Introduction
A number of factors aligned in recent weeks to encourage me to buy, and even read, Johan Hari’s latest book, Magic Pill: The Extraordinary Benefits and Disturbing Risks of the New Weight Loss Drugs. These included:
- On Hari:
- On Weight Loss Drugs:
- The recent South Park special on the new weight loss drugs
- A series of interesting interviews and conversations with a digital pharmacy which in recent months has been pivoting some of their operations and business model towards the supply of some of these drugs
So, this post will be about two things: Hari’s effectiveness and role as a popular science writer; and the potential effects and implications of the new weight loss drugs.
Johan Hari: Superior Inferior Superior Storyteller
In one adjectivally loaded term, that’s what I make of Johan Hari. Like George Orwell’s self designation as ‘lower upper middle class’, and Donald Rumsfeld’s infamous unknown unknown/known unknown/known known typology, I expect (and to an extent hope) this term is something that may initially appear as word salad, but on closer inspection and unpacking turns out to be insightful. So let’s unpack:
- superior (inferior (superior storyteller)): a superior storyteller, who’s somewhat inferior, but as an inferior superior story-teller is somewhat superior.
The term superior storyteller comes from the sociologist and historian Charles Tilly’s book Why? what happens when people give reasons… and why. Tilly offers a two-by-two typology of types of response to why questions, as shown below:
| Popular | Specialised | |
|---|---|---|
| Formulas | Conventions | Codes |
| Cause-Effect Accounts | Stories | Technical Accounts |
In brief, formulas offer if-this-then-that explanations, whereas cause-effect accounts offer that-because-this explanations. Formulas offer prescriptions for action, whereas cause-effect accounts how one thing leads to another. Science is largely concerned with the production, falsification and validation of precise and useful technical accounts. But both the level of detail required to describe causal influence in complex and complicated systems of relationship, and the need to do so with precision and without ambiguity, leads to a level of detail in terms of elements and their influence on each other that can be cognitively overwhelming, and a way of expressing such accounts - such as through algebra and graphs - that is rareified and so publicly inaccessible. Stories, by contrast, are often simpler in terms of the number of elements and types of relationship being proposed as causally related, and looser and more ambiguous in terms of the precision of language used, but have lower cognitive demands and require less training to parse and interpret.
Tilly’s term Superior Story is something like a reasonably efficient and accessible storyfication of a technical account, i.e. a reasonably accessible story based around a reasonably accurate, though often highly simplified and stylised, retelling of a technical account. In Tilly’s words:
Superior Stories? Like everyday stories, superior stories simplify their causes and effects. They maintain unity of time and place, deal with a limited number of actors and actions, as they concentrate on how those actions cause other actions. They omit or minimize errors, unanticipated consequences, indirect effects, incremental effects, simultaneous effects, feedback effects, and environmental effects. But within their limited frames they get the actors, actions, causes and effects right. By the standards of a relevant and credible technical account, they simplify radically, but everything they say is true. Superior stories make at least a portion of the truth accessible to nonspecialists.
(pp. 171-172)
So, aside from those popular science writers focused on taxonomy (raw material for the pub bore, the compulsive quizzing factshitter), the aim of a science writer should be to be a superior storyteller. And I think in Magic Pill Hari does largely manage to place himself in this camp.
But then why is Hari an inferior superior storyteller? Quite simply, because he doesn’t seem to be especially good at doing science, or employing scientific reasoning himself. This is most obviously the case when it comes to discussing risks of adverse events, such as this passage (quoted at length) on elevated thyroid cancer risk:
The third [risk] is far more serious. A few months after Daniel told me there was no safety signal attached to these drugs, one was raised for the first time. The European Medicines Agency – the regulatory body for the European Union – announced ‘a thyroid cancer safety signal’ for all GLP-1 agonists. This means that they were beginning to monitor the drugs for potentially causing thyroid cancer. They did this because of a worrying piece of research that was published in France by Jean-Luc Faillie, who is a professor of medical pharmacology at the University Hospital of Montpellier and also in charge of the National Pharmacovigilance Survey of these drugs for the French Medicine Agency. He told me that for several years it’s been known that when GLP-1 agonists are given to rats and mice ‘they have shown an increased risk of thyroid cancer’. It is also known that human beings ‘have GLP-1 receptors in their thyroid tissue’, so it’s conceivable that messing with GLP-1 might mess with your thyroid. So Jean-Luc decided he and his team needed to dig into this. France has one of the largest medical databases in the world, so they went back and analysed the data for all the patients with type 2 diabetes who had taken these drugs for one to three years, in the period between 2006 and 2018. They then compared those patients to a sample of diabetics who had not taken these drugs. Their findings were startling. He said bluntly: ‘We show there is an increased risk of about 50 to 75 per cent more’ of you developing thyroid cancer. He told me it’s important not to misread this. This doesn’t mean that if you take the drug, you have a 50 to 75 per cent chance of developing thyroid cancer. It means that if you take the drug, your chances will be 50 to 75 per cent higher than they would have been had you not taken the drug. Nonetheless, this seemed to me to be a disturbing increase. In most of the commentary on this study, it was repeatedly argued that it was a low risk. I said to Jean-Luc that maybe I was being dumb, but those figures don’t seem low to me. ‘Yeah. It’s not low,’ he said. ‘In epidemiology in general, when you have a 50 per cent increase, it’s quite a thing.’ But then he explained why many scientists would still reasonably continue to describe this as a low risk. ‘The incidence of thyroid cancer is very low. It’s not a very frequent cancer.’ (Currently, around 1.2 per cent of people will get thyroid cancer in their lifetimes, and 84 per cent of them survive it.) ‘So if you increase [levels by] 50 per cent, there is an increased incidence, but it remains low.’ But he added: ‘Given the exposure [of these drugs] to millions of patients, there will be some cases of thyroid cancer that maybe we could avoid.’
