The US has a pretty strong regimen for getting drugs approved, a system that has great credibility after it kept out the nausea and morning sickness combatting drug thalidomide that in the 1950s wreaked havoc on babies in other countries. The gold standard for new drugs to be allowed to be prescribed is, as I understand it, that they have to show that they do not pose significant risks and also undergo double-blind clinical trials on humans that demonstrate that they work better than a placebo. The process is long and expensive.
But this causes problems when someone is critically ill and there is the prospect of a new drug that may work but the patient may die before the trials are completed. There is a provision known as ‘compassionate use’ that was developed in the 1980s during the height of the AIDS crisis, when retroviral drugs were being developed but not yet approved but AIDS victims were dying after having exhausted their other therapeutic options. The protocol requires approval on a case-by-case basis by both the drug manufacturer and the FDA before physicians can prescribe these drugs as a last resort.
The June 2016 issue of Harper’s magazine (subscription may be required) has an article by Helen Ouyang on this issue. Ouyang says that this compassionate use policy sounds reasonable but is not without its downsides. A bad result such as death for a single patient can slow down or even halt development of the drug. Since the people granted use of these drugs are usually in a very bad way, a negative outcome is a very real possibility even if the drug itself did not cause it. Furthermore, each person accommodated because of compassionate use is one less member available for later clinical trials, and for many drugs the pool of potential test subjects is already small. In addition, these experimental drugs are produced in very small quantities so compassionate use reduces the amount available for clinical trials.
Ouyang writes about a parent of a child who got the adenovirus after a kidney transplant. Adenovirus usually creates mild symptoms but can be deadly for people who have weak immune systems, as is the case after patients receive transplants and are receiving treatments to combat rejection. A new company was developing a drug called brincidofovir to treat the adenovirus. They initially denied the parent’s compassionate use petition for use of the drug. The parent then created a social media petition for the use of the drug that went viral. The company finally agreed under pressure, the drug was administered, and the child recovered.
But later clinical trials of brincidofovir showed no difference when compared with a placebo, and even a higher death rate, which resulted in further trials being halted.
So this poses a problem. Did the child recover independently of the drug or because of it? More generally, does the fact that there was no statistically significant effect mean that the drug never works? That is the kind of question that is so hard to answer because humans are so complex and can respond so differently to the same stimulus and sometimes patients recover all on their own.
Take for example a hypothetical case where patients who suffer from some disease have a 10% chance of a spontaneous remission. Now suppose that there is a new drug being tested for that disease. Because we do not know, and cannot control, all the variables that might influence the potency of a drug, it may turn out that in the early stages just 10% of patients given the drugs in a trial have the other unknown factors that enable that patient to benefit from it. Under the standard protocol for testing, this drug will show no difference from a placebo and it is possible that further research on the drug would be halted. (Even if a drug fails to show benefits for the original disease it was being developed for, other uses might be found. It should be noted that thalidomide is still around and being prescribed under tightly controlled conditions for certain illnesses that exclude pregnant women.)
But if that research could be continued, it may be possible to isolate the factors that resulted in just the 10% benefitting, and future clinical tests that controlled for those variables would produce much better results. This is why research that is done by non-profit agencies is so important because they can further explore such questions that for-profit companies might forego since research and trials are so costly.
Another troubling question is whether such drugs, even if the tests show that they are no better than a placebo, should still be allowed for patients who are so critically ill that they will die without any treatment at all. If such a person wants to try the drug, should they still be allowed to do so and use whatever quantities that remain after production has stopped and the trials are halted?
From the point of view of the patient, the answer would likely be yes, because they think that they have nothing to lose. Furthermore, such use might provide valuable anecdotal information that could aid further research. But on the flip side, one has to be extremely cautious about allowing drugs to be used before their safety has been established..
These questions are not easy to answer and I am pretty sure that medical ethicists grapple with them all the time.
I thought the gold standard these days was not to test against placebo, but against the best known treatment, so as to not bother developing treatments that, while better than nothing, are outperformed by other treatments. I guess this has its own complications.
That doesn’t sound right. If 10% have spontaneous remission, then the effects of the treatment should still show up in 10% of the remaining 90% (or a bit less if receptivity to the treatment were somehow correlated with spontaneous remission). So you would see 10% recover in the placebo group, and 19% recover in the treatment group. The only way that the treatment would show no additional benefit is if the treatment-receptivity were perfectly correlated with spontaneous remission.
It’s not just about the ethics of drug testing in other countries or standards. Drug testing done in countries which have standards as high as the US (e.g. Canada, the UK) are rejected in the US -- usually, to protect drug companies and their copyrights. If tests were by other countries, the copyrights would run out sooner and the public would benefit from generics instead of lining corporate pockets.
Unfortunately, a drug testing scandal in 2014 showed that US drug companies were testing in Canada, offering money to doctors in exchanges for “positive results”, even when they were negative.
Ethics is not an absolute, and the precautionary principle is predicated on particular ethical premises.
Arguably, in the long term, it may better to test stuff whether or not it has a significant chance of success — either way, information is gained.
@1
Testing against other drugs is good, but not necessary. There are a number of drugs that have come to the market that are less effective than already available ones, but may have a different side effect profile, drug interaction profile, may just work for some people where the original drug doesn’t. So long as a drug proves that it is safe and effective it can be released, the rest comes down to prescribing criteria that doctors consider, not if it should be sold.
Not sure it is wise to trust the US for oversight. Bayer sold HIV and Hep C infected products to the United States in the 1980s. After some 10,000 were infected, they recalled the products and resold them in Asian and Latin American countries.
http://wearechange.org/bayer-us-government-gave-hiv-children/