Then vast majority of experimental treatments are worse than doing nothing. So, it's generally presented as your going to die either way would you like to help find a cure so the situation becomes less hopeless at some point in the future. However, in cases where there is a large demonstrated benifit they often end the contol and give everyone the treatment. Because extending lifespan by 10% is progress but you need a control the demonstrate it's actually better than random chance but a 10+x increase in lifespan needs no control.

PS: Wit is a great movie that deals with terminal illness from the patents perspective. If I recall correctly there is a vary minor subplot around a trial on the proper drug dosage. At no point is the idea that this might be a cure just possibility of slightly extending your lifespan at the cost of a lot of pain. It's basic cost/benifit analysis but taking higher doses is in no way presented as better.

disclaimer, I work on cancer drugs so I can shed some light on your question.

> a very high percentage of people would opt for a potentially fatal, completely untested course of action as opposed to imminent death. So who gets to try these treatments, who tells dying patients they are not allowed them, and is there a black market or large amounts of money changing hands for experimental procedures?

First, for many cancers you do not die of an imminent death. Sometimes it can take 5 to 10 years to die from Cancer.Unless you are affected by a cancer where no treatent is available, there are usually multiple courses of treatments you can follow before being out of options.So patients are actually trying existing methods before jumping to new treatments. New treatments are only available in small quantities and there are strict criteria for the selection of patients. And in Oncology actually we already give the chance for patients to try new drugs as soon as the safety evaluation begins, whereas for other therapies we go through healthy volunteers first. For a patient there is not much incentive to join Ph1 since the efficacy is not proven and there is a dose escalation method so you may not get the efficacy dose even if you enter the trial. Besides, there are safety risks involved and some patients die in Ph1. So every patient signs an informed consent that they are aware of the risks involved.

As for the hundred of thousands of patients, the companies cannot support them for many reasons: as mentioned, the drugs are only manufactured in small quantities in early stages and it takes years to scale up the process. Second, clinical trials cost a lot of money, and the more you have patients the more it costs. So you dont recruit more patients than what you need to do your evaluation. This being said, most pharma companies continue providing drug even after the trial is completed to patients who see efficacy with it. Third, the larger the trial, the longer the enrollment time and the longer it takes to get the results back and therefore the longer it takes before the drugs reach the market.

> Who tells dying patients they are not allowed them, and is there a black market or large amounts of money changing hands for experimental procedures?

The FDA regulates every aspect of clinical trial design. Dying cancer patients are always the first to receive untested drugs -- the FDA has set the bar that any new therapy must outperform what is currently used. There is no "black market" for experimental procedures, mostly because you must visit a hospital to receive them in the first place. Some drugs are intravenous; all are given with standard oncology examination. A doctor cannot write a prescription for a drug not on the market yet for you to pick up at the pharmacy!

> Surely with hundreds of thousands of desperate, dying, last chance sufferers, it is better to go to extreme measures and offer the most promising yet dangerous treatments to everyone. Is it simply a side effect of the way pharmaceutical companies have to do business?

In order for the testing of a Phase I trial to be complete, patients must meet certain criteria. The FDA (and, by extension, the pharmaceutical companies who want FDA approval) require enrolled patients in a clinical trial to have similar levels of disease. A patient that is clearly near end of life may not be able to physically swallow a pill...their outcome may bias the trial's results.

Yes, there are desperate and dying cancer patients, but a Phase I trial can (and should!) only accommodate dozens of them to achieve proper statistical power. Phase II may take hundreds, and Phase III will take even more -- but when a drug is approved, it has the possibility of being given to many thousands or millions of patients all at once.

As the trial population grows, medicine rears its stochastic side -- despite patients having pathologically similar disease, some will be helped and others will not. This is at the core of all FDA trial design: big populations are necessary to see just how effective your drug is. It should not be surprising, then, that the majority of failures happen at the bigger scales (PhII and PhIII).

When the treatment puts a large number of terminal patients into remission, there is no statistical reason for a placebo control. You know what percentage of such patients will die when offered a potentially revolutionary therapy that does nothing, because there are previous therapies that didn't work. If one third of your patients go into remission, one third show improvement and the treatment has no effect on one third, you know the treatment works better than a placebo.

