116 of 127 people found the following review helpful
The Nerd Rampant - great stuff,
Amazon Verified Purchase(What is this?)
This review is from: Bad Pharma: How drug companies mislead doctors and harm patients (Paperback)
This is an important book, which tackles a problem that affects all of us - the development and use of effective and safe medicines. It's well researched and passionately and engagingly written. He presents a number of tricky topics clearly (I think) without losing their essence.
The case he makes for publishing the data from all trials is made in vigorously sugar rush kind of a way. But the calm measured ways have been tried (as he reports) and Ben's entertaining blunderbuss of criticism to whip up public/political pressure seems like a good plan B to me.
The section I have to comment on is the area of clinical trials. I've worked with a lot trial designers and a number of trials. You could get the impression from the book that Pharmaceutical companies have many tricks they can pull to pretty well guarantee success. They don't, most drug development project fail and they fail because their trials fail, there's roughly a 80% failure rate after a successful 'first in humans safety study', but that's across all compounds and therapeutic areas. The more novel the compound and the harder the disease is to treat, the lower success rate.
On the aspects of trials that Ben criticises many are valid criticisms, but for trials that drug companies have to run to prove that a compound is an effective drug, many of the practices are ruled out by the regulators, in particular post-hoc sub group analysis, bundling outcomes and using uninformative outcomes. I'd also like to say that in my direct experience with the pharmaceutical industry, the clinical trials were designed to elicit the truth about the compound, not make it look good, by people that are smart, honest and diligent.
As a client emailed to us today, when we commiserated with a failed trial: "Thank you ... we believe that the results were credible and while we all wanted a positive study, our job is to uncover the truth."
Ben's problem with conducting trials on 'ideal' patients, this does not make it more likely for the drug to be successful, they allow it to be shown to be successful with fewer patients. The trials aren't conducted in subjects more likely to recover and make the drug look good - the patients in the control group will be recruited from the same population and recover too making it hard to show the drug is effective. The ideal patients are those most sensitive to treatment - the `in the goldilocks zone' of neither having so mild a condition that they have a high chance or a good outcome, or so severe a condition that even a good drug is unlikely to improve matters. Being able to study the drug and prove or disprove its effectiveness in fewer patients is also an ethical thing to want to do.
Ben's problem is that the medical profession needs to understand the drug in a broader population. The drug developer's problem is that drug development already costs hundreds of millions of dollars and consumed 10+ years of the drug's (nominally 20 years) patent life. Pharmaceutical companies may have been rich in the past but all are now merging, laying off staff and closing sites. So broader medical knowledge needs to come from the things Ben proposes - big simple studies using healthcare data gathered in the community and thorough and well researched meta-analyses.
Ben complains at the use of 'surrogate outcomes'. He gives the example of statins and control of cholesterol level. He gives that example because the only two 'surrogate outcomes accepted by regulators are cholesterol level and blood pressure. Perhaps giving up smoking and weight loss are also surrogate markers? Pretty much any other disease where you want to try to prove you have a drug that treats it, you have to show 'clinically meaningful' outcomes. The regulators are comfortable with cholesterol and blood pressure because of the unanimity of medical opinion that these were causes of heart problems. And studying the reduction of heart problems directly in trials requires studying thousands of patients for years.
A suggestion he makes is that trials for Diabetes treatments should be based on reducing deaths from diabetes, but diabetes sufferers also suffer from consequential liver disease, kidney disease, neuropathic pain, loss of fingers and toes and sight loss. Now I assume these are all consequences of the diabetics lack of control over their blood sugar levels, if that's correct and I was a diabetic I'd be happy with a drug proven to control sugar levels and would not want to wait until it was proven to reduce mortality. [Ben elsewhere criticises the UKPDS -trial - a big, long running study in Diabetes that is generally taken to show that stringent control of blood glucose levels and/or blood pressure reduce type 2 diabetes complications. The study is cited by many Diabetes groups. Ben's criticism is of a compound endpoint but I'm not clear if the criticism applies to the original outcome (1998) or the 10 year follow-up (2008) I could find the latter easily and they looked pretty convincing. Learning point: if Ben cites something its worth following it up.]
The biggest point I want to dispute is this business of stopping trials early. Looking at the data to in order to adapt the trial, in particular stop it is a very sane, reasonable and ethical thing to do and results in better science. Various statistical approaches to allow this to be done so that it doesn't also amount to cheating have been worked on in academia and industry for the past 25 years or so. The key aspect of all the approaches is that the rules by which decisions are to be taken are pre-specified and the consequences well understood.
Ben tells some tales of the worst behaviour by Pharmaceutical companies apart, highlighting the need for regulation and huge punishments in an industry where the rewards can be so great. But I think that is known.
