Thoughts on AbbVie’s Rova-T Implosion
blogs.sciencemag.org
blogs.sciencemag.org
To deny a terminally ill patient a drug that may save her life and is unlikely to cause further harm is the height of regulatory hubris.
Now it turns out that not only does it not work, it also actively made things worse. So your "terminally ill patient" and the "drug that may save her life" turns out not to be "unlikely to cause further harm".
The existing regulatory regimes allow for the patient to end up in a trial for a drug like this, but the "right to try" law now means the subject can be 1) charged for it and 2) not provide useful information for other patients because they're no longer in a blinded trial environment.
Current 'compassionate use' exceptions are even worse - only 1,200 of those are granted each year to the lucky few, out of hundreds of thousands of people who could use them.
Also, don't you think it's rather hypocritical to imply that being in a trial (if you win that lottery) is safe, while getting the drugs unfer RTT is not?
Yes, there will no doubt be failures, but to deny a terminally ill patient the right to choose what to do with their own body seems rather crazy to me.
Terminally ill patients and their families are very vulnerable and clinical data is difficult to parse even for doctors, and their are many people who would prey upon them if given the chance
Right to try laws allow companies to monetise desperate terminally ill patients. That's been shown repeatedly throughout history (and these days in the 'supplements' industry) to result in abuse of dying patients. And you get things like https://www.statnews.com/2017/11/21/herpes-vaccine-hotel-roo... which was clearly a great plan - and yeah that's not "terminal", but the exact same kind of thing can happen for terminal diseases as well.
The trial methodology removes the financial incentive, so that now you can't just take all of a dying persons money.
By keeping the treatment in the trial model you also mean that we get data that can actually help -- maybe find out that the drug does not in fact work, or maybe finding out that it does! A patient who is not being treated under trial conditions is much less useful as a data point (if at all) because the placebo effect, and a bunch of other related ones, have repeatedly been found to lead to incorrect feedback - the person says that the drug makes them more comfortable, even though they didn't live any longer, is that because it actually made them feel better? Or was it placebo? Or was it a desire to be "helpful"? If you use them as a data point you may end up with a drug that doesn't actually help, and that drug gets sold. for money. to people who it doesn't help. Cool.
Thus, it is completely unreasonable to say that because occasionally there could be drugs that are harmful, patients should be denied right to try them, especially if their only alternative is literally dying. It's like saying because medical care could "make things worse", we should ban doctors and medicine. Nobody is insane enough to propose this, but apparently there is a lot of people who are ready to say "because there's one drug that proved harmful we should let people die but not allow them to try other drugs that might be somewhat risky". Equally insane, IMO.
.
Also "safety" is relative to the clinical benefit. The CAR-T drugs recently approved for lymphomas have horrible, sometimes fatal side effects. But they also save lives that would otherwise be lost. The risk / benefit would be very different in things like diabetes.
Also, safety is a moving target. Late stage studies are large and expensive bc a lot of safety events are rare. So early studies dont give you a full picture of safety
People forget that FDA exists in the first place to 1) protect people from dying from shoddily studied drugs (ie incompetence) and 2) snake oil salesmen (malintent). While in theory less regulation could enable more drugs to get to patients faster, more likely it would just incentivize more snake oil sales and poor science
https://www.cancer.gov/news-events/press-releases/2018/immun...
https://clinicaltrials.gov/ct2/show/NCT01174121
https://www.forbes.com/sites/victoriaforster/2018/06/04/new-...
https://www.cbsnews.com/news/immunotherapy-treatment-saves-l...
[1] https://www.fda.gov/NewsEvents/PublicHealthFocus/ExpandedAcc...
With sufficient guidance and information, when available, I'm simply saying that informed patients may opt for highly speculative treatment options, on a case-by-case basis.
I can't speak to your question directly, will check it out...
Fundamentally the FDA is very against right to try because they believe, correctly, that it reduces the incentive for patients to participate in medical trials - why risk being in the control group if you don't have to? My (perhaps flawed) understanding is that the FDA doesn't not approve Compassionate Use very often for this reason, although they probably use safety as the official excuse.
