Ghostwriters of Science

The great invisible social science story of the modern era is the triumph of stealth in the engineering of professional opinion. Invisible to the radar of the dwindling number of guardians of truth, corporations have bought, bribed, suborned, twisted, bent, and perverted thousands of influential academics, public officials, and “public intellectuals.” Look at the ruins of the economics “profession” for abundant evidence.

The greatest triumph of this malign invisible movement is its normalization. An entire generation of knowledge workers now equates professionalism with material success. This is what Christopher Lasch termed the “betrayal of the elites,” but one cannot fault the elites entirely for pursuing greed over virtue. It is the stealth technicians who deserve the bulk of the blame. The slimy people who engineered the cunning, plausibly deniable schemes for warping the intellectual elite are the great villains of this story.

A mere three million bucks for all these rock-star academics? When you look at what Bill Clinton or Sarah Palin are/were? paid, you have to think of these professors as pikers.

The Gaddafi story reminds me of the time, years ago, when I overheard an antiquarian bookseller boasting that librarians could be bought for virtually nothing — contracts for carload lots of stuff cleaned out of the attics of Europe rarely cost him more than a dinner and a little flattery. This was, of course, back when many state university systems in the U.S. were expanding, and were flush with the money needed to buy themselves a proper accreditation.

Title suggestion: “Ghost Writers in the Sci…”

A correction and 2 questions:

The Correction: scientific content in support of our clients’ messages should be either:

scientistic content in support of our clients’ messages

or

scientific [sic] content in support of our clients’ messages

The questions concern this passage:

“We’ve never done ghostwriting, per se, as I’d define it”, says John Romankiewicz, president of Scientific Therapeutics Information, the New Jersey firm that helped Merck promote Vioxx with a series of positive articles in medical journals. “We may have written a paper, but the people we work with have to have some input and approve it.”

1) Since when is Romankiewicz Mr. Webster (or Humpty Dumpty)?

2) Isn’t it the case that nominal authors always have some input (if only to tell the story) and have to approve what has been ghosted for them?

The last comment need not be published (nor this one) since it duplicates one that I submitted a couple of hours earlier. I was curious about the moderation — hasn’t happened to me here before — so I thought to try again to see what was going on. I imagine that I pissed off Belle Waring, which I regret, esp. since I agree with several others in the comments from that post that it would be good if she posted more frequently.

“We’ve never done ghostwriting, per se, as I’d define it”, says John Romankiewicz, president of Scientific Therapeutics Information, the New Jersey firm that helped Merck promote Vioxx with a series of positive articles in medical journals. “We may have written a paper, but the people we work with have to have some input and approve it.”

Hmmm. Does medical ethics not require that **all** authors be credited in published papers?

Less here than meets the eye

However marginally these drug company efforts may end up influencing physician practice, patients shouldn’t have to worry that that practice is for sale, so, yes, it is a disgrace that the profession tolerates this crap, and that the industry is allowed to do this.

That said, the influence really is quite marginal. Now, there’s gold in them thar margins, in the sense that the medications still under patent that are pushed in this manner are priced so that very marginal increments in sales mean a lot, so there is ample incentive for the drug companies to engage in this sort of practice, no matter how little it effects practice overall. But the effect on the cost of medical care, or its quality (the idea that such techniques will get physicians to prescribe the wrong medications), is quite marginal. Big Pharma gets to gouge us because it is part of the overall health care cartel, because we allow cartels in health care, not because it uses these means to wangle marginally more use of expensive new drugs than is warranted by clinical necessity.

Dramatic cures tend to get all the attention, and dominate how people (even professionals, who ought to know better) tend to think of medicine. But mostly we are only able to fiddle around the margins of disease, get some marginally better result than not doing anything, than using Christian Science instead of seeking medical attention. Of course, when the stakes are life and death, even quite marginal improvements are worth pursuing, so, no, using Christian Science instead of medicine doesn’t make sense anymore (as it did up until a century or so ago).

