Why do cancer drugs get such an easy ride?BMJ 2015; 350 doi: https://doi.org/10.1136/bmj.h2068 (Published 23 April 2015) Cite this as: BMJ 2015;350:h2068
All rapid responses
It is salutary to be reminded that the 71 drugs for solid tumours approved by the FDA over the last 12 years prolong life by a median of only 2.1 months. The performances of individual drugs within this population will follow a distribution: we are not told the performances of those which appear in the upper 25th or 10th centile but these drugs may be judged worthwhile, once properly attested, by the medical profession or society as a whole.
As a physician, I also used to discuss treatment in terms of median response or survival. I am now retired and suffering cancer for the third time myself. In each encounter I have scrutinised survival curves, and the effects of proposed treatments on them where known. I have noted that the usual exponential decline in survival has levelled off in each case at a point above zero, whence the slope levels and may approach that in an age-matched population. In other words, there are always some long term survivors. And I ask myself “why should I not be in that group? Someone has to be”.
For my stage 3 adrenal cancer, for example I noted that the median survival (with and without surgery and adjuvant mitotane) was nine months, but that around 15% of the study population were still alive when published observation ceased at five years. That was in 2008. In 2012 I was treated for stage 2a gastric cancer. It is now clear that my progression-free survival from gastric cancer since then has proved well below median: I have relapsed with a large single metastasis and am more optimistic than the trials of palliative chemotherapy would support.
Of course the numbers still under observation at the end of a trial are always, inevitably, very small and the confidence intervals around them almost uselessly large. This justifies a plea for longer follow-up and routine second publication. The point I really wish to make is that treatments may offer similar median, but quite different upper 10th or 5th centiles, for prolongation of survival. Maybe their approval should be judged with both statistics in mind, and the patient offered an appropriately balanced chance to be optimistic.
Competing interests: I am the patient (as is clear)
please let me assure you- cancer is very much the evil and invincible predator you describe.
As I got to experience first-hand when I saw my husband die extremely painfully from Melanoma, only 15 days after his 37th birthday. I will never forget the relief when the tumours finally distroyed the nerves in his pelvis- the paralysis was nothing compared to the fact that we could reduce the morphines to a a degree that he was awake and we got a few more precious hours to talk before he slipped into a coma never to wake again.
You could also ask my youngest daughter- she was 4 at the time- what death from a combined liver/ kidney failure looks like, her recollection of seeing her father die is very accurate. Or you could ask my older daughter- although I am not sure she remembers that she climbed on to her father's bed just after he died to tightly hug him- we hadn't been able to do that for the last weeks as any touch had become too painful.
Before his death, my husband had taken part not in one but two clinical trials, plus an early access program- something most fervent promoters of the hardest lines of evidence will never be able to claim about themselves.
The first trial he participated in would probably have been very much to your liking and perceived high quality standards, considering it was randomized, although not blinded. Yet, undergoing the randomisation for this trial qualifies as one of the worst experiences of my life- just up there with him receiving a Stage IV Melanoma diagnosis without any prior warning and then, his death.
At medical school, I had been infected by the virus of evidence-based medicine- I remember very clearly sitting in the lecture theatre (lower part on the right) and the relief that there was a way to actually protect my future patients from my own lack of experience- one of the deep-seated fears any medical professional I know has, especially at the beginning of their career.
I then went on to a lab-based training in Molecular Biology, with serious emphasis on scientific rigour and the need for appropriate controls- I thought I knew what good science was about.
Now, on the other side, being subjected to the evidence-generating part of this brave new world of evidence-based medicine, my husband all of a sudden risked being randomized to a clearly inferior control arm.
This is what one of our patients said in a very similar situation: see screenshot
(you can watch her story here https://www.youtube.com/watch?v=H03vz24JhgM, including her oncologist's reaction when she decided to make use of her right to withdraw from the trial. He would only refer her to the next center if she accepted any treatment she was randomised to)
Melanoma is a very specific case. After years of no therapeutic progress, we all of a sudden have not one but two classes of therapies available. As based on insight into the underlying molecular mechanisms, these drugs present considerable advances over the previously existing 'gold'-standard of chemotherapy, DTIC- rather to be considered a standard of desperation and medical actionism but definitely not golden. (At the same time, I learned about evidence-based medicine, I also learned the Melanoma was both resistant to chemo and radiotherapy, so this is hardly news).
Yet, every single Melanoma drug we have had in the past few years has been tested in a clinical trial against DTIC- the last example that led to a positive recommendation by the European Medicines Agency only last week can be found here: http://www.ema.europa.eu/ema/index.jsp?curl=pages/news_and_events/news/2...
'After 12 months, the survival rate of patients treated with Opdivo was much higher than the rate of patients who received dacarbazine (73% compared with 42%).'
the motivation for randomized controlled trials (RCTs) was to deliver clinically relevant information and to provide insight in situations where there was clear uncertainty, not to produce statistics for statistics' sake.
A Phase 3 study delivering a result where one arm has nearly double (1.7 to be precise) the survival than the other arm allows for 2 possible conclusions:
1. The enormous difference was not picked up in the previous Phase 2 study. Which raises serious doubts about the professional capabilities of the parties involved.
2. The difference was clear when the trial started. Which makes this an equipoise-violation from day 1, violating the principles clearly stated in the Helsinki declaration http://www.wma.net/en/30publications/10policies/b3/- among them the binding declaration of physicians 'The health of my patient will be my first consideration'.
As a patient, neither choice is acceptable.
So the argumentation of this article falls in my opinion short on several levels-
1. Failing to recognize the diversity between the development in different cancers. Incremental progress in the past in one cancer should not kill patients in the process of evidence-generation in a different cancer at present. Tools need to evolve with scientific progress.
