NCRI 2015
How do we assess a drug's real worth?
Dr Antonio Fojo - National Cancer Institute, Bethesda, USA
As regards the economic impacts of cancer, there are very many and actually it depends on who you are, where you are, where you’re living, what cancer you have, all of that varies. I was going to focus a little bit on the American system, because obviously that’s the one that I am a part of, and point out that we generally think of it, may people in the world think of it, as the ideal system: nearly all drugs get approved, nearly all patients who have insurance have access to them, a lot of therapies are used off-label. It gives patients and physicians great flexibility, so that’s what it looks like on the surface. On the ground it’s a little bit different; it’s not as easy and not as well-greased, shall we say, as it appears from afar. But to me it’s important that what we do as academic oncologists is to make sure we ensure that the highest quality of clinical trials are conducted, that the data is interpreted properly and then we hand it over to everyone to use in the best way that they see fit.
For me personally, I started writing about this about a decade ago and I realised quickly that you didn’t gain so much traction writing about drug costs, it doesn’t go anywhere. You talk about it and then it sort of dies down and then the next thing is the drug is approved and it’s 10% higher than the previous drug. So if you look at the trajectory since I’ve started talking about this it’s been in the wrong direction and everyone who talks about it, not just myself but everyone else, we can confidently conclude that we haven’t had much impact on this because that trajectory is upwards. So I increasingly feel that my task is to focus on what makes a clinical trial good and what makes the outcome a valid and meaningful one. Then in that way you at least can feel confident that what is being approved is of value and then let the chips fall where they may.
What difficulties do clinical trials face?
Clinical trials, from my point of view there are several things. Increasingly, there is a pressure to accept surrogates as endpoints, meaning that if we really consider that the goal is to prolong a patient’s life actually the goal is to cure a patient. But we’ve scaled back from those high aspirations to prolonging a patient’s life. So if you want to prolong a patient’s life obviously the goal is the endpoint to be overall survival. But, heck, those take a long time to come by sometimes and sometimes not so long but still we want the answers sooner than later. So increasingly we’ve tried to use surrogates, progression free survival is one such surrogate.
I think it’s very important that then that be a really valid endpoint, that it be captured properly, that it not be subject to what, increasingly, PFS is subject to and that is just censoring at a level that is, in my opinion, impossible to discern what kind of an impact it’s having on the outcomes. So it’s in that sense that I think that we can be good for omissions is in ensuring that the trials are good. So we’ve had to focus a little bit away from the obvious endpoint, overall survival, about which it’s very difficult to quibble, or so I thought. Until now we get no difference in overall survival and then people will tell you, ‘Well, that’s because A, B, C, D and E.’ So we might have to even quibble about that. But my goal has been to try to understand surrogates and how those might be better captured and more stringently assessed.
Is your purpose to prevent people from cheating?
I really think people don’t cheat so everybody’s enthusiastic about what they’re doing. That doesn’t mean that they’re enthusiastically cheating, no. It’s just that actually things can happen that are unanticipated and that then lead to a result that is really not as valid as it should be. So one of the things that I worry about when I see a lot of censoring and I see a lot of patients who discontinued therapy because of toxicity, it’s impossible to not wonder how did this impact the outcome? And the simplest way of visualising this, what if the patients were more inform with more dense disease, more likely to progress? Going to the ones that suffer most from a therapy and discontinue it sooner, then they’re censored with regards to a PFS endpoint, for example, and that could bring an advantage to that arm. Now the opposite could be true, if toxicity is just capturing everybody, shall we say, or affecting everybody, you could be censoring patients who are otherwise in good health, would respond to the therapy and then it works against you. So there are a lot of variables and in the end the best way would be if there were no censoring. I think when Kaplan and Meier envisioned Kaplan-Meier plots incorporating censoring and how that would be beneficial they never thought 50% of patients would be censored for a given endpoint such as progression free survival. That’s not uncommon these days so it’s that sort of thing that I think we can strive to do better.
