This continues a Something Happened Series,
A Western moral order fractured between Luther’s nailing of his Credo Ergo Sum, I believe therefore I am, to the door of Wittenberg Cathedral in 1517, and Descartes’ Cogito Ergo Sum, I reason therefore I am, in 1649 the year of Charles I’s decapitation.
Justice and benevolence were central to the moral order monarchs held in place – bureaucracy is central to the moral order now. The interaction between human beings in Justice settings differs from that in bureaucratic settings. Law and regulation are different things. Justice is not an application of bureaucratic rules.
While a lot what now happens in courts is about whether we have infringed rules and while one side to an interaction in court might be able to bring more resources than the other to the interaction, and something hinges on the exercise of human wit on the part of advocates for either party, what we celebrate when we celebrate justice is the ability of a judge or jury to reach beyond the argument and exercise judgement, or discretion – their ability to make a diagnosis that moves things forward in the right way.
The trial of Walter Raleigh in 1603 produced a moment critical to our ideas about justice. Raleigh was convicted of treason, and later executed – on the basis of claims made by third parties who did not appear in court and could not be cross-examined. The judiciary recognised a problem and put in place a Hearsay Rule – evidence would not be admitted if the people offering it could not be cross-examined. Justice involves an exercise of judgement grounded in an interaction between people rather than an appeal to technique.
In contrast to justice, bureaucracy has a set of procedures that ideally are applied without discretion. It aims at sidelining judgement.
Medicine was one of the routes through which benevolence has traditionally been delivered. It made moral sense to tend to the ill and heal where possible – and later made what would be called economic sense as having people fit and able rather than drawing on welfare should all things being equal enrich a country. It also made sense when disorders were contagious – treating you might save me and my family.
Until recently, medical thinking was essentially the same as judicial thinking. Doctors faced with patients able to be cross-examined came to a view as to what was likely happening and both patient and doctor hoped the judgement call, the diagnosis, worked out.
When modern drugs came on stream in the 1930s, detecting adverse effects, Drug Wrecks, were one of the easier medical jobs – easier than diagnosing many illnesses. If a problem happens soon after a drug is given, and clears up if the drug is removed, perhaps reappearing if it is reintroduced, or varying with the dose of the drug, and if there is no other obvious way to explain what has happened, then it makes sense to diagnose the drug as the cause. This is still the standard view on how to go about establishing cause and effect in the case of drugs in judicial settings as laid out in the Federal Judicial Reference Manual for applying science to drug induced injuries.
When drugs later became precious commodities, worth more than their weight in gold, that changed. Before that point, if a doctor prescribed us a drug, there were only two of us in the room – the doctor and us. After that among others there were company marketing departments whose job it is to ensure our doctors don’t have a thought in their tiny little heads other than the thoughts put there by them or their competitors (this really is their view of doctors). The others included bioethicists, medical journals, medical academics, and politicians who were all singing from the same song-sheet as pharma.
The problem we have now is not caused by pharmaceutical company marketing – the lunches, the trips to conferences, the glad-handing, the making of second-rate medics into opinion-leaders. Pharma actively want you to think this glad-handing is the problem and are pleased when their critics rant on and on about conflict of interest.
Central to our difficulties is the bureaucracy we thought we had tied Pharma up in with the 1962 amendments to the FDA Act and especially a then new and poorly understood invention – randomized controlled trials (RCTs) – which were built into the regulations governing the licensing of drugs.
RCTs don’t work for the purpose intended, which was as a means to ensure Pharma could only bring drugs that worked on the market. They work for Pharma – nothing better has ever been invented for hiding Drug Wrecks.
And we have no easy back from what we did in 1962 – it would be easier to get doctors to believe the earth was flat that to get them to accept that RCTs are the source of their and our problems.
It’s not difficult to get doctors and pharma critics to believe conflict of interest is an issue, a little harder to get them to accept that ghostwriting of trials and sequestration of trial data is problem, but it will be like getting the Pope to give up Xtianity to get them to forsake RCTs.
The ghostwriting and data sequestration are a problem but pretty well all doctors and others, including the Chair of NICE, Chief Medical Officers in the UK and US, Ministers of Health in US, UK and Europe, the BBC in all its manifestations, New York Times in all its manifestations, the Pope and others, while accepting this is a problem can seemingly continue taking the sacraments as though there was nothing wrong, and where once they encouraged us to do so as well, they now seem to be gearing up to force us to do so.
Both JAMA and the New England Journal of Fake News two weeks back came out with articles claiming that being anti-statin was the same as anti-vaxx and that mistrust of the Fake News they publish (they call it Science even though they know its ghost-written and there is no access to the data) is a threat to the physician patient relationship.
Conflict of Interest was the stick critics (let’s say Puritans) used to beat the industry with some two decades ago but industry is no more bothered about this than Donald Trump by an association with Stormy Daniels or Jeffrey Epstein.
With Puritanism making little headway, a few took to pushing for access to the data. This however feels a touch like Catholic or Protestant pastors in Germany in the late 1930s and early 1940s making it clear they were not entirely happy with an elimination of the unfit – a move that was too little and too late. If industry are forced to grant access to the data, there are ways to ensure what becomes the data delivers the message that industry wants.
But even beyond this, the deeper problem here is the declaration that RCTs are infallible – that they offer gold-standard knowledge. Sure there are problems brought about by industry use of RCTs, many critics will say, but RCTs themselves are the best source of knowledge we have. This is the problem.
History of RCTs
The first RCT was of streptomycin in tuberculosis. Prior to that there had been a standard clinical evaluation of streptomycin in tuberculosis that produced a much more accurate picture of this drug than the later RCT. Both showed the drug worked. But the standard evaluation also showed that patients became tolerant to streptomycin pretty quickly and some went deaf.
