You might be forgiven for thinking the new discipline of maskonomics was created by government to facilitate to masking of awkward and tiresome economic results. In fact, it is a close cousin of economics, the field of study that makes astrology look rigorous, and its purpose is to mask, or silence, mask dissenters by providing a barrage of argument in favour of wearing face masks, or coverings, in public to reduce covid–19 spread. Its high priestess in England is none other than the utterly formidable Prof Trisha Greenhalgh, a medical OBE whose gaze alone is sufficient to curdle fresh milk at twenty paces. Lesser prelates include Dr Venki Ramakrishnan, a more benign but no less mask committed Nobel Prize winning structural biologist who is for the time being Chief Pongo at the Royal Society, where he has found applying structural social engineering to promote mask wearing very much to his liking.
The central premise of maskonomics, partly explored in posts passim, is that in the face of an unprecedented etc etc global public health threat, it is very much a case of all hands on deck, of any port in a storm. In maskonomics, that means no need to bother with all that randomised controlled trial nonsense. Instead, any evidence, however oblique or remote, can be called into play, if it helps the cause. And cause indeed it is. In her May 2020 bonfire of the straw men paper, Greenhalgh rallies us all with a stirring call to arms: “…[W]e also need to remember our moral accountability to a society in crisis. The relentless, day on day stories of avoidable deaths from this dreadful disease sicken me. I will do whatever I can, as an academic, a doctor and a citizen, to reduce that death toll and help get society back running again.” Readers, especially medical readers, will note how strikingly unusual such noble words are in a medical paper. Noble words that, a bit of Dr No can’t help wondering, are perhaps too noble.
There is a strong sensibility on display here, and there is always a risk that strong emotions in medical practice can cause the glasses to fog up. As Jane Austen reminded us in a now 200 year old narrative, sense and sensibility are not the same thing. Greenhalgh presents the backbone of her arguments — not letting the great (RCTs) be the enemy of the good (other evidence, including narrative evidence, which Greenhalgh finds very much to her liking), parachute evidence (no need for RCTs when an intervention very clearly saves lives) and the precautionary principle (why not wear a mask if they might do some good?) — as if they are pragmatic good sense, the sense of Austen’s Sense and Sensibility. But if we look again the middle part of Greenhalgh’s noble words, and add some emphasis — “The relentless, day on day stories of avoidable deaths from this dreadful disease sicken me” — what we find is a paper driven not by sense, but by sensibility.
There was a time when sensibility meant an aesthetic virtue, a refinement of the senses, an ability to appreciate the finer things in life, but here Dr No uses sensibility as Austen meant it, as an excess of sensitivity, what we might today call hyper- or over-sensitivity, which gets in the way of rational thought and progress. Greenhalgh’s “narrative rebuttal” (“narrative rebuttal” is her whole description of method in her straw men paper) has its origin in narrative (“relentless…stories”), which generates in Greenhalgh an acute sensibility (“dreadful disease”, “stories [that] sicken me”) that then drives the rebuttal. The danger here is that of falling into a variant of ‘the map precedes the territory’. Greenhalgh’s narratives of sensibility become her map, which then precedes, and so defines, her territory.
In passing (this and the next paragraph can be skipped without losing the core of this post), Dr No has somewhat stretched the meaning of ‘the map precedes the territory’ here, as a polite variant of here’s the answer, now where’s the evidence. The original ‘map precedes the territory’ stems from Baudrillard, an all but completely opaque French postmodernist philosopher. His account of the Precession of Simulacra (five star opaqueness from the get-go), has it that a simulacrum (representation, image, map, model) precesses (precess is not a variant of precede (comes before), instead it means to rotate, or move around something, in this case reality) through four stages: first it is the reflection of a basic reality, then it masks and perverts a basic reality, next it masks the absence of a basic reality, before finally it bears no relation to any reality whatever: it is its own pure simulacrum, which now exists in and of itself.
To put some flesh on the bones of this turbid Gallic waffle, what Baudrillard is trying to describe here is the process by which the video of a wedding on Instagram becomes more important than, and then ultimately replaces, the wedding itself, or the selfie of me with Mandela posted on facebook becomes more important than, and then ultimately replaces, my meeting with Mandela. We can also see in Baudrillard’s four stages glints of other mask narratives at work here: the mask that perverts reality, the mask that hides the absence of reality, and the mask that masquerades as reality. Such a precession might have more than a little to do with the way in which mathematical modelling — as used by Ferguson and others: perhaps the model precedes the pandemic — but that intriguing possibility will have to wait for another day, if this already long post is not to be overlong.
Dr No has been plugging the N-word (narrative) because it describes a particularly interesting form of ‘other evidence’ recruited by Greenhalgh. In sound hands — Dr Jonathon Tomlinson, an NHS GP working in Hackney, London is an excellent example, and his recent post on consulting during covid is well worth a read — narrative research can be profoundly useful, since it can examine parts of medicine not amenable to more conventional research. But that is as far as it goes. To co-opt narrative research as if it were quantitative research — which, if you look carefully at Greenhalgh’s “narrative rebuttal”, is what her paper does — is to conflate the qualitative and quantitative, and so to abuse both.