Do you think all negative data should be published by law?
Absolutely, absolutely, all negative data should be published. In fact, I would say something that appears in abstract and is never published is, by default, negative. Maybe I’ll put in a plug here, put in a plug when being interviewed by one journal for another, it’s unheard of, right? So I actually work for The Oncologist and we have a section that actually is designed, CTR, results, designed to capture, ‘Go ahead and send us your negative data.’ We’re interested in negative data because everybody needs to know negative data otherwise we’ll all in the community give that drug that may work and we don’t realise that somebody already gave it to twenty people and it didn’t work. But try to publish that in the average journal and it goes nowhere. So it’s important that those be published. Actually, as another sort of plug, I work in rare diseases; it’s difficult to publish in rare diseases because you’re not going to have hundreds of patients in a randomised trial. One has to have a different bar for that sort of trial. You also need to draw different conclusions, you’ve got to be careful that you’re drawing a conclusion about twenty patients. But especially if it’s a negative study you should publish it. So if you have a rare disease and you have twenty patients and it’s negative, oh my goodness, where are you going to publish that?
Do you think there is a culture in the US to keep treating cancer patients, regardless of likely survival?
Yes, I think it’s only a matter of time before other countries will catch up to the United States in that regards. That’s a problem for which everyone shares some blame, such as it is. As a physician, as an oncologist it’s extremely difficult to see a patient dying of cancer and you’ve been trained to treat cancer and not be able to do about it. So as oncologists we sometimes treat at the end of life when we shouldn’t and we know we shouldn’t. For a patient it’s very hard to accept dying of cancer in many cases and that’s understandable. I actually in the abstract point out the fact of something that I hear so often, patients saying, ‘What can be worse than dying?’ To me the answer, as an oncologist who has seen this, is very clear is dying anyway or maybe even dying sooner with a lot of toxicity from a therapy. Unless you have experienced that you haven’t experienced what is worse than dying but you only experience that once. So patients ask to be treated because they think we have something that we can offer them that is valuable. We feel conflicted, we feel that it’s “easier” sometimes to treat than to have that very difficult end of life conversation.
Then I also think patients show up in offices increasingly having read in the lay media about the next greatest therapy that is out there and that’s just got to be controlled. You never see… I shouldn’t say never… you rarely see an article of a patient who went for the greatest and latest and everything went wrong or it had no effect, you see the outlier who had the greatest results. Then every patient thinks he or she will be the outlier and, of course understandably, come looking for that.
So there are a lot of things that are in this mix that make it difficult. If it’s not here in the UK… I think it is. There are undercurrents of that in the UK and in Europe, I see that increasingly. The colleagues that I talk to in Europe say, ‘We’re becoming more and more Americanised in that way.’ Not that Europeans want to be Americanised but we’re becoming more as the Americans have been in that regard.
Do you think the US can learn anything from NICE in the UK?
I’ll tell you, there’s no perfect system in the world but NICE, in my opinion, is as close as it comes to being perfect. If you had a system that was perfect everybody would adopt it but the fact that we have different systems is because of that. It’s also because of societal pressures and so forth. NICE gets a lot of push-back in the UK, I know, but there’s also a lot of people that defend it. It’s a superb system. Will it occasionally make a mistake? Yes it will. And not to trivialise life or the duration of life, if it makes a mistake on something that was marginal, NICE is not about to turn down the cure for pancreatic cancer, it might err on something that has a month or two of survival advantage but it won’t err on clearly transformative, good therapies. So it’s as close to a perfect system as it is and that it can dissociate itself and consider really not just the benefit to the patient but the benefit to society as a whole is laudable in my opinion. So I just hope that it keeps on keeping so that we can eventually emulate some form of it throughout the world. So, as you know, in the United States we don’t have that and the FDA correctly approves drugs irrespective of cost. That’s as it should be but then after that there’s no check and balance for that system.