There are also ethics guidelines that deal with control group selection. If an effective approved drug exists, usually that has to be the control arm, not a placebo. So a patient will get a drug that works in the control arm, though maybe not in the active arm
The context for "right to try" is a person with a terminal condition that does not respond to any conventional therapies. In essence, the choice is between experimental treatments and hospice.
> If an effective approved drug exists, usually that has to be the control arm, not a placebo.
There will always be a placebo in a stage 3 trial.
FDA guidelines are you need "at least two adequate and well-controlled studies, each convincing on its own, to establish effectiveness" [1]. Well controlled can be placebo controlled or active control
[0]: https://www.prnewswire.com/news-releases/upadacitinib-monoth...
[1]: https://www.fda.gov/downloads/Drugs/GuidanceComplianceRegula...
There is precedent for this in the world of molecular diagnostics. Certain tests can be marketed without FDA approval. However few companies are successful because insurers wont pay and patients wont pay enough
Isn't that known, and just something terminal patients need to balance with their doctors? If I had 6 months to live, I would probably roll the dice for a cure even if 50% of the time the drug would actually only give me 3 months to live.
I doubt anyone believes that trying experimental drugs can only help, and never hurt the patient. If that was the case we'd have been allowing experimental drug usage for a long time.
If I've got six months to live, would I take a 50/50 gamble on living 20 years, where the downside is that I might only live 3 months? Probably - the expectation value of that gamble is highly positive.
On the other hand, if you only have 6 months to live, 3 months is really a lot to lose...
But I agree. This isn't a valid argument against Right to Try.
For the first person to try the drug, it's a big gamble, but as each additional "right to try" participant takes the drug, everyone else gains knowledge as to its effectiveness.
Is this as good as stage 3 trials at assuring safety? No.
But you have to count all the people who could have been helped but weren't because of delays against a no right to try policy.
The plural of anecdote is not data. We have come to understand what it takes to run a good study for effectiveness of a drug: double-blinded, randomized, high-power control studies, with further guards against p-hacking (did you fail a phase III but find a subgroup that was responsive? great, do another phase III on that subgroup; turns out it's mostly spurious). The data that comes from right-to-try is mostly not useful because it's not going to tell you how much of the benefit comes from things other than the drug (such as the placebo effect).
This is ignoring 50% being (i would say optimistic) for stage 3 trials, but right to try kicks in before stage 3 trials so suddenly you're dealing with stage 2 trials: and iirc only like 10% of candidates make it from stage 2 to stage 3, so now you've got maybe a 5% chance of it helping, and a much greater chance of it making your situation worse. And you're still not providing useful data that might help other people.
It also ignores that there are typically multiple drugs with the same or similar targets running trials at the same time: which do you take given currently there's no evidence of which (or any) is effective, no data on how dangerous any (or all) of them are.
> I think Right to Try laws are largely useless, and have the potential to cause actual harm disproportionate to the good they might do. I think that there are sincere, principled advocates for them, but at the same time, there are many people on that side of the issue who are either ill-informed about medical research or who are simply grandstanding for votes and popular support.
> In short, my biggest reasons for thinking this way are: (1) most investigational drugs don’t work. The failure rate in human clinical trials is around 90 per cent, and it’s been that way for a long time. (2) there are disease areas where a tragedy-of-the-commons situation will mean that controlled clinical trials (which are still the only way to determine if a new drug works) will be much more difficult to run under a “why wait for proof?” system (I have to note that the current bill restricts right-to-try to those patients who are ineligible for clinical trials, which is a good provision). (3) opening up unproven therapies could lead to a moral hazard, where it becomes lucrative to put useless-but-harmless “drug candidates” into “clinical trials” so you can charge desperate patients at “cost”. I mean every quotation mark in that last sentence. I am not the only person who thinks this way. My own suggestion for the last point has already been made: antioxidant phytonutrients – i.e., dried grape juice pills – for Alzheimer’s. Won’t do a thing, but hey, you have the right to try ’em.
