Ask your doctor about buried evidence
March 29, 2018 9:37 AM Subscribe
Clinical trial results are not always reported or published, especially when they do not show that a drug is safe and effective. Doctors need access to all clinical trial results in order to make informed choices about drugs and provide the best care to their patients. The British Medical Journal's new Unreported Clinical Trial of the Week initiative is calling attention to this publication bias issue with the ultimate goal of improving reporting rates.
The 2007 FDA Amendments Act requires that certain clinical trials in the U.S. report results, but compliance is poor and the FDA has never levied a fine for non-compliance even though it has the authority to do so. The FDAAA TrialsTracker identifies every trial from ClinicalTrials.gov that has gone unreported.
Dr. Ben Goldacre, founder of the AllTrials advocacy campaign and a key contributor to the FDAAA TrialsTracker project, is a fellow at the Evidence Based Medicine DataLab at the University of Oxford and is widely known for his 2003-2011 "Bad Science" column in The Guardian.
The 2007 FDA Amendments Act requires that certain clinical trials in the U.S. report results, but compliance is poor and the FDA has never levied a fine for non-compliance even though it has the authority to do so. The FDAAA TrialsTracker identifies every trial from ClinicalTrials.gov that has gone unreported.
Dr. Ben Goldacre, founder of the AllTrials advocacy campaign and a key contributor to the FDAAA TrialsTracker project, is a fellow at the Evidence Based Medicine DataLab at the University of Oxford and is widely known for his 2003-2011 "Bad Science" column in The Guardian.
This post was deleted for the following reason: Poster's Request -- Brandon Blatcher
Just to calibrate people: To my knowledge, In order to get approval for a particular indication (i.e to sell a drug), in the UK and US, clinical trials MUST be published.
To my knowledge, if a clinical trial fails or is stopped with the FDA's approval process, then they have to notify the agency of that stopping or failure too......
The question is, how much - if any - material is published prior to submitting to regulatory agencies, or from 'speculative trials' or details from failed trials.
The answer is : not enough.
posted by lalochezia at 11:59 AM on March 29, 2018
To my knowledge, if a clinical trial fails or is stopped with the FDA's approval process, then they have to notify the agency of that stopping or failure too......
The question is, how much - if any - material is published prior to submitting to regulatory agencies, or from 'speculative trials' or details from failed trials.
The answer is : not enough.
posted by lalochezia at 11:59 AM on March 29, 2018
The results are buried along with the patients!!!!
posted by Burn_IT at 12:29 PM on March 29, 2018 [2 favorites]
posted by Burn_IT at 12:29 PM on March 29, 2018 [2 favorites]
At least in the US, which is what I'm all too familiar with on this topic, this is a truly hairy problem, and attempts to solve it have had some recent, weird, consequences for human subjects research more broadly. This is going to be long, and requires some definitions first. Clinical trials are a subset of human subjects trials more broadly that specifically focus on the clinical efficacy of a drug, device or treatment, whereas human subjects trials are any experimental manipulation involving human subjects. Clinical trials are regulated by, among other things, FDAAA (42 CFR part 11). Human subjects trials, across the board, are subject to what is usually called the Common Rule (45 CFR part 46).
So, FDAAA does require registration and reporting of designated clinical trials, and has required it since 2007. The thing is, no one really wants to spend the time to report failed clinical trials, and that's a problem, particularly for federally-funded research, because the National Institutes of Health (NIH) really doesn't want to fund known dead-ends. The really big offenders here are oncology trials (supported by the National Cancer Institute, NCI), and failures to report in the oncology space got bad enough that the National Institute of Medicine got grumpy about it in 2010, which resulted in language in the 2012 appropriations bill that told NIH as a whole to figure out a way to mandate compliance here, because there's no reason for this not to just be a thing. The goal is "Ok, you're going to do a clinical trial. Great! Tell us what you're going to do first, and then tell us what happens so we can move science forward."
That's all well and good, and while NIH was kind of slow getting off the blocks after 2012 to get things moving, they were making slow if solid progress towards a global policy here. Things went a bit off the rails last year, when NIH decided that, as far as they were concerned, all human subjects trials were, essentially, going to be treated as clinical trials going forward. Since NIH supports research that ranges from "test this new cancer drug" to "find the T among Ls" here, this went down about as well as you'd expect, which is to say not well. This was not helped by NIH being decidedly unclear about what was or was not a clinical trial by their new definition, and by bifurcating their grant application system into "clinical trial permitted" and "no clinical trial permitted." Things have actually gotten (somewhat) better as of a week or so ago, with the 2018 Appropriations Bill, which basically tells NIH that yes, they should make the actual clinical trials register and report their results they way they'd been planning to for years, but that they shouldn't overreach to try to encompass all human subjects experiments.
