied to the medicines at the centre of this study. In her response to the Davis et al(2017) article, NICE Associate Director, Linda Landells, confirms NICE's role in assuring that "only cancer treatments that are clinically and cost effective become available in the NHS"[vi].
Without the ability to use surrogate endpoints and early access regulatory schemes, new treatment options for cancer patients would be fewer and arrive more slowly. The Institute for Cancer Research raised this concern in their review of the Davis et al (2017) article. "The ICR and many other organisations have been arguing that we need to be flexible in the way drugs are assessed, in order to speed up the approval process and get the best new treatments to patients more quickly."[vii] The imperative is continually to improve the measures we use, the methods we have to collect them and the research designs to provide the evidence for regulators and health technology assessors to ascertain the value of a medicine. This assessment occurs over the medicine's lifecycle, recognising that the value of a medicine continues to evolve and expand as new indications for treatment are developed. Research has shown that seven out of 10 cancer medicines approved by the EMA between 2003 and 2005 had additional approved value expansions following their initial indication approval, including for additional types of cancer.[viii]
Davis et al (2017) drawn important attention to the measures, and specifically the lack of progress in the use and development of quality of life measures. ABPI and the Academy of Medical Sciences (AMS) held a symposium in July 2017 on precisely these challenges for endpoints in clinical research for cancer.[ix] Participants debated the development and role for endpoints in clinical design and the need for greater progress in designing and using quality of life measures. Although the discussion identified many areas for further development, the need for an understanding of what constitutes a meaningful endpoint, beyond overall survival was broadly supported.
Taking the long view, cancer survival is improving and has doubled in the last 40 years in the UK, thanks not only to cancer medicines but also to diagnosis, screening and other treatments (chemotherapy, radiotherapy, new surgical treatments).[x] Incremental improvements have been part of that history of progress, and have an important role to play in the future alongside the breakthroughs. Regulatory science and clinical research design also have a critical role to play, particularly as new technologies in cancer care emerge and need new mechanisms for measurement and evaluation. However, most importantly, we need to progress the dialogue on these new technologies and approaches for assessment more fully with all key stakeholders, and particularly with patients. By doing so, we can perhaps find greater agreement on how to establish and evaluate the evidence of a medicine.
[i] Davis et al (2017) Availability of evidence of benefits on overall survival and quality of life of cancer drugs approved by European Medicines Agency: retrospective cohort study of drug approvals 2009-13, BMJ 2017; 359:j4530, http://dx.doi.org/10.1136/bmj.j4530
[ii] Kim et al (2015) Cancer Drugs Approved on the Basis of a Surrogate End Point and Subsequent Overall Survival: An Analysis of 5 Years of US Food and Drug Administration Approvals. JAMA Intern Med 2015; 175:1992-4. Doi:10.10001/jamainternmed.2015.5868
[iii] Pignatti, Francisco (2017) Rapid Response to Davis et al (2017), BMJ, published online 12 October. "Importantly, if a sufficient effect is observed based on this endpoint [progression-free survival], it is generally considered to reflect an intrinsic clinical benefit and is not a "surrogate" for survival requiring subsequent confirmation, as Davis et al. seem to imply."
[vi] Landells, Linda ( 2017) Rapid Response to Davis et al (2017), BMJ, published online 9 October.
[vii] Dick, Sam (2017) Evidence vs. access – addressing the balance in assessing new cancer drugs, Science Talk: the ICR Blog, posted 11 October; accessed 13 October. https://www.icr.ac.uk/blogs/science-talk-the-icr-blog/page-details/evidence-vs-access-addressing-the-balance-in-assessing-new-cancer-drugs?utm_content=buffer23897&utm_medium=social&utm_source=twitter.com&utm_campaign=buffer
[viii] Rejon-Parilla, JC et al (2014) The expanding footprint of oncology treatments. Office of Health Economics, May.
[ix] https://acmedsci.ac.uk/more/events/endpoints-in-clinical-research The report of the meeting is forthcoming.