> There is empirical evidence for all of these positions. Many states have passed their own right-to-try laws (in various forms), and if there has been any decrease in human suffering or advance in medical knowledge through them, I have missed it. To get specific, when Colorado passed its own law in 2014, one of the first companies that announced plans to offer treatments under its provisions was a stem-cell company, Neuralstem. And guess what? None of their treatments ever worked. Most things don’t.
Note that Lowe talked about harm which is disproportionate to the good, not the un-nuanced statement that that there may be harm and therefore bad, as you summarized it.
His objections are good, but this mindset will probably have to go in order to make headway on more efficient medicine. It would also alleviate the concern in #3, because once you give up that mindset, you actually have tools besides the (expensive) controlled clinical trial that allow you to make statements like "dried grape juice pills won't do a thing for Alzheimer's".
The pharmaceutical industry would pay billions for a cheaper or more effective replacement.
The other method I had in mind is to put more focus on basic theory / models. Models in the sense of individual elements that can take input and produce output, not aggregated statistical models. These models can inform (and be informed by) later data. They're pretty useful with and without controlled experiments, e.g. either particle physics or astronomy. When applied to drugs, a functional model of how a drug would interact with the organism will tell you whether a drug works or not -- clinical trials and observation both can help refine the model as its outputs match up against reality. Of course good models of biology are hard, many mechanisms are poorly understood, but you need more focus at this stage before you even consider getting to the clinical trial. Unfortunately less effort seems to be spent on the modeling, more effort on getting quick iterative feedback via petri dishes then animal studies, and then disappointment when something fails in expensive human trials. With a lot more focus on the modeling side, with more willingness to use the data from experiments that aren't "gold standard" controlled clinical trials and to use tools like causal analysis, is it unreasonable to imagine that a lot less money could be spent for a lot sounder results?
I'm not sure we can conclude such broad things about what the pharmaceutical industry wants or would and would not do. It's a mess of a beast. With the side effect of astronomical costs that results from this mentality of thinking controlled clinical studies are always necessary and the only way to know for sure, and more directly that you need them for every new drug to get to market because of regulations, is it any wonder that most new drugs are "me-too" drugs that companies "researched" from existing drugs to be just different enough to earn a new patent and get past the clinical trials that its predecessors did so they can sell it and make a profit? (Rent-Seeking and Redundancy, http://levine.sscnet.ucla.edu/papers/imbookfinal09.pdf)
Some of those studies are already done. Viagra's effect on erectile dysfunction was based on collecting data about side-effects when evaluating its use as a treatment for angina.
But I don't see how they can be applied to most drug tests.
For example, in the news recently we learned of a new treatment which lead to a "complete response in breast cancer patient." https://www.eurekalert.org/pub_releases/2018-06/nci-nat06011...
This requires an intervention and special personalized analysis of DNA and RNA from both a tumor and her normal tissue.
There won't have been any observational data which lead to Taxol because it's not something that cancer patients would normally inject into themselves, outside of a clinical study.
Most new drugs require complex and expensive synthesis, many may require special treatment methods or schedules, and most are for uncommon diseases. I don't see how a observational data from a non-clinical study can be effective.
Similarly, radiation therapy, chemotherapy, and other such treatments are not normally found outside of clinical studies, and I don't see how it could be otherwise.
As for basic theory / models, I follow the blog "In the Pipeline", by Derek Lowe, a pharmaceutical chemist. Links to his blog come up not infrequently here on HN. I get the very strong impression that models are used all the time, that model building is an important part of pharmaceutical chemistry, and the reason for iterative testing in assays, cell cultures, and the like is because the mathematical models aren't that good, even after decades of attempts.
A DDG search for 'Boldrin' on that blog shows several articles which reference "Against Intellectual Monopoly", including http://blogs.sciencemag.org/pipeline/archives/2006/04/02/dow... . I am too ill-informed to be able to provide any commentary on my own.