It’s not as if any of these paid shills, no matter what his or her academic credentials, or however they dress up their opinions, could make more than a very marginal difference in clinical practice. Clinicians are going to respond dramatically only to evidence of dramatic differences in outcome. If that evidence exists, the issue will never get to the RCT (randomized controlled trial) level, never get to the point of generating evidence so complicated that ordinary clinicians will need an expert to sort through it for them. It’s only ethical to randomize treatments where there truly is no already existing reason to think that either treatment is better. We’ve never had an RCT on anti-biotics for bacterial meningitis, and never will. All we ever had on that question were a few case series, a very weak study design. Not that clinicians waited for even those to be published before starting to use antibiotics for bacterial meningitis. Evidence of differences in outcome so dramatic that all you need is a case series and historical controls, doesn’t need an expert to interpret and change everyone’s practice overnight. IN contrast, if the differences in outcome are so weak that you need some RCT that found some marginal difference in one statin over another to prove the point, only a marginal few clinicians are ever going to pay heed, whether the expert is some shill or a truly disinterested expert.

“We’ve never done ghostwriting, per se, as I’d define it”… “We may have written a paper, but the people we work with have to have some input and approve it.”

God I want to hear that said by Will Ferrel at the pool party scene of Anchorman while talking to the blond.

Well, as I said, the marginal nature of the effect on practice is absolutely no reason to not ban it. You can’t expect the average patient to know enough about how medicine works to understand that the effect is marginal, and no patient deserves to contemplate the spectacle of the standard of care, on which his life may depend, even apparently for sale.

But the marginal effect on practice is why it is tolerated, why most physicians are rather blase about it.

Personally, the reason I started, some years ago, categorically refusing to partake of any of the little bennies (Free food! You risk getting kicked out of the AMA if you refuse free food, however .) and tchotchkes the drug reps make available, is that they get to write this stuff off as a business expense, then jack up the prices they extort from widows and orphans for their products.

As for the idea that academic integrity is somehow threatened, well I won’t argue against that idea. I thoroughly agree. But I’m rather old-fashioned in that respect, coming as I do from a long line of physicians who practiced back when you couldn’t make any money at it, but you did it for the prestige and place in society the avocation would bring the family, a family that had to have other means. Well, nowadays you can make quite a bit of money at it, so there’s that temptation. And we do live in a society in which the ability to make a lot of money is systematically confused with competence, and the failure to make a lot of money is confused with lack of intelligence and/or energy and ambition.

Now, in that milieu, merely academic physicians don’t really rank too high in any sort of hierarchy, because they don’t make much money, not like sub-specialty surgeons. And ordinary practitioners don’t have any particular reason to feel betrayed if an academic has found some sweet deal with a drug company, because most practitioners don’t really pay much attention to the literature. Please don’t be shocked. As I explained in my first post, anything that needs an RTC to prove it probably is an effect too small to be clinically significant. This habit of physicians in ordinary practice drives the Evidence-Based Medicine people to distraction, but frankly, all the important issues in medicine are resolved prior to the quantitative methods that you need an MPH or the academics to sort through. Yes, MDs should all get MPHs, and Epidemiolgy is way more relevant ot medicine than, say, organic chemistry, but, frankly, neither effects practice much.

Thanks for the insight.

I would be interested in your thoughts on the direct marketing of prescription drugs to end users. Or, as a bumper sticker I once read put it, “ASK YOUR DOCTOR IF GETTING YOUR MEDICAL ADVICE FROM A TV COMMERCIAL IS RIGHT FOR YOU.”

While this is considerably more disturbing than the more run-of-the-mill kind of ghost writing, I’m struck by why we tolerate ghostwriting at all. Why are we supposed to pretend that presidential aspirants and sports stars and tv personalities are really writing the books that are issued under their name? Why is it unacceptable to admit that an institution is responsible for a piece of writing rather than the person who is the figure head for that institution? My guess is that in your average book store the non-fiction section is awash with ghost written works.

More on point, thank goodness we’re decided to turn life and death decisions into profit making opportunities.

“That said, the influence really is quite marginal.”