As for offering this to desperate, last chance sufferers, unfortunately the history of potentially revolutionary cancer treatments shows this is a bad idea. If the treatment kills 90% of people within hours or days of treatment (sadly this is a realistic outcome before dosing is worked out and in some cases, regardless) then it is morally unacceptable to roll it out to thousands of people. You have to do the trials first to establish whether it works, is safe, that it doesn't cause unnecessary suffering and hasten death and what the optimal dose and protocol is. You also have to figure out how to deal with patients who do have an adverse reaction to the treatment.

What if the side effect of the untested treatment leads to everyone in the hospital becoming infected or dying?

Well in this case it's obviously different. I'm talking about actions that have consequences for the dying individual only.

They injected a cancer patient with 100 billion units of the measles virus, this is new stuff. It may have unintended consequences.

I recommend to watch Dallas Buyers Club, it addresses the very point you make.

One could argue, how could the trials NOT be morally randomised? Are we doing a disservice by not conducting the trials in a way to study the efficacy of certain drugs.

It should be noted that one of the components to many drug and observational trials is the ability to halt the trial and switch patients on or off drugs as it becomes clear that one treatment is vastly superior.

The original studies of aspirin and heart disease were, for example, halted because the benefits were so clear that it wasn't considered ethical to keep the placebo arm untreated.

Several HIV trials have similarly been halted because the treatment was showing no or negative effects.

A third problem which hasn't been discussed: If a new treatment is hapharazdly applied and a lot of patients die from it, funding will dry up and continued research will stop. Not to mention that the US legal system would be a problem for the doctors involved. So the researchers who are at the front of developmenst like this are really cautious, which further slows the process.

It's valid morally. However, science runs on grant money and grant money will usually only be given to projects that show a proper scientific method. It would be nice if there were some sort of open market, where terminally ill patients could opt in to clinical trials. But these trials must be tightly controlled, or their results are meaningless.

I would argue that as a terminal patient, I certainly would not want to be part of the control group taking a placebo or no treatment at all, and that the scientific method could still be applied later when you still have a lot of people that took the treatment compared to a lot of people who didn't, for one reason or another. I imagine scientific grants could be funded on an individual basis, if the individual is rich enough.

Usualy placebo arms get less than 50% of the patients anyway. So most people on trials get the drug. And no, you cannot "apply the science later" because you specifically want to see the difference in PFS (progression free survival) and OS (overall survival) on both groups, and for it to be comparable they need to start with the same baseline.

But we're talking about people who have been told "this disease is going to kill you." If a treatment causes even 5 percent of them to be cured, how is that not significant? What if 40 percent have tumors shrink? Remember, any reversal of the disease is a good thing. If you're just measuring a statistical increase in life span (say 3 years treated vs 1 year placebo), I'd say you're not really aiming high enough. If you've got something with real promise you'd know it without a control group. The person in this article is exactly such a case.

It's more that the controlled placebo group will also be controlled for other medications that could compromise the research. If I have cancer A and I'm taking three or four medications + an experimental medication then it's difficult to know which factor affects the results. So the control group will be on the same types of medication in a proper study where as a sort of ad-hoc grouping of patients would be less easy if not impossible to accurately report.

I assumed we'd be comparing to the predicted outcome from no treatment, but I suppose if you wait until Hospice time there's not going to be much success with even the best virus compared to doing nothing. Which raises other questions about just how long and at what cost we try to delay the inevitable.

For many drug trials, the "control" group isn't getting no drugs - they're getting the current standard of care, whatever that is.

It's not just that. If it's not properly and exhaustively tested, then they open themselves up to litigation by people who recover less than 100% or who have a new symptom.

It's also worth noting that there is a constant sussurus of "miracle cure for foo!" that doesn't actually pan out in the long run.

Both very good points. To the first though, I imagine you could sign away your rights to post-treatment legal action beforehand.

The main problem with that is you will have a lot if nefarious groups abusing the situation. Desperate people will sign away their rights without understanding their risks. And even though they may already be dying, they can always suffer more. I fear this would lead to dangerous experimentations.