A narrative is at heart a personal story, an n=1 account, which holds its value in depth and words, which cannot be added, without making a nonsense, to studies based on observations, which hold their value in breadth and numbers. The anecdote of the mask wearing man on the plane (no one gets infected), or the non-mask wearing choir (lots get infected), are just that, anecdotes. To accord them the status of narrative research is to mistake reportage for narrative. Once stripped of its flummery, we can see that Greenhalgh’s straw men paper contains but one narrative, the narrative of an academic GP driven to distraction by relentless stories of a dreadful disease. Noble perhaps, but not evidence of the effectiveness of face masks (or coverings as Greenhalgh prefers to call them, presumably to distinguish them from medical masks) in public to reduce covid–19 transmission.
In a moment Dr No will provide two examples, one a medical treatment, the other a public health measure, where advocating a common sense approach without adequate evidence backfired badly, by which Dr No means unavoidable deaths in their tens of thousands. Before that, he feels compelled to mention another anomaly in Greenhalgh’s straw men paper that bothers him, the use of quotes and quotation marks. Conventionally, quotation marks are solely for direct or verbatim quotes: this is what was actually said or written. Indirect, or reported, speech, or writing, travels without quotes. The distinction is crucial, because in the second indirect form, we the readers are made aware the current author may have tampered with meaning, whereas in the first, bar such wheezes as quoting out of context, we are told this is what was actually said or written. For some reason — perhaps there is a narrative for this as well — Greenhalgh puts several reported/indirect quotes from her straw men in quotes, but they are not direct quotes. For example, we have (double double quotes as Dr No is quoting something which is itself in quotes):
“”Masks, suggest Martin et al, are an example of a complex intervention in a complex system. Their effects are impossible to predict, therefore we should not introduce them[.]””
In passing, another two conventions as Dr No understands and uses them: square brackets in a quote mean something added, as in the full stop above, or changed eg lower case to upper case, an ellipsis means something omitted, usually in the interests of conciseness, in both cases by the current author.
What Martin et al (the two straw men and one straw woman targeted in Greenhalgh’s narrative rebuttal) actually wrote turns out to be considerably more nuanced (there are two versions because largely the same material appeared in two different places):
“Face masks (and measures to secure their uptake) are a complex intervention in a complex system: the results of a change of this nature are emergent, unpredictable, and potentially counterintuitive.”
” Face masks are a complex intervention in a complex system: their impacts are emergent, unpredictable, and potentially counterintuitive.”
Neither in two versions, nor in the vicinity, is Greenhalgh’s added “therefore we should not introduce them”. Instead Martin et al counsel an urge to caution, to not introduce an unevaluated measure. Greenhalgh — also known on occasion at the BMJ as Greenlaugh — may intend to lay straw men to rest, but appears instead to be tilting at windmills. More than that, she is misleading her readers.
Dr No favours Martin et al’s implementation of the precautionary principle: not to introduce an unevaluated measure until there is sufficient good evidence of both useful benefit and absence of harm, over Greenlaugh’s something must be done, even if the evidence of useful benefit and lack of harm is roundabout. He favours the former because there are all too real examples from medicine where implemented precautions based on common sense, rather than solid evidence, have turned out to cause considerable harm, measured in tens of thousands of deaths. Two well known ones are sufficient to make the point.
The first is the use of precautionary use of corticosteroids to reduce inflammatory changes in acute severe head injury. The condition was common (worldwide, millions of cases every year), the threat real (death or permanent brain damage, often in young people) and the common sense logic clear enough (corticosteroids reduce inflammation). There was even a small number of non-conclusive but suggestive RCTs that leant lightly in favour of corticosteroid use. Unsurprisingly, corticosteroid use became common practice.
Then came the CRASH trial, a properly conducted adequately sized RCT that revealed the sickening reality: corticosteroids significantly increased mortality. All cause mortality at two weeks in the treated group was 21.1% (paywall), compared to 17.9% in the untreated group, rising to 25·7% (paywall) and 22·3% respectively at six months. A Lancet commentary (paywall) concluded that corticosteroids use in severe head injury had caused more than 10,000 avoidable deaths during the 1980s and earlier.
The other perhaps even better known example comes from a public health recommendation, that babies should sleep face down. In the mid to late 1950s, among many others, the hugely influential Dr Spock, promoted prone sleeping for babies, arguing —without evidence — that ‘If he vomits, he’s more likely to choke on the vomitus. Also, he tends to keep his head turned to the same side—usually toward the centre of the room. This may flatten the side of his head.’ No mother wants a lopsided baby come the morning, and unsurprisingly putting baby to sleep face down became common practice.
Then came the evidence. By 1970, it was clear there was a significantly higher risk of cot deaths (SIDS) from front sleeping. Yet, even with evidence of the harm available, the face down recommendation continued, only fizzling out in the late 1980s. Estimates put the number of avoidable deaths at over 10,000 in the UK alone.
These two deeply painful and admittedly extreme lessons from history are a poignant reminder that neither sense nor sensibility can stand in for science. Both sense and sensibility, like the human that stands behind them, are fallible. Even good science isn’t fallible either, but it does go some way towards mitigating the real dangers of arguing from inadequate evidence. As the CRASH commentary argues, applying interventions with unproven effectiveness is like flying blindly — and no one should be forced to do that, either by influential and well meaning but distracted and misguided academics, or by government.