Glen – I assume from your comments that you are a physician. My experience as a practicing physician and clinical researcher with some experience in drug development is different. These kinds of things do make a real difference in physician prescribing habits. Its not just the ghost-writing, its the whole use of academics as “opinion leaders” as part of a marketing strategy. Its not difficult to find examples of this opinion leader approach substantially enhancing revenues.

Parenthetically, there are many examples of RCTs having a substantial effect on clinical practice. Its correct that there were never RCTs for antibiotic treatment of bacterial meningitis, but there have been RCTs of vaccines for bacterial meningitis, and the latter have had a large beneficial impact.

“We’ve never done ghostwriting, per se, as I’d define it”, says John Romankiewicz, president of Scientific Therapeutics Information, the New Jersey firm that helped Merck promote Vioxx with a series of positive articles in medical journals. “We may have written a paper, but the people we work with have to have some input and approve it.”

That’s the most awesome sentence pairing ever to appear in a quotation, ever. “We’ve never done ghostwriting as I’d define it. We’ve done ghostwriting as you’d define it. … Next question?”

… “Not to my knowledge, no, I’ve never misused the phrase ‘per se’ as I’d define it. I’ve misused ‘per se’ as you’d define it. … Next question?”

“that water the underlying steel”

er, “that water does not corrode the underlying steel”

The Raven,

You will have to stand in line behind the entire Evidence-Based Medicine movement in your concern that practicing physicians mostly ignore current research. My take is that while they and you undoubtedly have a point, it’s not the huge scary issue you might think at first glance that supposed men and women of science don’t pay much attention to scientific research.

The reason for the ignorance, and why that ignorance isn’t really critical, is that the scientific literature in medicine, at least these days, confines itself to reporting the results of clinical research, and that research, and the quantitative methods by which it is conducted, don’t come into play until after all the important issues are settled. You can’t count to apply your quantitative methods until you’ve already decided what you’re counting, until you can distinguish your apples from your oranges. But it turns out that that’s the key step, that’s what takes all the skull sweat, deciding, at the profession-wide level, on a nosology, a list of diseases and their defining clinical characteristics, and at the indvidual patient level, on a differential diagnosis, deciding which, if any, of those defined clinical entities your patient has, or might have with a high enough probability to make it worth while ruling that diagnosis in or out. All of that is pre-quantitative, a matter of gestalt and pattern recognition, with no help from numbers until you’re so close to a diagnosis you don’t need algorithmic help anymore anyway.

Take, for example, prostate cancer. Now, given that this disease is now detected by way of a blood test, and then staged by way of biopsy material sectioned, stained and examined under a microscope, these technological trappings mislead many people, including practitioners, who ought to know better, into imagining that we are in some Physics-like environment where all the answers are as well-defined and clear as the pathology reading of the biopsy specimens. But the technology only tells you prostate cancer vs not prostate cancer, not what to do about prostate cancer. Trial and error had given us an answer to the question of what to do about prostate cancer — remove the prostate. But because that answer was based on empirical observation of what happened if you did vs did not remove the prostate in response to clinically-discovered prostate cancer, and was not built from any deep understanding of how prostate cancer actually works, that answer did not survive the new blood test, the PSA, which detects prostate cancer at a rate higher than it was observed in the past to ever become clinically apparent. It is conceivable that that means that prostate cancer is becoming dramatically more common, but that seems very unlikely, and most people believe it means that the PSA, while detecting prostate cancer that is fated to kill the patient unless his prostate is removed, also detects quite a bit of prostate cancer that will never go anywhere, will never cause the patient the least harm. It is no longer at all clear that whacking out the prostate when prostate cancer is discovered by way of PSA is the right thing to do, because such whacking puts at risk all sorts of plumbing and wiring that run close to the prostate. This question that has arisen because the PSA redefined the nosology, made it necessary to distinguish clinically-discovered from PSA-discovered prostate cancer, is not amenable to answer by quantitative methods. To do an RCT on the question would involve randomizing some men with PSA-discovered prostate cancer to no treatment, when we do not know that this will not condemn them to a death that prompt prostatectomy would have spared them. That won’t be ethical until clinical experience, the same sort of empiricism that taught us that prostatectomy was necessary for clinically-dscovered prostate cancer, tells us that it the surgery isn’t necessary in all cases of PSA-discovered prostate cancer. At that point, yes, I’m sure people will jump in with RCTs, but they will only serve to put a fine point of what exact cutoffs of pathology results (“how much” cancer is present, how high the Gleason Score is) and PSA level dictates which approach, not to tell practitioners that watchful waiting is or is not appropriate for low scores. That question has to be settled, at least roughly, before we can do an RCT.