The RCT showed randomisation could be used as an aid to evaluating drugs but it would not necessarily get as good answers as standard clinical evaluations.
The first RCT of a drug before it came to market was done on thalidomide which sailed through this trial and came out on the far side as safe and effective.
Still the mantra took hold that no doctors would ever be able to work out if a drug worked were it not for RCTs. This gets repeated every hour of every day even though every hour of every day, patients, or doctors or both combined decide if a drug is working or not and medicine simply would not be possible if one or other of them weren’t right pretty well always.
We might say “To err is human, To really foul things up needs an RCT. And we are totally screwed if RCTs are given an infallible status.”
The idea is that RCTs deliver objective knowledge, which doctors on their own or patients can’t. Compared with clinical judgements, RCTs aren’t objective. The idea that they are is a myth. They are mechanical and impersonal.
One basis for their supposed objectivity lies in Ronald Fisher’s first thought experiment involving randomisation in 1925 which he expressed in terms of statistical significance. Fisher’s original idea was that statistical significance would indicate we knew what we were doing so well that only chance could get in the way of the outcome we predicted. But when a doctor today figures on giving an antidepressant to someone, there is no better than a 50-50 chance it will suit them (never mind work). Despite a statistically significant result in trials, giving an SSRI is no better than a crapshoot.
Imagine walking into an emergency department with a broken arm, being told they are randomly applying plaster casts to broken limbs and ending up with a cast on your leg. This RCT would show randomly applied casts beat placebo (one in 4 times the case would be on the right limb versus 0 in 4). But to practice medicine this way would be obviously nuts. This however is increasingly the modern practice of medicine.
In response, some defenders of RCTs – and most defenders are non-industry folk who figure RCTs are the one true way to knowledge if we could just get industry’s hands off them – will say pooh to statistical significance, we use confidence intervals.
Confidence intervals come from efforts in astronomy around 1810 to come up with a way to decide whether the differing measurements we ended up with came from two different stars or one star imprecisely measured. The bright idea was that measurement errors would cluster predictably around a mean – a distinguishable second star would fall outside this cluster.
This works for stars but not for human disorders, where diabetes, depression, back pain, breast cancer, parkinson’s disease and pretty well everything else can be forty different conditions rather than one. And it works even less well for trials of drugs, where even if the condition were one my response to a beta blocker might be exactly the opposite to yours.
The Gold Standard
The only reason RCTs are a gold-standard is that they are the standard through which industry makes gold. They work for industry – and not just because industry work them. Within 3 years of RCTs being built into regulations in 1962 as the way to keep ineffective drugs off the market, company salespeople were encouraging doctors to prescribe in accordance with RCT evidence – RCTs that the companies had not run. Companies did a lot to create and have been the biggest promoters of Evidence Based Medicine (EBM) ever since.
RCTs are not totally worthless. They are like a microscope or telescope – helpful in seeing things that are not obvious to the naked eye such as how many people on active treatment end up dead compared to those on placebo when treatment extends 5 years and needs thousands of people recruited to the trial to spot a very distant or miniscule difference. But just as you wouldn’t use a microscope or telescope to work out who it was you were talking to, any more than you’d have this kind of lens on your eyes when walking down the street, so RCTs can get badly in the way of dealing with someone right in front of you – or with yourself.
All of the above applies to the benefits of RCTs, which is where RCTs are supposed to be particularly helpful in shielding us from bias. Their intense focus on one thing to the neglect of everything else, is clearly risky, but this might be a risk worth taking if they got the right answer to the question of whether there is a benefit to this drug but more often than not they don’t.
When it comes to adverse events, Drug Wrecks, things get exponentially worse. We’ll deal with this next week.
But cutting across the effect of RCTs in helping us get a handle on either the benefits or the hazards of treatment is the effect of RCTs on our confidence in ourselves and our judgement calls.
And this is of a piece with a removal of judgement from everyday life that picked up pace in the 1960s.
This sequence of posts would be interminably long if our turn to procedure was explored in detail but in brief what gets called neo-liberalism emerged at exactly the same time as neo-medicalism – in the mid-1970s. Neo-medicalism is typified by the operationalism of DSM III – with blood pressure, blood sugar and other measurements playing the same role in the rest of medicine.
Both neo-medicalism and neo-liberalism embody thermostat functions – simple algorithms – that reduce complex problems that should call for judgement to simple functions – if X do Y. If the supply of money grows to a certain point, cut it – regardless of the damage this will do to a country and its people. If someone can tick 5 out of 9 boxes, they have depression regardless of whether the boxes they tick all stem from a flu or a pregnancy.
This is a bureaucratic – procedure-based – approach to complex problems. Judgement and benevolence are replaced by a slot-machine – if 3 lemons line up you are entitled to an antidepressant without anyone intervening to ask whether this is “wise” or not, “honorable” or not – any call to judgement based on any of the virtues (pagan or religious) that used to guide us at important moments involving justice or benevolence or other things is sidelined.
RCTs fit into this bureaucracy perfectly and did more than anything else to ensure a triumph of neo-medicalism in the 1980s. Its this that I think the Roy Porter review glaringly misses. Porter and others, especially those writing history, were very aware of neo-liberalism and were at the vanguard of those raising concerns about it but they missed its manifestations in medicine in a manner that suggest they and perhaps others never understood what neo-liberalism is – and to this day the word is a piece of jargon that few people ever try to define.
Understand neo-medicalism and how to get to grips with it and we might understand and be able to roll-back neo-liberalism.
This too long post will continue next week – returning to the decapitation of Walter Raleigh and why this matters now more than ever before.