Placebos (testing new medicine)

Factual Questions
21

Historically I’m sure there was a much greater willingness to use placebo controlled trials (even in circumstances which I found surprising) in the US than there was in Europe. The “advantages” of placebo control are that it gives an absolute measure of benefit; and it will produce a bigger margin of benefit than comparison to current standard of care will. The latter is more important than you might immediately think. IIRC someone upthread pointed out that this makes showing the superiority of a new treatment more difficult; if you can still do it, it will require larger studies (therefore putting more subjects at risk, at least in theory) and thus cost more to develop the new drug, possibly impacting viability of the development. The better medicines get, the harder it is to prove a new one is an improvement.

Not that I’m arguing for a return to the good old days of placebo. I’ll go with the Helsinki Declaration: Ethical Principles For Medical Research Involving Human Subjects.

Not that every regulator is entirely happy with this. See also Wikis:

and

j

22

Has the drug been approved in the meantime?

23

That was many years ago. It was actually a new use for an old, generic medication, so all he needed to do was convince other doctors it was worth trying for this condition. Not only did he succeed, but that discovery led to new research in the field and that drug is now obsolete for that condition, as better treatments have been developed. And what used to be an always-deadly disease is now often a manageable chronic condition.

He discovered it because he’d given the drug to someone who had that disease (primary biliary cirrhosis) and also had a condition that was commonly treated with that drug (methotrexate). And a year later, the guy was still alive and his liver actually improved. So the next time he had a patient with primary biliary cirrhosis he asked the guy, “are you a gambler”, and gave it to him. And he improved, too. Then he set up a formal, double blind study.

24

If I didn’t make it clear in the post about my friend-- he’d reached end stage, and had gotten to the point where “standard treatment” was palliative only, and prepare for death. You sometimes get to a point where you are either too weak for chemo, and it will kill you faster than the cancer itself, or where the cancer just doesn’t respond to it, so there’s no point in doing it.

The treatment he was in a trial for could become standard for people in earlier stages, if it showed that it made some difference for people like him, even if the differences were mainly measurable at microscopic levels (he had to agree to allow an autopsy that involved lots of biopsies).

Living a year, when his life expectancy was three months was a pretty big deal. He got another birthday, another Christmas, and saw a nephew born.

I honestly have no idea what happened with the treatment he had, he had a cancer that was almost always fatal, even when it was caught fairly early. And it caused a severe decline in quality of life for people who had it. What was really important about his last year was that his decline stopped. He didn’t regain any function, but didn’t lose more, until maybe a week before he died.

So it might’ve been something that replaced standard care, or was administered in addition to it, I guess.

Just from what I know about cancer research (and if I have the details wrong, someone please correct me), because DH works in it (he’s a lab rat who prepares sample for examination from biopsied tissue and blood draws), sometimes drugs are tested first on “hopeless” patients to see if there are “microchanges”-- in other words, they may not live much longer, but there might be statistically significant changes to tumors observable under microscope, that show the treatment was effective.

The step after that is a trial where it’s given alongside standard treatment-- or not-- in a double-blind study.

25

I have a rare form of blood cancer and I’ve been asked to join a new drug study. If I choose to join the study I will have a 50/50 chance of being given the new drug. If I’m not in the chosen 50% I will be asked to simply continue taking my usual medicine and I will be monitored as part of the control group.

26

I’m sorry for your loss.

It sounds like your friend was part of a Phase II trial, which may or may not involve RCT. In his case, it sounds like it did not. Since you indicate that there is no actual standard of care treatment at the stage his disease was in, it sounds like he wasn’t being given the experimental drug in place of anything - he was receiving the drug in addition to the current standard of care, which in his case was purely palliative.

27

I was wondering about the timing of this, because these days the conditions under which the trial would be stopped are all laid down very precisely before hand in the study design stage, along with complicated statistical methods that take the potential early stopping times into account.

But getting to the OP, the study design of a trial has to pass rigorous vetting for ethical consideration, and any trial for which one arm got less than the current standard of care would never get approved.

28

Another reason for this is that, although it may sound morbid, high mortality rates make smaller studies possible. The statistical power of a study is approximately equal to the square root of the number of events (deaths). So testing the drug on a set that has poor prognosis gives you more bang for your buck. Finally these patients might not have much to lose, so there is less danger that the even if the drug proves to be to worse than the standard of care.

29

Interesting. This must have been in the 1970s.

30

Yeah, I kinda figured that, when they were taking people who were “palliative care only.”

ETA: Actually, it quadrupled his life expectancy-- 1 year instead of 3 months. So either he was very lucky, or he got the drug, and it was pretty effective. I don’t know if you can extrapolate, but if someone had three years, and got 12 instead, that would be a really significant gain-- that’s enough time for an even better treatment to come along.

31

Thanks for the information, everyone.

@RivkahChaya, I’m sorry to hear about your friend.

32

Thanks. It was a long time ago-- almost 20 years. My son is named for him.

33

It really depends on the disease. When they tell a person they have 3 months to live, what they really mean is that the median survival is three months, so that half of the people live less than three months and half live more. But it could there could be a substantial portion that live a lot longer.

For example look at the following plot.

This is what is known as a Kaplan Meier survival curve. The the lines show the estimated probability (y-axis) of a patient surviving a given number of years (x-axis) If you look at the red line, only 50% of patients live 6 months ,so the doctor will tell them they have 6 months to live, but an estimated quarter of the patients live 7 years.

You hear these sorts of situations touted as miracles or proof of some alternative medicine or that doctors are quacks, but really its just luck, statistics and poor communication.

34

I know how statistics work-- but he also was a lot more comfortable during his last year than he had been, and his decline slowed way down. Except for his last week or so, his loss of function was almost arrested-- he didn’t regain what he’d lost (he had a malignant brain tumor), but he didn’t have any more loss.

Anyway, his life expectancy was based on his rate of decline, and how long he could expect until his nervous system failed, and he stopped being able to breathe or have his heart pumping, his other organs functioning, and so forth.

35

Just to be clear, my final comment wasn’t directed in any way shape or form at you, I misread your ETA as a question rather than a comment. Given that he was on a experimental treatment it is entirely possible that it did him good and was the reason behind the slowing of his progression.

My final sentence was directed in frustration at the snake oil salesmen and religious fanatics who treat anecdote like proof.

36

And of course, extrapolation is always fraught with peril, and extrapolating from a single data point is a fool’s errand. If the drug gave him 1 year instead of 3 months, then maybe that means that the drug quadruples life expectancy, and so someone with a 2 year expectancy would get 8 years. Or maybe it means that the drug adds nine months, and so someone with a 24 month expectancy would get 33 months. Or maybe it means that everyone who gets the drug lives for one more year, and so if someone had a 2 year expectancy before, the drug would halve that.

This is why scientific studies are based on lots and lots of data points, not just an individual anecdote.

37

Yes. I know. But there was probably reason to be cautiously optimistic.