What Happened to those Who Were Suicidal in Study 329? And to the Learned Intermediaries?



The Doctor exhibited 1891 Sir Luke Fildes 1843-1927 Presented by Sir Henry Tate 1894 http://www.tate.org.uk/art/work/N01522
“The Doctor,” exhibited 1891, by Sir Luke Fildes

In May 2014, the RIAT team asked GSK what the children who became suicidal in the course of Study 329 have since been told.  (Marty Keller’s “take” on this is at the bottom).

The consent form says that anyone entering the study would be treated just the way they would be in normal clinical practice.

In Study 329, the children taking imipramine were by design force titrated upwards to doses of the order of 300 mg, which is close to double the dose of imipramine given in adult trials by GSK or in normal clinical practice.

In normal clinical practice it would be usual to inform somebody who had become suicidal on an SSRI that the treatment had caused their problem.

  • It is important to the person’s image of themselves, that they are aware this problem might have come from the drug rather than from themselves.
  • It is important that they be told that they are likely to react in the same way to any other serotonin reuptake inhibitors – including some antihistamines, isotretinoin and some antibiotics and that they should be cautious about such treatments in the future.
  • It is important the patient’s family be warned about this adverse event as some blood relatives will be more susceptible to commit suicide on an SSRI than the rest of the population might be.

Responding to our letter, GSK’s Dr James Shannon made it clear that 20 years later the company have still not informed any of the participants in Study 329.

One reason he offered was that it had only recently been agreed these drugs posed risks. [One of the features of GSK’s response to the published study appears to be a public acceptance that SSRIs do cause suicidality in at least this age group].

But the key reason for not doing so that he offered was that:

As I have mentioned in my earlier letters, it is standard in clinical trials carried out according to good clinical practice guidelines for our trial investigators and treating physicians to be responsible for patients’ medical care during and after a trial.  This would include the management of any adverse experiences that arise during the trial. Being closest to patients’ medical histories, they are best placed to do this and we are confident of their commitment to provide the care patients need.

This may be the first recorded appeal to the use of the Learned Intermediary doctrine in clinical trial settings.

One lesson of the 329 story is that without access to data, the learned intermediary doctrine is supremely dangerous for patients – and for doctors.

Learned What?

The notion of a learned intermediary arose in the 1950s when the first reliably effective drugs emerged, when an ‘ethical’ pharmaceutical industry distinguished itself from a patent medicines industry by advertising to physicians only, when Congress decided to make all new drugs available on prescription only.

Doctors then were seen as a bulwark against pharmaceutical advertising, by inclination and by training more likely to resist the pressures of advertising than the rest of us.

This was a Marcus Welby world, in which the forerunners of today’s mega-corporations were still divisions within chemical corporations and were run by doctors or scientists rather than by marketing people.

Just as prescription-only status is a police function, the idea of a Learned Intermediary is a legal notion.  Neither have anything to do with the practice of medicine.

The term ‘learned intermediary’ was first used in 1966 in Sterling Drug Inc v Cornish, 370 F.2d 82 (8 Circuit):

‘we are dealing with a prescription drug rather than a normal consumer item.’ In such a case the purchaser’s doctors is a learned intermediary between the purchaser and the manufacturer. If the doctor is properly warned of the possibility of a side effect in some patients and is advised of the symptoms normally accompanying the side effect there is an excellent chance that injury to patient can be avoided’.

Based on the doctrine, it is held that:

  • Warnings about a medication’s hazards need only go to physicians because they are the only people that know both a particular patients medical history as well as the risk profile of the drug being prescribed.
  • That directing routine prescription drug information through the doctor in this way preserves the physician patient relationship from outside interference.
  • That the complicated medical terminology necessary to explain the risk benefit profile of prescription drugs is difficult for ordinary patients to understand.
  • It is more effective for drug companies to have to communicate with physicians only rather than directly with all potential patients.  [This assume drug companies will tell physicians the truth].

329 & Learned Intermediaries

The doctrine is premised on the notion that the physician is an objective intermediary who will draw an independent judgment about the best course of treatment for his or her patient.

Two factors have been put forward as compromising the objectivity and independence of doctors – gifts and Direct to Consumer Advertising (DTCA).

But 329 opens up a whole new dimension.

When the learned intermediary doctrine was introduced most information about drugs came from doctors writing up their experience in case studies or running trials where they were in possession of the data and they wrote the manuscripts about what the trial had shown. There was some sales pressure but almost no marketing pressure.

