99% of that article tries to get you to feel sorry for someone. The remaining 1% presents a few facts:
1) Clinical trials are currently considered necessary to prove drugs work.
2) Some people think the clinical trials are cruel to the control group.
... That's it. Here's a genius quote:
-- Some melanoma specialists familiar with the drug would have traded the data for faster access to the drug. “I know all that I need to know based on the results we already have,” said Dr. Keith Flaherty of Massachusetts General Hospital, who led the early clinical testing. “My use of this drug is not going to be informed by testing it against a drug we all hate and would rather never give a dose of again in our lives.” --
So, he saw some initial results, but not enough to be conclusive to any objective observer. That was enough to make his mind up. He doesn't care to follow science because he -believes- the drug will work now.
Thank goodness he's not -my- doctor.
I don't disagree about the cruelty of putting people in the control group. You're basically denying them medicine that they need. But we don't have an alternative at the moment.
It's also worth remembering that everyone that goes into a study like this is told that they have a 50/50 chance of not getting the medicine. They KNOW this going in. They choose to take that chance because their current medicine isn't doing the job... Or whatever other reasons they have. Getting upset that got into the wrong 50% is pointless. That was part of the deal.
And finally: How does he KNOW he's in the control group? Maybe the medicine simply doesn't work on him. Even the doctor wouldn't know if this were a double-blind test.
It's not a blinded trial, according to the article: “Because of the different ways the drugs were dispensed — one by mouth and one by infusions — doctors and patients, it was decided, would both know who got which drug.”
(Note: didn't read the article, just the HN comments.)
"I don't disagree about the cruelty of putting people in the control group. You're basically denying them medicine that they need."
The problem is, you don't know this until you do the trial.
I'm not up on current figures, but sometime ago a read that the efficacy of placebos in treating organic (i.e. very real and physical) heart damage was 50% plus or minus a bit (I think it was plus). (Given all the connections we know of between the nervous system and things in the body that could allow this the finding was not entirely surprising.)
They aren't entirely heartless, so to speak: sometimes they will stop trials involving life or death conditions and start giving everyone the new drug if the initial results are fantastic enough and they've got enough control evidence (including how populations not in the study have responded to the old treatment(s)).
And there are enough phase 3 (efficacy) trial failures it's clear we have to do this; the alternatives are worse.
There's also the "first do no harm" rule, which I'd argue is in alignment with this process. Giving someone a new drug that's worse than existing treatments is harmful and one sad fact of side effects is that by definition the more rare they are, the more people you'll have to give the drug to to be able to find them. At each level, phases 1, 2, 3 and general marketing you expose it to larger populations, which is the only way to find out :-(.
>Even the doctor wouldn't know if this were a double-blind test.
Actually there's plenty of evidence that doctors do tend to know which groups patients are in. In SSRI trials, they can be about 80% accurate (can find the ref at work tomorrow if needed). Patients also tend to know — probably because of the different side effect profiles in drug vs placebo arms. This is one of the major arguments for running trials with active placebo arms — that is, a placebo which does have side effects but which is assumed or known non-efficacious for the problem under investigation.
Without reading the article, but as someone who has a condition where lots of clinical trials are done, I will offer this opionion:
Clinical trials can be pretty idiotic because a) they don't control for enough factors and b) the act of qualifying for the trial can confound the results, something that seems to be completely ignored.
On a cystic fibrosis forum I belong to, one member wanted to participate in a clinical trial. A precondition was that she not be on a particular drug that she was on. So she goes to the forum and asks for non-drug alternatives so she can get off this drug and qualify for the study. If she does better during the study: Is it because of the drug being studied? Is it because she stopped taking this other drug to qualify for the study? Is it because of the non-drug alternatives she began using as a substitute? Is it two or more of those reasons and how much of the change can be attributed to each factor?
If she qualified, I seriously doubt the people making the study would ask any of the above questions. If they did, I have no clue how you could determine meaningful answers. It looks like junk data to me.
>he act of qualifying for the trial can confound the results
The example you give doesn't confound the result of the trial (provided it's properly randomised) but it does make it hard to generalise the result to the general population.
What you describe is actually common — antidepressant trials typically exclude people with comorbid medical or psychological conditions, despite the fact that the rate of comorbitity in depression runs at >60%.
Hello this is malignant melanoma it does not take a genious to use scientific method to compare the results without a control. The end result and time for that result is not in question. Once it sperads you are going to die in a very predicta le amount of time.