Just because most practitioners don’t pay much attention to the literature, doesn’t mean that they are not responding to empiricism, not following an appropriate scientific method. It just means that in medicine, the most important empiricicm happens before we can trundle in the quantitative methods that the literature traffics in. Those methods only put a fine point on what we already know, and ususally that point is so fine as to be clinically insignificant, no matter how statistically significant.

Medicine isn’t physics, though they both started out in the same proto-science mishmash. The latter got to choose what it studied, and chose only phenomena amenable right out the gate to quantitative methods. Medicine will never get to choose its subject matter based on its ready quantitative digestibility, it gets its subject matter presented to it by patients, and has to come up with the best answers that empiricism allows under those conditions. So medicine has followed a more devious and obscure path since the two separated.

That said, I personally line up with the Evidence-Based folks. I have my MPH. But frankly, I find the extra level of critical thought about the medical literature it opens up useful mainly as a corrective to any tendency to set too much store by anything in the literature. I can’t be bamboozled into the false attribution of practice-changing findings to every study that rolls down the pike because I have the critical tools, while most of my colleagues cannot be so bamboozled because they don’t bother to read the literature — but the result is largely the same.

The more I hear about the pharmacutical industry, the more convinced I get that we probably ought to burn our entire IP regime in this area to the ground, then salt the soil where it grew. Ultimately, the biggest buyer of all the products of pharma is the goverments of the world, so what we are currently doing is paying both for the research and for the marketing that distorts the research out of public coffers, which is simply madness. It would be far more efficient to void all patents from all medicine, and simply foot the bill for medical research directly out of the taxpayers pocket. Do it right, and it should be possible to realize both cost savings, a bigger total research budget for medicine and far, far higher standards of medical research. Among other things, publicly funded research might be more oriented towards cures than symptom treatments..

Roger,

Because of the monopoly protection we give drug makers under patent protection, you don’t have to shift the prescribing habits of many practitioners at all, or very much at all, in the direction of some new med, to make it worthwhile to the drug companies to go through their whole drill — the opinion leader journal articles and CME presentations, the CME trips, food and tchotchkes, etc., etc. But I can’t point to a single instance of this sort of opinion leader, the drug company shill, having more than a very marginal effect on practice.

Now, since there is not any good at all to balance this very small harm, yes, we should still go ahead and end these practices. I think other harms, particularly the harm to patient confidence created by this circus, are even greater than the very small harm of the very small untoward effect on practice, but in the absence of any good created by allowing Big Pharma to do any marketing at all, they should be prohibited from all such.

Now, if we agree on the practical outcome, then there is perhaps little point in arguing over the rationale for that desired outcome. I only maintain my point about the marginal nature of the influence from these opinion leaders and their sometimes ghostwritten material, for two side reasons — to explain the blase reaction most practitioners have to this circus, because they think of it as something quite marginal, and to avoid falsely inflating this particular circus into an appreciable, much less major, cause of the money and quality problems that medicine is having right now.

To find an example of opinion leadership that actually did cause a more than marginal untoward bias in practice, you have to turn from these drug company shills to opinion leaders organized enough to actually change practice within the profession on a large scale. For many years, most of the relevant professional societies urged estrogen replacement therapy as a cardioprotective for post-menopausal women, despite the best evidence for this at the time being cohort studies that showed RRs no better than 1.3:1. The rule of thumb is that, because of the potential for large imponderable biases inherent in observational studies, you don’t even begin to think about attributing causality until you get past 2-3:1. To have something with only a 1.3:1 touted as the basis for a definite treatment recommendation was a much clearer distortion of usual standards than anything I’ve heard these shills advocating. And unlike their efforts, the recommendations from the wide chorus of eminently respectable and well-regarded professional groups who backed ERT as a primary prevention for all post-menopausal women, had a huge impact on practice for over a decade. Of course, when they eventually did the RCT on the question, that 1.3:1 disappeared, as everyone who understood the weakness of observational studies should have expected.

