In Epidemiology of Autism Spectrum Disorder on RxISK, I mention that for me a key paper in a reference list of 27 references was one by Ann Bauer and colleagues, who recognized limitations to the epidemiological data but nailed their colours to the mast in saying that there was sufficient evidence here that pregnant women should be warned of a possible link between acetaminophen – paracetamol and neurodevelopmental delay.
In the post, the approach Bauer and colleagues took was contrasted with a GlaxoSmithKline (GSK) approach that dug deep into technical details and had multiple procedures that appear designed to give an appearance of objectivity. When challenged on key points, the GSK spokesman said, as many in the industry say, that he and the process in place was not in the business of coming to value judgements. Their task was to amass evidence.
This choice of words is wonderful. Getting involved in values sounds like being sucked into something religious rather than sticking to science and of course we want our scientists to stick to the science.
Value Judgements
Many colleagues whose work I admire and company I enjoy seem to see things is a similar way to the GSK dude cited here. They aren’t keen to get involved in issues I take positions on. This ‘attitude’ is most apparent when it comes to possible involvement as an expert in a legal action. Their body language, and tone of voice, suggest they see this as muddying their scientific hands.
Being a legal expert, even on the side of patients injured by drugs, is as bad a conflict of interest problem for many of the ‘good’ anti-pharma guys as being consulted by a pharmaceutical company. It has a going over to the dark side quality to it. People will be able to say of you that you only have these views because you are being paid – and of course some people will say this.
A.I. sides with industry and with my friends. Putting a current Healy legal report into ChatGPT and asking it what the problems are with it, turns up a 4000 word, 8 page ChatGSK response nearly as long as the core expert opinion section of the original report.
Throughout the report, Dr. Healy’s tone betrays a strong advocacy stance. Phrases like giving the patient “a poison in the hope of bringing good out of its use” to describe an SSRI, or calling the medical literature “ghostwritten… adverts”, are inflammatory. Such language is not typical of a dispassionate expert witness [and] could raise questions about the expert’s impartiality.
An effective expert report should help the judge by illuminating facts.
All the above issues – scientific overstatements, legal missteps, inconsistent reasoning, and intemperate tone – ultimately affect how persuasive and admissible the report is as expert evidence. An expert opinion must be grounded in recognized science and delivered with objectivity and care to assist the trier of fact.
There are many amusing insights here. The bolding is Chat GPTs not mine. I’ve picked two bits out. One to illustrate the emphasis on tone – scientists should be dispassionate, disinterested, dealing with figures not people. Does the bolding suggest ChatGPT’s tone is slipping?
The second is the word fact. And especially the phrase ‘the trier of fact’.
Purity of tone comes with problems when we are asking to participate in trying facts.
Scientia and Science
At the end of last year, Britain’s General Medical Council struck off a doctor for claiming the Covid Vaccines Lacked Effectiveness and Safety and that it was astonishing that so many healthcare folk were supporting vaccinations. The doctor, GMC said, was offering opinions rather than facts and inviting the public to believe his opinions because he was a doctor. It was in the public interest that he be stopped from making claims like these implying they were facts rather than opinions supported only by his status as a doctor.
Before we got science c.1660 we had Scientia – a Latin term for pure knowledge, as is found in logic and some mathematical theorems.
Logic is an operational, algorithmic, procedure that enables us to come the same objective conclusion in a specified situation that everyone else for all time will come to.
In contrast, daily life from time immemorial involved having opinions, which were viewed as a lesser form of knowledge, highly subjective and ephemeral – even though they were critical to survival.
After the rise of science, philosophy continued to seek eternal truths. One path to this was logical positivism. (Someone hopefully will be able to comment on the interface between A.I. and logical positivism).
In contrast to Scientia, how does science, which is constantly subject to revision, get to objectivity or truth? The answer is it borrows from Law, which, some decades before science, decided the route to objectivity was to get a jury of 12 people with disparate opinions, beliefs and backgrounds to attempt to arrive at a consensus on observables in front of them.
This consensus created a fact. Facts are open to revision if new evidence comes along.
There were some stumbles on the way. Walter Raleigh was executed on the basis of what came to be called Hearsay – testimony jurors could not observe being delivered and cross-examined firsthand. Pharmaceutical company assays (studies) are a modern exemplar of Hearsay. The Rules of Evidence in legal cases since Raleigh ban Hearsay.
For nearly 3 centuries, science aimed at getting juries with disparate opinions to attempt to achieve a consensus on demonstrations happening in front of them. Replications, and tweaking of the experimental machinery to see what happened, were encouraged and accepted as a route to revising views as to which facts were established.
The peer reviewed literature was designed to allow others on the far side of the world to repeat the demonstration, making for an enlightened universality.
Medical Science
After 1800, clinical practice grounded in a medical model, became scientific in this sense. Public health initiatives, case conferences and clinical consultations (ideally with both doctor and patient participating) became venues where facts were tried. If the consensus was correct, the patient might have a better chance of surviving. If wrong, the patient might die prematurely. Progress in medicine, especially the dramatic improvements in life expectancy after The Great War, linked to public health and surgical developments, suggested medicine was on the right path.
In the 1950s and 1960s, a set of new pharmaceuticals brought two operational elements into the mix – randomized trials and a set of statistical procedures. These operations claim to offer an objectivity but it is not the objectivity of science.
