Clinical Trial Ethics
In the New York Times Amy Harmon has a long and fascinating article on the ethics of medical experiments on humans. She considers the argument that it is unethical to have a control group which doesn’t get the wonderful new experimental treatment.
I just note that the ethical rule imposed by the FDA and IRBs seems to be
first: do no harm.
second: do no harm.
third: do no harm.
The immense effort and delay required for approval of clinical trials basically forces anyone who doesn’t have tenure already to devote his or her efforts to treating sick mice and rats.
The article discusses two cousins undergoing melanoma cancer treatment, one of whom is in the control group and dying. However, if any reader has malignant melanoma with metasteses but no metasteces to the brain or heart and no history of auto-immune diseases, I personally can arrange participation in a clinical trial without a control group.
I am absolutely 100% serious. Commenters with melanoma are most extremely welcome and eagerly sought.
update: Link added. This post does not discuss the ethics of including control groups in clinical trials. My points are that it is very extremely horribly difficult to get approval for clinical trials, and that this is a moral problem too.
Also, if you contact me soon, I might be able to get you into a phase 1 trial of treatment of melanoma. Phase 1 means no control group. This is an absolutely serious claim. I did not post contact information for the trial, as I don’t want the team burdened with general debate about ethics.
Conducting clincal trials without a control group is dangerous and irresponsible. Humans are used as test subjects without the ability to draw any valid conclusions, and the experimental treatment could well do more harm than good.
From the individual viewpoint though, if I am dying I really don’t care much about the statistical validity of the study, I will take my chances with the experiment since I have nothing to lose.
I agree with you. Controlled trials are necessary. Therefore some people will just have to be in control groups.
I might add that even in phase 1 (a study to make sure the drug isn’t toxic which is required before the controlled study) people are excluded because treating them would confuse the science even though the alternative is certain death. Part of the reason it is very tempting to devote one’s career to treating sick mice is that you don’t have to tell people that you can’t given them one last chance, because the FDA and the IRB won’t let you.
I don’t think this is a case of bad policy or bad people or unethical rules. I think this is a tragedy — something which is horrible but unavoidable, because valid moral rules clash and you can’t both do everything you can for everyone in need and contribute to the great world changing progress of medical science.
yes sorry. I forgot to post the link. But I mean it is the front page of the New York Times
I strongly disagree Larry. I don’t think that most sciences can model results without experiments. I’d say physics can and that’s it. In any case biology plainly isn’t one of those sciences. Ask any biologist.
It is very clear that many pharmaceuticals which, in theory should work don’t work. Also pharmaceuticals have unexpected side effects. It just isn’t true at all that the state of biology is such that controlled trials are not needed to answer questions about safety and effectiveness. In fact, dangerous pharmaceuticals are sometimes approved (Vioxx, Thalidomide except in the USA many others).
Living things are immensely complicated and people are different from other animals.
“Whether or not in vivo studies improve the understanding of in vitro results is questionable.” Oh really ? Can you name anyone with any expertise who asserts that ? Anyone ? I mean any biologist not just anyone who has anything to do with clinical trials.
“My points are that it is very extremely horribly difficult to get approval for clinical trials, and that this is a moral problem too.”
Perhaps this is not entirely a bad thing. There’s often more enthusiasm to experiment on people than competence. On a related point, I think it was Lee Sechrest I heard say, years ago, that if you’re ever randomized into a medical trial, hope for the placebo.
“If the elimination of contol groups potentually would save a few lives now, sometimes, with the cost of more lives later, then is this a reasonable treadoff for them.”
It’s not clear which group of the dead or living is represented by “tradeoff for them.” I suggest that those in the group of the “more lives later” might feel a bit perturbed at anything less than a rigerous experimental design. That is if they can feel anything given their places in the more lives later group.
I’ll add that pharmaceutical trials have a bad enough histroy without compromising the safeguards built into the experimental procedures. Better to find a process that fits within the ethical dimension of human “trials” a nice and inocuous way of saying human experimentation. Maybe it is time to look more closely at the concept of placebo controls in such biological studies. Do we really believe that there is a likelihood of a placebo effect within a study of biological effects? Effects on psychological phenomenon are not a basis for assuming that such effects may or will occur on physiological phenomenon.