  • From the 1980s, companies ran trials, increasingly hiring second rate investigators to tick the protocol boxes, and latterly locating trials in third world settings where there can be no guarantee the patients exist.
  • From the 1980s, an increasing proportion of the clinical literature about on patent drugs has been written within companies or by ghost-writers.  That proportion is likely now over 90%.
  • From the 1980s, companies have sequestered clinical trial data so there is no independent oversight of what this data shows.  Not even FDA get to see all the data.
  • From the 1980s, the pharmaceutical companies were on their way to being the profitable corporations on earth, managed no longer by doctors or pharmacologists but by business men with a background in marketing.

In this new world:

  • Few doctors can distinguish between sales and marketing.
  • Few doctors appreciate that RCTs, adopted in 1962 to contain companies, are now the major marketing tool of companies.
  • Few doctors appreciate how adept companies have become at marketing diseases as a means of selling their drugs.
  • Few doctors appreciate that companies make most money from marketing risks such as marginal elevations of cholesterol levels or reductions in bone density, where if the patient takes a medicine they take on risks with little likelihood of benefits.
  • Few doctor appreciate that Safe and Effective in Keller et al 2001 means Safe for GSK and Effective for GSK not safe and effective for their patients.

While there is something to be said for a Learned Intermediary in medical extremis, when a person’s judgment may be compromised, there is much less to be said for having a third party make a judgment call that requires you to accept risks they would not personally accept, for benefits that are unlikely.

Cause and Effect 

When the Learned Intermediary doctrine was introduced, doctors were trained in how to determine cause and effect in terms of the adverse effects of a medicine.

The way doctors assessed such events through to the SSRIs and Suicide was in line with what the Federal Judicial Reference Manual outlines to this day. Broadly speaking if exposure to a drug produces a problem (challenge) and the problem clears up on stopping the drug or reducing the dose (de challenge) and reappears on re exposure to the drug (re challenge) this is definitive evidence that the drug can at least cause the problem in some of those who are exposed to it.

However under company marketing pressure most doctors have been persuaded that such evidence is anecdotal. That the only evidence of cause and effect that counts is the evidence that comes from controlled trials. For 25 years doctors across medicine have been systematically educated to override the evidence of their own eyes.

  • As a result major problems that came to light quickly in the 1960s may now take 10 or 20 years to be accepted by the field as caused by treatment.
  • As a result drug induced death is now a leading cause of death.
  • As a result, even in the face of Black Box Warnings a majority of doctors may still believe there is no evidence that these drugs can cause this problem.

This extraordinary situation has arisen because in their defense of Prozac 25 years earlier, Lilly deployed selected data from sequestered RCTs and a sophisticated understanding of doctors, to counter compelling clinical evidence that Prozac could cause suicide.

This is caught in this quote from Leigh Thompson, who was coordinating Lilly’s efforts in 1991, contrasting the fate of Prozac with that of the 1980s Oraflex (Opren):

Today at PSC was Medical’s finest hour. Dave Thompson and Gene Stap told me that it suddenly gave them a glimpse of how far medical has come and the vision that they knew (about global databases, super handling of ADE, proactive excellent relations with FDA, complex analyses and presentations made simple, DEN, GPT etc) but had never really had burned into their brains the elegance and mastery of the complexity!

So many of us were not here for the Oraflex , Moxam, etc crises, that it is very hard to measure the progress over the last few months on so very very many fronts.

When you battle the media and politicians, the only thing that counts is the first word. The rebuttals are always on the last page and forgotten. You have to get out front and enlist your allies. The rapid flights to Boston to visit Teicher, the trips to FDA, the consultants coming in, the huge complex database, having so many large trials, the ability to quickly perform elegant analyses, DENs mastery of ADEs, have all come together in a significant effort.

I’ll try to give a global overview of our past (Oraflex and Moxam especially) and our present and our future (with Mobius, Scientology etc after us) tomorrow at DEN. Please pass on my congratulations and profound thanks to your spouses/friends for tolerating your extra work/pressure and to those colleagues whom I have left off the list of addressees in my rush to get out this note.

I’d like to have some buttons or mementos of other kinds made with a logo along the lines of: “I saved Prozac.” Suggestions please for design, memento and words. (Exh 10 in Deposition of J Potvin in Fentress Vs Eli Lilly).

From a company risk management point of view, an aggressive management of adverse events that might jeopardize the sales of a Flagship Brand and in so doing sink the company is a No Brainer.  The trouble is it plays straight into a medical blind-spot – Marty Keller and Stan Kutcher and the rest of us would prefer not to think that something we have done might have injured our patient.