The great ERT fiasco is what I use as the benchmark for opinion leadership in medicine gone astray, and by that benchmark, the shills are small potatoes. That should be seen as reassuring for the basic soundness of the empirical approach of the average clinician. They rely more on consensus than clinical research evidence, so they don’t pay much attention to individuals waving bits of research at them, whether those individuals are shills or otherwise. And consensus formation is beyond inappropriate influence except by understandable biases such as those that caused the consensus of opinion leaders to get ERT wrong, idealistic biases such as a desire to promote primary prevention, and Women’s Health, etc.

…. and the work by Ben Goldacre.

Dr. Tomkins, your argument is probably very much like the arguments deployed against the work of Semmelweis and Pasteur. Doctors of that period felt that their clinical intuition trumped the statistical and microbiological evidence.

Hidari, this is a big part of the rationale of evidence-based medicine. Doctors are subject to “illusions of control” (a technical term in psychology) which leads them to believe that their interventions are effective when the results are random, or even negative. It has happened multiple times in the history of medicine that some ineffective, or even dangerous practice, is widely believed by medical professionals to be effective. Sometimes this can contribute to inaction, as well as action. In Semmelweis’s time, doctors refused to believe that their own unwashed hands were causing childbed fever, since they believed their interventions were healing. Doctors even argued that Semmelweis’s valid claims were unscientific, since they could not see a a mechanism by which his aseptic practices worked. In our own time, unnecessary surgery is still a problem. It is less than a problem than 30 years ago, but this is largely due to pressure from insurance companies, and that based on actuarial evidence.

Humanity is still at the dawn of scientific medicine. The wide acceptance of aseptic practices is only 100 old and their incomplete practice is still a problem. There is still much blundering through illusion and forgetfulness. Dr. Tomkins, you argue for clinical intuition above technical evidence. The advocates of evidence-based-medicine argue just the reverse. I say both are valuable, and it is illusion and error to give either sort of knowledge too much weight.

Which, returning to the original subject, means I regard the filling of the medical journals with well-buttressed marketing literature as a very serious problem.

Glen Tomkins: “The rule of thumb is that, because of the potential for large imponderable biases inherent in observational studies, you don’t even begin to think about attributing causality until you get past 2-3:1. ”

IIRC, the ‘only RR’s of 2 or more count’ was something promoted by the tobacco – well, fraudsters.

Subject to correction, I believe that in practice, the only infectious diseases subject to eradication are those that can only be transmitted from one human to another.

Strictly speaking, I think you could also drive the nonhuman vector species extinct. But that’s hard to do with mosquitoes.

Hmm — it might depend on the life cycle. If a human-infecting phase is indispensable to the life cycle *and* you can get 100% herd immunity in the humans, then you might succeed in breaking the life cycle and driving the pathogen extinct without driving any of the host species extinct.

But this won’t work if other species can substitute for the humans in the pathogen’s life cycle. If the disease can promiscuously infect several species and jump from one to another in arbitrary order, you’re screwed. The other host species are always going to be a reinfection reservoir unless you drive them extinct or vaccinate every last one of *them*.

roac: Malaria has been eradicated in a lot of places; the first great triumph of the CDC as currently shaped was the eradication of malaria in the southeastern US, and there’s a fair amount of Europe which had malarial swamps a hundred years ago and doesn’t now. And not always only because the swamps are drained. The tertian and quartern agues that you find in Pepys and Shakespeare are reasonably likely to have been malaria, and people were using quinine against it in England as early as the late 17th century.

Yes, the malaria parasite has some really nifty biochemistry for injecting itself into cells and remaining encysted within the cells and insulated from the immune system – it’s remarkably evolutionary static biochemistry, and antibodies against bits of it are one of the better hopes for a malaria vaccine.