Its 1991. Two trains are heading toward each other on the same piece of track. One has a shiny badge that was new for trainspotters – Evidence Based Medicine (EBM). It had been assembled in part in Canada’s McMaster University by David Sackett and Gordon Guyatt among others and in the UK by Iain Chalmers and others. Richard Smith (photo), then the editor of the British Medical Journal (BMJ), was an enthusiast for this new marvel.
In 1980s, pharmaceutical companies were perceived like Sauron in Mordor as growing in strength. Organisations like No Free Lunch appeared. Physicians were hearing they should be wary of the Wooden Horse sitting outside the gates of medicine.
A medicine under siege breathed a sigh of relief when EBM arrived in town and seemed to solve the problem. Doctors were given weapons to fight pharmaceutical companies with. The vampires with their sales and marketing tactics would shrink back faced with the crucifixes and garlic of evidence held up in front of them.
It wasn’t put like that. More a case of in the face of the obvious, the lion would voluntarily lie down with the lamb. Bash industry? Not us. Just scientists here.
In 1984, the US pharmaceutical industry generated $24 Billion of revenue per year. A decade later, a single drug alone could generate $10 billion per year. The pharmaceutical industry had found a way to make the powder in their capsules worth more than its weight in gold.
Prozac which launched in 1988 was one of the exemplars of the other train rolling down the track toward the shiny new EB Medicine. If you wanted a decent drug – something that would be at least as strong as the pills we already had, and ideally safer, or maybe able to treat something we had not been able to treat before, this wasn’t that pill. It was less effective than already available and cheaper pills and there were good grounds from the get-go to think it and anything related to it might be hazardous.
Should have been no contest. EBM should have rolled right over Prozac.
Especially as Prozac was almost derailed before it had fully picked up speed. Pesky patients went and got themselves suicidal on it. An article appeared in the American Journal of Psychiatry in 1990 outlining 6 cases of people with suicidality that had emerged on treatment. The paper came from Harvard. One of the authors, Jonathan Cole, was just about as distinguished and at the same time sensible as anyone could be. Thousands of patients lined up across the US claiming they had become suicidal on the drug.
The company had three responses – these are anecdotes, its the disease not the drug and it’s all just being rustled up by the Church of Scientology.
The trains met in September 1991 – thanks to Richard Smith.
Central to Lilly’s efforts to defend Prozac was a meta-analysis (a new term then – part of a lexicon EBM was introducing) of the clinical trials of Prozac, looking at rates of suicidal events on the drug. Very few journals had seen data handled like this before.
Whether the novelty of an article like this meant it didn’t work for the first journal it was sent to, or the very obvious increase in risk on Prozac, it was turned down and ended up with the BMJ.
The BMJ reviewers demurred at the proposed title which suggested Prozac reduced the risks of suicide. The claim was made on the basis that Hamilton Rating Scale scores fell on the drug but ignoring the fact there were more suicidal events on Prozac.
Looking at the paper later published in the BMJ, with Charles Beasley as first author – HERE – in placebo controlled trials there were 5 suicidal events in 1398 patients taking Prozac and 1 in 645 taking placebo. More on Prozac – but it was not statistically significant – which everyone was told meant there were no suicidal events on Prozac in other words. Move on – nothing here to look at
However looking at the small print the placebo suicidal event happened in the washout period so there were in fact no events on placebo and the risks on Prozac were statistically infinitely greater than the risk of placebo. It doesn’t look like the reviewer can have looked at the small print.
Richard Smith (for whatever reason) certainly didn’t look closely. He had in his hands a major company signing up to embrace the EBM he was promoting at every turn. If all companies ended up following (being forced to) this example, the world would spin more happily on its axis from here on.
Oh Happy Band of Brothers, you will be able to tell your children that you were here with me and the BMJ on St Candida’s Day.
The article came out in the September 20 1991 edition of the BMJ. coinciding with FDA hearings on Prozac and suicide. Given the patient groups springing up all over the place in the wake of Prozac’s launch, with hundreds of reports to FDA of suicides or suicidal events, FDA was forced into a public hearing.
At the hearing, company people and company experts (people), notably Charlie Nemeroff, were brandishing the Beasley paper – look it says here in the BMJ there is no risk from Prozac. Which are you going to believe the anecdotes or the science. The plural of anecdote is not data.
The Little People
The FDA committee recognised the compelling testimonies from the little people whose husbands or children or parents had blown their brains out or who had died in other horrific ways. A few weeks after starting Prozac, Joseph Wesbecker, who featured in the news last week, two years and one week before, had killed 12 at the Standard Gravure plant in Kentucky, before blowing his own brains out.
But their hands were tied. There was no basis in the science to say Prozac had caused this.
It wouldn’t even be a good idea to warn about a possible risk because that might put people off seeking treatment and the benefits they could get from Prozac.
Prozac sales had been falling. They recovered after the FDA hearing.
The effect on medicine and the rest of us was far-reaching. Journals like the BMJ stopped taking articles like the Teicher and Cole paper outlining 6 convincing cases of suicidality induced by Prozac. FDA had just ruled this is not evidence. Journals might be sued for publishing them – besides which they made no money from them.
A premium was put on ghostwritten clinical trials with no access to the trial data and even more on meta-analyses – companies paid vast fortunes for reprints of these making BMJ and other journals profitable indeed wealthy if they wanted.
Richard Smith could have demanded access to the raw data in exchange for publication in BMJ. He didn’t.
But he got a sales-force for EBM. Companies now encouraged doctors to prescribe according to the evidence and to stop paying heed to the appearances of people in front of them or to what they say. Everyone knows that patients and doctors can’t work out what drugs do – only controlled trials can do that.
Patients became invisible – just as doctors and everyone else in healthcare did too. If practice is going to be scientific it has to be according to the evidence, embodied in guidelines which dictate what drug is given in what sequence.
If this is the case of course who needs high cost prescribers. We can have nurses, pharmacists, AMPs or whoever – told keep to the evidence based guidelines or you get fired.
And we can replace everyone with locums. Who needs relationships if whoever is there sticks to the “Evidence”.
As a result it now takes 30 years to get recognition of the common problems on patent drugs cause.
Clinical Science is incompatible with a top-down-archy. Hierarchy is not the right word here – as hieros means sacred or moral. Its Gallilean or nothing.
The first task is to work out whether the drug this person took has caused the effect they now mention or not. Having reviewed all the evidence if we both think it has and this is at odds with what the Evidence Base appears to suggest, the second scientific task is to explain the discrepancy. In the case of the SSRIs and suicide, there was no discrepancy once we got to see the clinical trial data. If it is to have any credibility, EBM has to be Data Based Medicine.
Waving or Drowning
Well done Richard on a Bruce Willis award.