- Hari, Johann. Magic Pill (pp. 106-107). Bloomsbury Publishing. Kindle Edition.
So, Hari appears somewhat baffled by the distinction between relative and absolute risk, the kind of distinction that an hour or two reading and reasoning through the examples and methods proposed by Gerd Gigerenzer decades ago in Reckoning with Risk and similar publications would have helped to clarify. Instead, he appears throughout the book to adopt something like an Oracular Perspective with regards to scientific experts. i.e. he approaches scientists as something like Delphic Oracles, who themselves have access to some kind of arcane knowledge called science that he, a mere mortal, cannot hope to access himself. Instead, the best he can do is find these oracles, ask them questions, and try as best he can to understand the wise but gnostic utterances they deliver in response. In short: Hari doesn’t do science; he speaks to scientists.
So can a science writer who doesn’t do science be a good science writer? I guess the answer’s yes and no. (Very gnostic, I know!) Despite largely relying on scientists to ‘do the work’ in terms of making sense of data and findings, Hari’s book manages to be generally well structured and well reasoned, tackling questions about and raised by these drugs and their blockbuster popularity from multiple perspectives: the benefits and risks of the drugs themselves, their apparent popular adoption through channels of downwards cultural diffusion (starting with rich celebrities, then cascading to the middle classes, then hopefully to people who actually most need it), their broader obesogenic environment in Anglophone nations, the conflict between body positivity and acceptance and mitigating obesity-related health harms, and the curious case of Japan and its cultural resistance to Western trends in obesity. Hari manages to cover the right bases, in roughly the right order, and (I expect) draw broadly the right conclusions, while making broadly the right hedges and caveats. All the while still not really doing science.
And now the final qualifier: superior. Why is Hari a superior inferior superior storyteller?
Because as a storyteller, he’s bloody good! His book reads very well; it’s a joy to read. He brings personal anecdote (perhaps too much), drama, mystery, complex characterisation, and intrigue into a book ostensibly about stabbing oneself in the belly once a week with some drugs. By presenting scientists as oracles, he presents himself as a journeyman going on a sacred quest full of intrigue and surprise. By centring himself as the journeyman (perhaps too much), he flips back and forwards in time to invest the reader in his origin story as a ‘bad eater’, and the fallen soldiers to obesity he’s loved and lost along the way. Hari knows how to tell a story, the Campbell-style mythic architecture and elements that need to be in place to tell stories that are engaging and compelling, and he manages almost flawlessly to use this ability throughout Magic Pill.
And this ability and commitment to storytelling in non-fiction was, of course, part of Hari’s ‘downfall’ as a British journalist: sometimes Hari had a tendency to prefer the ‘truthy’ to the true. Infamously he was found, on multiple occasions, to have ‘quoted’ interviewees saying things that they did not, technically, say. Instead, having read and researched the authors, he sometimes confected ‘truthy’ quotes from the interviewees that expressed, to Hari’s mind, views he thought the interviewees held, but in more articulate and compelling ways than they happened to express in person on the specific occasion he interviewed them. Hari appears to have engaged in the linguistic equivalent of ‘airbrushing’, attempting to make the words used by interviewees prettier, pithier and punchier than those that were, technically, said and heard. On other occasions he appears to have confected quotations for dramatic effect, to punch up the stakes involved in the narrative. The fact such words were not, ‘technically’, uttered as stated appears to have been secondary to the compulsion to spin a good yarn.
And in the UK, Hari’s wrongdoing was ‘found out’, leading to some frustration and incredulity that, in the USA, Hari’s reputation appears almost entirely untainted. Like Jeremy Clarkson, Hari appears to have ‘failed upwards’, being heavily rewarded rather than punished for his misdeeds. Hence, perhaps, in part (alongside Hari’s unwillingness or inability to ‘do science’), the degree of vitriol and opprobrium directed towards Hari in the Studies Show podcast?
And the ‘Magic Pills’ themselves?