As I understand it, most of the "me-too" drugs are because multiple drug companies are working with relatively few drug targets. If company X comes out with a drug, and company Y comes out with a 'me-too' drug two years later, most likely what happened was that companies X and Y started looking for a drug some 15-20 years ago against the same target, and found similar compounds. Then when X got its patent, Y had to figure out some way to extract profit from its decade of work, which would otherwise go down the drain.
It's not so much that Y "researched" X's existing drug to make a me-too compound.
In addition, just because some drug appears to be a "me-too" drug, that doesn't make it useless. People's bodies are different. Quoting from a comment at http://blogs.sciencemag.org/pipeline/archives/2004/08/31/me_... , "Some people may not respond to Zocor but may respond beautifully to Lipitor. Or they may experience less muscle weakness under Crestor." See also http://blogs.sciencemag.org/pipeline/archives/2004/02/29/mor... .
This ties back to the Right To Try stuff though. Suppose there's a new drug researchers have reason to believe might work, maybe it's even synthesized with an individual in mind, so they administer it to a patient after getting consent and (more importantly in the right-to-try discussion) legal approval. The patient either gets better, gets worse, has no change, or develops some new problems. Whatever the case, the data are recorded, it's a study of some sort. I might tend to see this as an observational study but I should probably use the more precise jargon and follow the split of only no-influence being observational and any-influence being "experimental", so maybe the best category is "field trial". Still, it's definitely not a (randomized double-blind gold standard) controlled clinical trial, but can we still learn anything from it? Did the drug work? Did the drug cause any state change? If another drug had been taken would the result have been the same? Causal inference people say you can make (qualified) claims from such data. But it seems that the regulatory mindset at least (which will lag practice, though I don't think practice has changed dramatically enough to influence regulation yet) is "we can't answer these questions at all" and so they require the very expensive (source of the majority of costs for new drugs) controlled clinical trial.
This reminded me of an old thread I keep bookmarked... https://news.ycombinator.com/item?id=3819346 "The entire thrust of our medical regulatory system ... is the belief that it's better for 1000 patients to die of neglect, than 1 from quackery. ... The entire process we call 'drug development' is an attempt to gain six-sigma confidence that we are not practicing quack medicine." My real complaint is with this sort of mindset, which I took Lowe to be spreading with his statement about controlled clinical trials.
I know that trials are not always between placebo/"no-influence" and "any-influence." They are sometimes between current best practice and possible new best practice.
I believe the recent case of the woman treated for breast cancer, which I mentioned previously, is an example of your "suppose" case.
Another might be the first trial of injectable penicillin was on Reserve Constable Albert Alexander. It was clear within 24 hours that it was effective.
So certainly it doesn't require randomized double-blind trials. Then again, Lowe said "controlled clinical trials", which I suspect may include more than the 'gold standard' of clinical trials.
There's an old joke about how there hasn't been a controlled trial about the effectiveness of parachutes. https://www.bmj.com/content/327/7429/1459 . I don't think the medical field as a whole agrees with "the most radical protagonists of evidence based medicine" (quoting from the abstract) to conclude that we therefore can't be certain about its effectiveness.
I read the comment from 2012 that you pointed to. I think it is an opinion with only a tenuous basis in evidence.
Other countries have different requirements for care. I pointed out elsewhere in this thread how people from the US go to Mexico for treatments which are illegal in the US. (Note: Mexico is not full of Nazis as moldbug exaggeratedly suggested must be the case for that to happen.)
If it were so simple as closing the OODA loop, then we would expect amazing treatments to come out of these countries. Or as Lowe pointed out in the quote I gave earlier:
> Many states have passed their own right-to-try laws (in various forms), and if there has been any decrease in human suffering or advance in medical knowledge through them, I have missed it. To get specific, when Colorado passed its own law in 2014, one of the first companies that announced plans to offer treatments under its provisions was a stem-cell company, Neuralstem. And guess what? None of their treatments ever worked. Most things don’t.
Since the lower standard doesn't seem to help, why lower the standard?
What evidence is there that 1000 or more patients die of neglect for every one which dies of quackery?