Addendum: Clinical trials are aimed at proving the efficacy of an agent. Why not start the series with a focus on the ill effects of agents which are hypothisized to be effective in the treatment of some biological dysfunction. Look for the side effects first?
many trials have outside obsevers who are looking at the unblinded data. their role is to either stop the trial is the drug is hurting people, or occasionaly, to require that all patients start getting the drug if it is showing a large survival benefit and it would not be ethical to continue withholding it from the other arms of the trial.
many trials are not run vs placebo but rather against existing standard of care. this mitigates the ethical issues somewhat.
keep in mind that those getting new drugs are also taking real risks. sure, it might work, but it might not and could cause serious problems as well. they are taking substantial risk in hope of unproven benefit.
most trials fail, so all in all, you are likely safer and better off in the standard of care control group.
that mitigates the ethics somewhat (unless you are using pure placebo).
it cuts a couple of ways though. sure, long arduous and expensive trials may weed out a few dangers, but they also keep pharma companies from even working on small applications.
a disease that affects 10k people a year, even if it is debilitating or fatal, is just not economic to research. you’d never get your $300-500 million in trial costs back.
there are some fast tracks and reduced requirements for “orphan” indications, but it’s still mostly prohibitive to try to bring a drug to market.
so, pharma works on blockbusters and ignores everyhting else and we cure erectile disfunction (with 3 or 4 drugs) but have nothing for a vast number of fatal indications.
that seems to me to be a strong reason to loosen standards a bit for the small indications. otherwise, you’ve effectively legislated the people who have them to death or debilitation.
they do that. that’s phase 1 and then phase 2 dosing ranging.
you cannot look for side effects until you know what dose is medically effective. that’s th point of a dose ranging study. you see how much it takes to be effective and how the side effects increase as dosage increases and then try to find a sweet spot of good clinical efficacy and tolerable side effects.
I’d agree that you can make a reasonable utilitarian argument for reducing evidence requirements for treatments that will be infrequently used (assuming you can count on treatment-outcome follow-up). In contrast, treatments orphaned simply because they’re for some reason too cheap to be profitable represent a type of good appropriately funded through governments. Those things aside, it strikes me that a scientific establishment that charges half a billion to test a drug likely has cost issues that extend beyond any rational concern for human-subject protection. [Deleted and reposted to correct a typo.]
the half a billion is from FDA requirements intersecting with liability etc.
i’ve seen a fair bit of how drug development works and there is just no doing it much cheaper (in the US), at least not for a big drug. for somehting more specialzed, you can do better, but i’m starting to see a lot of trials done overseas to save money. you can get a trial done in china for less that 5% of what it costs here. now that they are starting to have a medical establishemnt that runs at international standards and can be trusted (recall that doctors convicted of bribing and making up data submitted to the chinese fda were executed) you may see some significant regulatory and liability arbitrage in drug trials.
the big issue is NCe’s (new chemical entities). it’s much less expensive to lauch a “me too” which is why you get 4 sleeping pills and 3 ED drugs, but something genuinely new gets subjected to so much scruitny that soup to nuts you spend $500mn-$2bn to get the NDA done.
Adams C, Brantner V (2006). “Estimating the cost of new drug development: is it really 802 million dollars?”. Health Aff (Millwood) 25 (2): 420–8. doi:10.1377/hlthaff.25.2.420. PMID 16522582.
then you might not even get aproved. this is making innovation VERY diffcult and funneling it into huge applications only. how else can you recoup a billion dollars?
it seems to me that if we could allow small, inexpensive studies for drugs with a small market and then use pahse 4 follow up once they are on the market we could take a lot of the cost out of the small drug development process, but we really need to get the costs on NCE’s down too. this process is horribly out of hand.
“[R]ecall that doctors convicted of bribing and making up data submitted to the chinese fda were executed….“
I don’t, as a rule, support the death penalty for data fudging.
“[T]he big issue is NCe’s (new chemical entities)….[Y]ou spend $500mn-$2bn to get the NDA done. “
The cost of trials _shouldn’t_ be a large portion of that–though, who knows, maybe they are these days. If you can’t run a husub for a grand or two, not counting the manufacturing (not development) cost of the treatment, you really ought to be in another line of business.
“[I]f we could allow small, inexpensive studies for drugs with a small market and then use pahse 4 follow up once they are on the market we could take a lot of the cost out of the small drug development process….“
Absolutely. And any law or rule that prohibits doing so is incoherrent in the context of a system designed to balance risks against benefits.
“….and tolerable side effects.”
It may be a necessary concept, but it sure is one scary idea.
My colleagues and I have just concluded one of India’s most in-depth studies on “Clinical Trials in India”. We have looked at various areas such as:
– Market Trends
– Growth Drivers
– Regulatory Bodies and Framework
– Major Players.
We interviewed over 200 individuals and firms to collect the data in what we believe is one of the most detailed study on the subject in India. If you are interested in a copy, you may email me at email@example.com.
the problem to me is that instead of selecting a treatment for a patient based on what is best to him or what he/she believes is best for him, we randomise it with only criterion the control of statistical bias. As a statistician, I feel that our field needs to work harder, and the regulatory authorities should be more sensitive to the patients and not only to society in general.