This is the really long way of saying oh my, this is a big problem at the moment, and it's harder to solve that anyone would really like it to be. I know as much as I do about it, because I'm a basic science researcher (as I've mentioned before, looking at visual perception and driver behavior) and I spent a lot of last fall helping to push against NIH's attempt to make everything a clinical trial. I'm fully supportive of registering studies and reporting results, because that's how science moves forward (and it's a matter of respect for your subjects), but not everything is a clinical trial, and it's good to bear that in mind.
posted by Making You Bored For Science at 12:47 PM on March 29, 2018 [5 favorites]
So, FDAAA does require registration and reporting of designated clinical trials, and has required it since 2007. The thing is, no one really wants to spend the time to report failed clinical trials, and that's a problem, particularly for federally-funded research, because the National Institutes of Health (NIH) really doesn't want to fund known dead-ends. The really big offenders here are oncology trials (supported by the National Cancer Institute, NCI), and failures to report in the oncology space got bad enough that the National Institute of Medicine got grumpy about it in 2010, which resulted in language in the 2012 appropriations bill that told NIH as a whole to figure out a way to mandate compliance here, because there's no reason for this not to just be a thing. The goal is "Ok, you're going to do a clinical trial. Great! Tell us what you're going to do first, and then tell us what happens so we can move science forward."
That's all well and good, and while NIH was kind of slow getting off the blocks after 2012 to get things moving, they were making slow if solid progress towards a global policy here. Things went a bit off the rails last year, when NIH decided that, as far as they were concerned, all human subjects trials were, essentially, going to be treated as clinical trials going forward. Since NIH supports research that ranges from "test this new cancer drug" to "find the T among Ls" here, this went down about as well as you'd expect, which is to say not well. This was not helped by NIH being decidedly unclear about what was or was not a clinical trial by their new definition, and by bifurcating their grant application system into "clinical trial permitted" and "no clinical trial permitted." Things have actually gotten (somewhat) better as of a week or so ago, with the 2018 Appropriations Bill, which basically tells NIH that yes, they should make the actual clinical trials register and report their results they way they'd been planning to for years, but that they shouldn't overreach to try to encompass all human subjects experiments.
This is the really long way of saying oh my, this is a big problem at the moment, and it's harder to solve that anyone would really like it to be. I know as much as I do about it, because I'm a basic science researcher (as I've mentioned before, looking at visual perception and driver behavior) and I spent a lot of last fall helping to push against NIH's attempt to make everything a clinical trial. I'm fully supportive of registering studies and reporting results, because that's how science moves forward (and it's a matter of respect for your subjects), but not everything is a clinical trial, and it's good to bear that in mind.
posted by Making You Bored For Science at 12:47 PM on March 29, 2018 [5 favorites]
This BMJ initiative is a welcome, yet tiny, counterpoint within what may be the most severely subverted realm of science of all - health and medicine. The scientific method is a vital tool for developing the knowledge and wisdom needed to inform our actions as a species, but it is open to corruption and error. I can't think of any industry with a more corrupting effect on a body of science than the pharaceutical sector.
I recommend Goldacre's book 'Bad Pharma' to anyone interested in how to make sense of medical evidence with all its inherent flaws and biases.
Having sat on medical research ethics committees I was already cynical about the quality and impartiality of much of the science informing medicine before reading his book, but Goldacre made me realise things were far worse that I'd supposed.
Thanks for posting xylothek
posted by Neil Hunt at 9:28 PM on March 29, 2018
I recommend Goldacre's book 'Bad Pharma' to anyone interested in how to make sense of medical evidence with all its inherent flaws and biases.
Having sat on medical research ethics committees I was already cynical about the quality and impartiality of much of the science informing medicine before reading his book, but Goldacre made me realise things were far worse that I'd supposed.
Thanks for posting xylothek
posted by Neil Hunt at 9:28 PM on March 29, 2018
« Older Not So United Front | The Stourbridge sound Newer »
This thread has been archived and is closed to new comments
posted by greenish at 10:01 AM on March 29, 2018 [4 favorites]