2. Mistaking statistical significance for clinical relevance. From a clinical point of view, it is irrelevant whether a new drug is 71%, 72.5% or even 75%- in the Melanoma example above, it is 73.8%- better than the previous standard as a patient will be getting the new therapy anyway.
3. Failing to recognize the limitations of and alternatives to randomized clinical trials. RCTs are designed with optimal comparability between groups (and favourable outcomes for the sponsor) in mind, not with representativeness with respect to the disease. In Melanoma, this concretely meant that patients with brain metastases were systematically excluded from clinical trials for a very long time. In a disease where cerebral involvement is the major driver of mortality and 80% of our patients have metastasis to the brain at death, this raises serious doubts about the clinical relevance of such a design.
On the other hand, alternative designs are available but not widely used- Bayesian trial designs e.g. allow to minimise the risk to expose patients to inferior treatments but are statistically more demanding. Having said that, certainly a patient's life is worth the effort?!
Increasingly, HTA bodies and payers demand data on effectiveness to see how a new drug performs in the real world, not only in an idealised/ inappropriately optimised setting. Currently, systematic data gathering after the approval of a drug focuses on the safety aspects of a new medication, neglecting the potential benefits- the faster we generate evidence into how and when to use new medication, the better for all of us.
4. An unhealthy amalgam of evidence and price. If new cancer drugs were available at a quarter of the current price, this discussion would in all likelihood be mute. While even CEOs of large pharmaceutical cooperations recognise that the current price level is unsustainable- to quote Jimenez, Novartis CEO, from his interview with Forbes:
'What you know is not going to happen is the ability to stack therapies on top of each other at the current price and expect people to pay,” he says. “ The whole oncology pricing structure needs to be rethought because it’s reached the level that is not going to be sustainable for the long term.'
health care systems now use evidence- or rather the lack thereof- in order to negotiate for better prices. The call for more evidence in justification for outrageous drug prices provides a short-term negotiation advantage but then paradoxically increases the costs of the development even further, contributing to the very effect it was intended to control.
While the current system is clearly financially unsustainable, we have no societal contract in place stipulating that patients should die in order for the health care system to negotiate for better drug prices. Effectively, this means that with the diagnosis, a patients looses parts of his or her right to live.
The right to live is one of our fundamental human rights- which effectively turns a trial like the one mentioned above into a human rights' violation. Not to mention that the costs for drugs amount to around 10% of the healthcare spending in developed countries and one could argue that healthcare systems are conflicted as e.g. investments to reduce administrative costs or diagnostic redundancy are harder to implement and do not deliver the same short-term positive budget impact as a saving in drug spending.
5. Logically following on to that and to preempt the point that patients are not forced to take part in clinical trials. In a disease like Melanoma, the most promising new drugs offering the best chance of survival are still in clinical development, the standard of care for stage IV Melanoma is therefore still clinical trial participation. Arguing that a patient actually has a 'choice' to rather die on the standard of care instead of entering a clinical trial with a potentially life-saving drug is cynical. In many diseases, clinical trials ARE treatment and for our patients, clinical trials are part of their treatment strategy.
6. Inaccurate temporal perception. Patients need to take treatment decision at a single point in time, so the perception that 'we just need to accumulate more evidence and then treat the patient in an optimal fashion' is incorrect- because that very patient will in all likelihood be dead or past the ideal time point of treatment. This process is also too slow, especially in fast-progressing diseases like Melanoma where the historic survival after a Stage IV diagnosis was 6- 9 months, failing not only today's but also tomorrow's patients.
So instead of letting patients 'die in safety', we need to invest more efforts into processes like adaptive licensing which grant earlier access to a wider patient population while systematically gathering evidence- this will deliver better data faster and hopefully also in a more cost-effective way.
It's time for the 21st century
And last but not least- are processes like randomisation, blinding and placebo-use actually still considered ethical in the 21st century? Or are they rather the remnants of the paternalistic concepts of clinical medicine of the last?
In the end, it is up to the individual patient to decide.
As the Helsinki declarationhttp://www.wma.net/en/30publications/10policies/b3/ clearly states in Article 8:
'While the primary purpose of medical research is to generate new knowledge, this goal can never take precedence over the rights and interests of individual research subjects.'
Patients are full members of our societies, tax-payers, voters and contributors to our healthcare systems. And they are increasingly insisting on their rights- or directly taking it into their own hands.
We need a broader discussion in society how to provide sustainable healthcare in the future for all of us and high-level strategies instead of short-term fixes for innovation- and accountability for the losses of lives caused by the current system.
Above all, we need ways of generating high quality, clinically relevant evidence in a cost-effective way that does not kill our patients in the process because the ultimate question remains:
Would you want to be the patient on one of these trials?
Melanoma Patient Network Europe
The comment is also available on
Declaration of interests:
Bettina has lost her husband to Melanoma in 2012 and since then, several friends. All of them died a horrendous death and all of them had participated in clinical trials.
Currently, she holds a researcher position in Evolutionary Biology at Uppsala University.
She founded the Melanoma Patient Network Europe in 2013 and her special interest is patient centric clinical trial design. In that capacity, she interacts with different stakeholders- the pharmaceutical industry, research societies, EMA and HTA bodies.
Currently, the basic costs of the Melanoma Patient Network are covered by Bettina's honoraries for advice/ consultancy/ talks; currently she does not receive a personal salary nor any remuneration for her work at MPNE.
Conferences and meetings are predominantly covered by unrestricted grants by the pharmaceutical industry, balanced between companies with competing interests in Melanoma, and without any influence on concept, faculty or other. Project work with e.g. EMA or GetReal was covered by the respective project partner.
More information on MPNE's activities can be found on the MPNE website http://www.melanomapatientnetworkeu.org.
Competing interests: No competing interests