The definition of insanity to trying the same thing over and over expecting different results well this qualifies.
I am usually horrified by the lack of true controls in medical research because of poor understanding of the variables and unintentional(or gaming) results. To me this is just one more example of medicine not understanding scientific method and statistics.
Please not if you catch it early it is often trivial to effectively treat. Be aware and see a doctor befor there is noting to do.
7 comments
[ 1.9 ms ] story [ 32.9 ms ] thread1) Clinical trials are currently considered necessary to prove drugs work.
2) Some people think the clinical trials are cruel to the control group.
... That's it. Here's a genius quote:
-- Some melanoma specialists familiar with the drug would have traded the data for faster access to the drug. “I know all that I need to know based on the results we already have,” said Dr. Keith Flaherty of Massachusetts General Hospital, who led the early clinical testing. “My use of this drug is not going to be informed by testing it against a drug we all hate and would rather never give a dose of again in our lives.” --
So, he saw some initial results, but not enough to be conclusive to any objective observer. That was enough to make his mind up. He doesn't care to follow science because he -believes- the drug will work now.
Thank goodness he's not -my- doctor.
I don't disagree about the cruelty of putting people in the control group. You're basically denying them medicine that they need. But we don't have an alternative at the moment.
It's also worth remembering that everyone that goes into a study like this is told that they have a 50/50 chance of not getting the medicine. They KNOW this going in. They choose to take that chance because their current medicine isn't doing the job... Or whatever other reasons they have. Getting upset that got into the wrong 50% is pointless. That was part of the deal.
And finally: How does he KNOW he's in the control group? Maybe the medicine simply doesn't work on him. Even the doctor wouldn't know if this were a double-blind test.
"I don't disagree about the cruelty of putting people in the control group. You're basically denying them medicine that they need."
The problem is, you don't know this until you do the trial.
I'm not up on current figures, but sometime ago a read that the efficacy of placebos in treating organic (i.e. very real and physical) heart damage was 50% plus or minus a bit (I think it was plus). (Given all the connections we know of between the nervous system and things in the body that could allow this the finding was not entirely surprising.)
They aren't entirely heartless, so to speak: sometimes they will stop trials involving life or death conditions and start giving everyone the new drug if the initial results are fantastic enough and they've got enough control evidence (including how populations not in the study have responded to the old treatment(s)).
And there are enough phase 3 (efficacy) trial failures it's clear we have to do this; the alternatives are worse.
There's also the "first do no harm" rule, which I'd argue is in alignment with this process. Giving someone a new drug that's worse than existing treatments is harmful and one sad fact of side effects is that by definition the more rare they are, the more people you'll have to give the drug to to be able to find them. At each level, phases 1, 2, 3 and general marketing you expose it to larger populations, which is the only way to find out :-(.
Actually there's plenty of evidence that doctors do tend to know which groups patients are in. In SSRI trials, they can be about 80% accurate (can find the ref at work tomorrow if needed). Patients also tend to know — probably because of the different side effect profiles in drug vs placebo arms. This is one of the major arguments for running trials with active placebo arms — that is, a placebo which does have side effects but which is assumed or known non-efficacious for the problem under investigation.
Clinical trials can be pretty idiotic because a) they don't control for enough factors and b) the act of qualifying for the trial can confound the results, something that seems to be completely ignored.
On a cystic fibrosis forum I belong to, one member wanted to participate in a clinical trial. A precondition was that she not be on a particular drug that she was on. So she goes to the forum and asks for non-drug alternatives so she can get off this drug and qualify for the study. If she does better during the study: Is it because of the drug being studied? Is it because she stopped taking this other drug to qualify for the study? Is it because of the non-drug alternatives she began using as a substitute? Is it two or more of those reasons and how much of the change can be attributed to each factor?
If she qualified, I seriously doubt the people making the study would ask any of the above questions. If they did, I have no clue how you could determine meaningful answers. It looks like junk data to me.
What you describe is actually common — antidepressant trials typically exclude people with comorbid medical or psychological conditions, despite the fact that the rate of comorbitity in depression runs at >60%.
The definition of insanity to trying the same thing over and over expecting different results well this qualifies.
I am usually horrified by the lack of true controls in medical research because of poor understanding of the variables and unintentional(or gaming) results. To me this is just one more example of medicine not understanding scientific method and statistics.
Please not if you catch it early it is often trivial to effectively treat. Be aware and see a doctor befor there is noting to do.