In the hands of pharmaceutical companies, these operations are spun as science and contrasted with anecdotes (opinions, likely misinformation) but this ‘science’ is better viewed as Scientia with the proper contrast being Facts not anecdotes.
Scientia (RCTs and statistical tests) generate non-existent averages. There may offer an element that allows drugs to be licensed but they tell a doctor nothing about how to treat the person in front of them. They operate in a virtual world rather than the real one.
Few if any of the statistical operations have demonstrated a solid grounding in clinical reality. They may assist us but their use should always depend on an act of consensual judgement to decide whether their use in particular situations helps or hinders.
These operations intruded into clinical care and became a serious, and at present unresolved problem in the early 1990s. In 1991, the Evident capacity of SSRI antidepressants to cause suicide in some people, using traditional clinical methods to establish causality, lost out to some extent at least in clinical settings to the new Scientia.
Can Medical Science be Saved?
Juries and judges remained somewhat immune to the argument that there are no grounds to say that this man murdered his wife, just because no clinical trial has ever or will ever show that on average husbands murder wives, even a trial or epidemiological study restricted to men with a criminal records, legal and financial problems and wives that no-one likes.
Any non-medical and non-expert trier of fact on a jury in a legal setting will ask for the details of the case.
Guided by regulators (bureaucrats), hemmed in by threats of legal actions and losing their jobs, when it comes to trying facts in clinical settings doctors unfortunately are more inclined to practice according to the Scientia and less and less inclined to engage with their patients in trying facts.
ChatGPT doesn’t seem to understand this.
No legal system to date has addressed the fact that a great deal of the evidence presented by experts in cases involving pharmaceuticals is Hearsay and derived from an operational logic that is not designed to help a trier of fact either in a clinical or legal setting.
When it comes to drug-induced injuries, justice requires someone, informed by medical experience, to help a jury see how many doctors shaped by their experience are likely to read certain clinical signs and symptoms. Increasingly, however, this needs to be supplemented with some account of how we got to a position where what might be very Evident to a jury just wasn’t evident to the doctor treating the patient.
See
Our current problems stem from importing medieval operations stamped with a guaranteed-to-offer-objectivity imprimatur. Legal and clinical systems need to return to working with what is Evident in the case in front of them rather than what gets branded as Evidence. A jury of us and our peers is much more likely to be fooled by Evidence, especially Evidence designed to fool us than by what is Evident in front of us.
We also need a chronicling of how we managed to turn the clock back on medical science.
Pharmaceutical companies specialize in commissioning other companies set up to manufacture statements appearing to offer a professional consensus from this or that medical body as to why this or that pharmaceutical sacrament (something that can only do good and cannot harm) should be consumed by the faithful as assiduously as possible.
In contrast, it is very unusual these days to see a bunch of clinicians and researchers come out with a consensus statement on a treatment’s hazards and figure women need informing about these possibilities so that they will be in a better position to engage in a trying of the facts when needed.
This is what led me to pick out the Bauer and colleagues consensus statement as exemplifying what to my mind we need more of – not just courageous researchers but also journals willing to publish statements like this.
We are up against a semi-Religious War on Science launched by forces whose slogan is Science will win
We need to hold onto our Values. Science is not value-free. It values data and people are the data in medicine. The figures from clinical trials are 2-D abstractions from the data. If it is not possible to interview the people from whom the figures comes, we have no context with which to understand the figures.
There is another value that lies at the heart of medicine and that medicine contributes to modernity. Science fostered the enlightenment and its mission statement Sapere Aude – Dare to Know.
Medicine challenges us to Fidere Aude – Dare to Trust Me.
In recent decades doctors have lost the ability to recognise and work with someone who brings them an insight that is not in their books. While we want some doctors to remain alert to the possibility that some people are trying to game the system, medicine cannot work if most of the people who bring a problem to us are not believed.
As Challenging my Doctor to Disclose says – a key question for those of us going to a doctor these days is likely to be – Why Do you Not Believe Me?
When challenged on key points, the GSK spokesman said, as many in the industry say, that he and the process in place was not in the business of coming to value judgements. Their task was to amass evidence.
ChatGPT doesn’t seem to understand this.
The trier of fact. Totally gives the game away. AI will always betray it’s ignorance. It’s like receiving a scam email from Amazon.
Two examples come to mind on Judgement of Values.
Last year there was an article in the Daily Mail, about the death of Thomas Kingston. It centred largely on large debts of his company. BUT in the article it named David Healy as the expert witness who was of the final opinion that Thomas Kingston, his death, was due to the actions of his doctor, who made a profound mistake, with her prescribing, and Thomas shot himself in the head because of a severe reaction to his antidepressants. The comments section did not make one remark about SSRIs, they all centred on huge stress from money worries.
The horrifying attempted suicide of Dr. David Cartland, whose crime was to believe that Covid jabs were not safe and effective. It also appears that he has been relentlessly bullied by three anonymous doctors. David Cartland was struck off recently by the GMC.
The Value of Judgements, can go from one extreme to another.
A spokesman for GlaxoSmithKline, Alan Chandler, said: “We were surprised at the verdict. This is a tragedy caused by severe depression and not its treatment. There is no reliable scientific evidence linking the use of paroxetine to the events in this case or to violence generally.”
https://pmc.ncbi.nlm.nih.gov/articles/PMC1173337/
Meanwhile, Romain and Dexter, remain in limbo.
Joining the ‘dark side’ is well worth the effort.