More importantly Ben highlights some real and pressing problems that require pretty big changes to bring about. First pharmaceutical companies operate in a highly regulated environment, where those regulations don't always work in our best interest. Companies need to get the data on their drug necessary for licensing; this is already a big and onerous task (and much bigger and more onerous than it was 10 years ago). This data is not a good match for physician's needs - this mismatch Ben documents vividly. The best solution is to better manage drugs and their information after they've been approved, with the access to the full data, meta-analyses and big studies using public healthcare records. To move the requirement to pre-approval without also changing the patent laws will effectively make drug development unaffordable.
Secondly - and this is an area where I've no direct experience, the management and analysis of post registration trials needs management - these seem to be essential to broaden the evidence base to patient groups not studied for registration, to amass evidence for rare but serious safety issues, to study other outcome measures (such as quality of life, and long term mortality), to identify sub-population effects. From Ben's tales of clinical trial malfeasance or incompetence it would seem these aren't to the same standard as drug development trials.
Lastly this data needs collating and analysing by cool heads, in particular neither swayed by patient advocacy groups funded by the drug companies nor needing a juicy finding in order to justify publication.
Tracked by 4 customers
Sort: Oldest first | Newest first
Showing 1-10 of 15 posts in this discussion
Initial post: 28 Sep 2012 09:52:20 BDT
This is a great review- thank you for taking the time to write it
In reply to an earlier post on 28 Sep 2012 17:07:18 BDT
Thank you - yes it did take a while to write, somewhat longer than I planned!
To properly review Ben's book would probably result in a review 4 times larger than the book!
Posted on 28 Sep 2012 22:33:57 BDT
In reply to an earlier post on 29 Sep 2012 07:08:33 BDT
"In modern non-scientific use, however, it is generally not treated as a plural. Instead, it is treated as a mass noun" from the defintion of 'data' in the OED.
The context of 'non-scientific use' I interpret roughly as being 'when not writing a an academic peer reviewed paper'.
Posted on 29 Sep 2012 09:40:19 BDT
J. Murdoch says:
In reply to an earlier post on 29 Sep 2012 12:02:52 BDT
Dr. G. S. Pyne says:
Not worth bothering with the book then...I think it's even longer.
Posted on 29 Sep 2012 23:58:19 BDT
Thank you for such a comprehensive and mostly balanced review, but I detect a slight naivety when you suggest that pharma executives are interested in !the truth about the compound." The truth is that the truth is subjective and whilst I don't believe that all pharma executives are corrupt, I do think that many of them are corrupted by the industry which they serve, albeit in subtle ways. I have friends in the sector that are decent honest people who nevertheless are only too ready to justify the unjustifiable because they are immersed in industry hype. Yes drugs are costly to develop and yes that needs to be paid for, but those costs are invariably inflated by inefficiencies, some of which are documented in this book.
I would agree that the author gets carried away a little - criticising a flu drug trial on the grounds it did not include efficacy data on curing other illnesses is perverse, and some of Goldacre's own arguments are not exactly based on peer-reviewed data - but overall I was surprised at the relative lack of hyperbole
In reply to an earlier post on 30 Sep 2012 17:33:31 BDT
Last edited by the author on 30 Sep 2012 17:34:13 BDT
Thank you - I guess I wanted to reflect how I find the people in Pharma that I work with, because its at odds with the image I imagine people get from tales of the pharma industry's worst episodes. Of course I work with scientists rather than exceutives, working on the scientific side and not the commercial side. Given the sorts of things Pharma have been fined for, these companies clearly can have a 'dark side', I give thanks that I've never encountered it directly.
Posted on 1 Oct 2012 09:49:45 BDT
Bic Ballpoint says:
"this business of stopping trials early"
The issue isn't stopping the trial early. It's stopping the trial early WITH THE SPECIFIC INTENTION TO MISLEAD. (Sorry, not sure how else to express emphasis here). And also not announcing that beforehand that this option is available to the sponsor.
In reply to an earlier post on 1 Oct 2012 11:00:32 BDT
Who knows what someone else's intentions are? They may even be confused about them themselves, or have misled themselves as well as others, with a better outcome from the early termination regarded as a "Happy coincidence" rather than the reason for it. Well designed trials should state at the outset what triggers for early cessation will be but it is not always possible to foresee every circumstance and the fact that many trials are not well designed is not evidence of wilful misconduct. The trial may also be stopped early for commercial reasons - scientists are not always comfortable with that but at the end of the day all of this has to be paid for somewhere.
Of course there is also ill intent in the mix as well, but probably not quite as much as you might think and a good percentage of early terminations are IMHO down to poor design in the first place, which is more due to negligence than anything else. That, combined with arrogance, which of course is often associated with ignorance and leads to an unwillingness to consider outsiders in the design (like patients, maybe)?