Control and Consent 

Since prescription only status was copper-fastened in place, and since the emergence of the Learned Intermediary doctrine, legal cases in the 1950s tackling the use of radical mastectomy and other drastic treatments for breast cancer and the use ECT within mental health gave rise to a Doctrine of Informed Consent.

Informed Consent is incompatible with notions of a Learned Intermediary whose role includes deciding whether to withhold information from a patient based on his judgement about what is in that patient’s interests.

When the Learned Intermediary doctrine arose it was not unreasonable to think that while patent medicine makers found no difficulty in creating an advertising industry in order to communicate with the general population, there were legitimate grounds to think it might be a problem to convey complex information about an entirely new class of compound that contained ingredients that for the first time ever could save lives and restore function but could also kill.

A case could be made for engaging doctors in communicating that information.

But with the development of the internet, the extension of education, and 75 years of familiarity with modern medicines, the pharmaceutical industry appears to have little problem communicating with the population at large.

And so, there no longer appears to be a good reason to exempt drug manufacturers any more than other manufacturers from a duty to warn – and a duty to make public the data behind their claims.

Instead of helping patients, the Learned Intermediary doctrine now helps companies.  It allows companies to keep their data under wraps. It shifts liability from a careless manufacturer onto an intermediary who is responsible for distributing a defective product to a vulnerable citizen.

He (your doctor) is coached by companies in the arts of persuasion to play on his patient’s (your) anxieties so that they (you) take medications they (you) would never take if left to themselves.  He was once skeptical of the benefits of drugs and a bulwark against unwarranted treatment, but many of his patients (you) are now more skeptical than him even though they (you) don’t quite know (you’d never guess) the extent to which his recommendations to take a treatment are based on sand.

The fact of prescription only arrangements allows companies to sell to doctors only. In a world where medical education does not cover how companies market medicines, we have produced a generation of doctors more susceptible to marketing that the most designer label addicted adolescent.

Marty Keller & Summary Judgement

Pharmaceutical companies make medicines.  Because doctors use medicines and are trained in medicine, they might appear to be equipped to assist their patients in making decisions about those medicines.

This was true in the 1960s when doctors’ clinical experience (not RCTs) meant they knew more about taking risks with chemicals than their patients did.

Medicines are chemicals that come with information.  The chemicals are unavoidably risky.  The information component of a medicine and the culture about using chemicals in medicine was commensurate with those risks in the 1960s but the information and the culture have both been degraded and now systematically conceal risks.

Worse again this information is portrayed as objective and scientific. It has become the primary determinant of prescribing, over-riding the natural caution of both doctors and patients.

The problem that sales techniques such as gifts pose in compromising the doctor patient relationship is minor compared with this.

If doctors are now truly to function as Learned Intermediaries, it would be by informing their patients and the Courts that the bulk of the academic literature, especially that in the most distinguished journals, cannot be believed.

Sitting in the midst of a Maelstrom in 2004 when New York State were suing GSK for Fraud, faced with demands to “modify” Study 329, on June 13 Marty Keller emailed some of his co-authors to get a united position about what they would say to GSK about the modification:

that [it must be] 100% clear in this paper that there is no way to read it and think that 329 is being criticized and that it was not written with complete integrity and accuracy given the data we had and should have had…..  We also want it to be crystal clear that any new data or analyses, case report forms, narratives etc. you have worked with since 329 was published, was not made available to the 329 investigator’s by SK, otherwise we could look foolish, naïve, incompetent or “biased” (the most likely accusation that will be made) to present things in a way that was favorable to SK, disregarding our responsibility to the proper scientific method,  to the public, children and their families.

So as MK & GSK see it, while the Learned Intermediary Doctrine rules, he and his colleagues have a responsibility to their patients.  Someone needs to do the right thing by the Paxil 12.



  1. It should be clear, that many of those children in this so called clinical trial died. The hapless victims of a fraud, whose only purpose was to deceive the public and regulatory agencies, into selling a drug, that would result in more deaths.

    It may be easier to deceive when deception is cloaked in scientific jargon. And then there is always the priest like status of medicine, that seems grant a religious like suspension of the critical faculties of reason, in a large segment of the public.

    Then there is big money, albeit blood money, that offers a certain amount of protection to organized criminals through legal strategies, until the evidence of their crimes become overwhelming.

    It is a process, that attempts to sanitize mass murder.

    Report comment

  2. Has anyone ever tried to contact the subjects of the study? I know their details are not available publicly (obviously) but there are sometimes ways to track them (e.g. via ads requesting contact if they’re willing to tell their stories).

    Report comment