Derek's big concern with right-to-try is that it's basically a way to advocate for taking the science out of drug approval. Data outside of a well-controlled study is hard to use, but the public will lobby for it to be used. And the concern isn't idle: when the FDA's response to the bill was "we're going to look into how to do this safely," the reply form a key Senator was "no, no, no, you don't understand, we're trying to tell you to get out of the way." The end goal would be to eliminate the FDA as useless bureaucracy, but we already have an example of what unregulated drug industry looks like: it's called the health supplements industry, which is so sleazy that many companies don't even bother to include the active ingredients they claim to include.
People seem to forget that there was a time where the FDA and trials didn't exist. And that the reason the FDA has to exist is because otherwise you get very expensive "drugs" that frequent, at best, do nothing, and oftentimes do harm.
Nobody is doing that - current RTT legislation only applies to treatments that have passed phase I, not drugs that are in or have not yet started phase I trials.
Do you disagree with all individual restrictions? Should we eliminate the concept of prescription drugs, so that anyone can go buy Adderall or Accutane without a doctor's approval?
Yes, and Yes.
I believe all drugs should be legal, and OTC. I own my own body and what I choose to consume is none of anyone business but my own.
The idea that you should have any say at all as to what chemicals I consume is absurd to me.
Do you support say making it illegal for a person to buy their own brakes for their car? Instead forcing people only to have their cars repaired by government certified mechanics?
It seems odd to me that we allow anyone to work on a car, something that could actually endanger others, yet we severally restrict what people can do to themselves
I am 100% for complete and total individual freedom provided that freedom is not directly harming OTHER PEOPLE, I am opposed to any restrictions upon my individual liberty which includes my liberty to harm myself up to and including suicide
We have experience of what sort of quack "medicine" appears when unscrupulous people take advantage of the hopes of the terminally ill for personal gain.
Is that what you want us to return to? Or do you have a way to prevent it?
Here's a real-world example - https://www.livescience.com/4040-dying-desperate-lure-quack-... .
People in the US in the last months of their lives, in desperation, go to Mexico for "treatment" by methods which have never ever worked.
> One hospital we visited offered homeopathic treatments. (Homeopathy is a form of quack medicine essentially based on the mystical principle of "like cures like.")
> Another, respectable-looking hospital offered such "alternative" treatments for cancer as shark cartilage, mega-doses of vitamins, and "prayer therapy." The hospital also offered Laetrile, a notorious cancer treatment discredited by repeated scientific studies.
> ... Donsbach's clinic had a reputation for providing questionable medical procedures, including "ultraviolet blood purification," colonics (a potentially dangerous colon-irrigation therapy), and the use of microwaves to "heat" cancer cells.
This is what you want, everywhere, yes?
These evil practices all work to stop people having access to things as basic as Elixir sulfanilamide[1], Thalidomide[2].
If only there were already some example of the FDA allowing victims of terminal diseases to have access to drugs that might save their lives if there was overwhelming evidence that they would at the very least work? Like imagine if a drug demonstrated an immediate overwhelming success in treatment but had not yet had sufficient trials to determine side effects and safety? HIV victims could certainly have benefited from AZT before the full side effect profiles were known. It's so sad that they didn't get access to it. Oh wait, they did [3].
[1] https://en.wikipedia.org/wiki/Elixir_sulfanilamide - "We have been supplying a legitimate professional demand and not once could have foreseen the unlooked-for results. I do not feel that there was any responsibility on our part."
[2] https://en.wikipedia.org/wiki/Thalidomide#Adverse_effects - It did work for morning sickness, but an evil scientist at the FDA had stopped it from being sold in the US, depriving expectant mothers of helpful relief
[3] https://en.wikipedia.org/wiki/Zidovudine - HIV and AZT actually did result in the laws allowing exemptions for terminal cases. The difference between those, and "right to try" nonsense is that the terminal disease exemptions still require a degree of solid evidence of function, and a demonstration of safety vs alternatives.
The FDA restricts "snake oil" sales by requiring products be marketed based on good data. Remove that requirement and patients will be bombarded by marketers and salespeople peddling exaggerated claims to get at the potentially lucrative right to sell a drug