Semaglutides appear ascendant in the Hype Cycle, and may still not have peaked. So, the medium and long-term effects of this class of drugs being used for weight loss and health improvement is likely to fall significantly short of the most hyperbolic and enthusiastic pronouncements being made about them. However, even a moderate fraction of the hyped peak of expectations will still represent a seismic change to obesity, to obesity-related health harm, and to public health, which could well represent the start of a turning point in adverse trends in the obesogenic environment. These drugs may well turn out to be to obesity as vaping is tobacco: a type of technology that could genuinely break or substantially attenuate the health harms caused by one of the primary ‘avoidable’ causes of morbidity and longevity loss.
And like vaping, there are plenty of reasons to find elements of this ‘solution’ unpalatable. Both are commercially generated solutions to commercially generated problems, and as a result their widespread adoption appears to risk ‘rewarding’ the private sector twice: both with the sale of addictive ‘poisons’ (cigarettes; ultra-processed foods) and with the sale of life sustaining ‘antidotes’ to these very same poisons (vapes; semaglutides). Although in the case of semaglutides the same exact private sector companies aren’t both providers of poisons and antidotes (with the possible exception of manufacturers of some antidepressants, whose side effects can include substantial weight gain and reduction in satiety), the idea that the private sector stands to win twice over - from harming and unharming public health - is clearly from first principles distasteful.
Then there are wicked issues of efficiency and equity to consider. In the US, manufacturers appear to have successfully price-gouged their way to excess profitability, with prices many times higher than in other high income nations. So, while the health need tends to be higher than in Europe, given higher rates of obesity, access appears, if anything, inversely correlated with need and benefit, not least because in high income nations obesity tends to be a disease of poverty, but this same poverty (together with the absence of an integrated and effective healthcare system) means those with greatest potential to benefit tend to have least means of access.
And in the UK, the situation also appears messy and complex. Like in the USA, the initial adopters may not have been those with greatest potential to benefit, with demand driven more by those seeking slimness than those needing to avoid the worst consequences of fatness. And though the margins appear lower than the US, the market appears largely driven by private spending and consumption rather than benefit maximisation from an epidemiological or public health perspective. And here’s the dilemma for both regulation and public provision. If there’s too much, or the wrong kind, of regulation then a grey or black market in these drugs risks emerging very quickly, leading to even greater variability in quality and consistency in what people are injecting and by which people. If there’s too little regulation, and too little active engagement from bodies like the NHS, then the risks of both abuse and suboptimal use are much increased: abuse in the sense of those with unrealistic body image (such as those at risk of anorexia) using the drugs to cosh their appetite to excess and starve themselves; and suboptimal use in the sense of the main consumers becoming those who do not have the greatest obesity- and (pre)diabetes-related health risks.
If, once the soufle of hype collapses, and it’s clear from a realistic evaluation of their effectiveness there’s still enough underlying substance, then this class of drugs has the potential to be to obesity, pre-diabetes and diabetes, as statins are to cardiovascular risk: something taken and prescribed widely, prophalactically, and continually, but to (albeit broad) subpopulations known to have elevated risk. Statins have, within a generation, managed to shift the curves on cardiovascular risk, and are almost certainly part of the reason cardiovascular mortality has given way to cancer mortality as the leading cause of death. But this class of drugs is also different to statins in a number of ways: they are both much more expensive, and have much greater misuse potential. No one takes statins for reasons of vanity, because it leads to changes in their appearance in ways that are generally considered socially desirable; whereas for many users of semaglutides this may be the primary reason to take these drugs.
Some Fermi Estimation: Would semaglutides-as-statins be affordable?
Let’s do a little bit of Fermi estimation to get a sense of the scale of the cost of broad spectrum NHS prescription of semaglutides in a similar way to statin adoption: In the UK, around one in four adults is obese. If we take the prices here as broadly indicative of cost, then at £200 a month the prophalctic use of such drugs for an obese person will be around £2400 a year. Let’s say instead the mass purchase of such drugs can lead to a halving of the costs of procurement, and to simplify say each individual’s treatment ‘only’ costs £1000 a year. If the adult population is around 56 million, and one quarter are obese, then there are around 14 million obese people. If the NHS were to use semaglutides similarly to statins, then the upper limit of the annual cost will be around £14 billion. According to this source the NHS England budget is around £155 billion. Let’s therefore assume the UK NHS budget is around £200 billion.
Very approximately, the adoption of a semaglutides-as-statins policy by the NHS could therefore increase the annual running costs to the NHS, at least in the short term, by around 5-10%. This is actually somewhat less expensive than I was expecting, with price of course being the critical factor. If the prices charged in the US were charged here, then this could balloon an order of magnitude: 50-100% of NHS budget! I guess mass public purchase of these drugs could both reduce the unit cost, through mass discounting, but could also (in the shorter term) increase the cost, by increasing demand against finite supply. And of course not everyone with obesity may take up an offer of semaglutides even if free at the point of use for them, which would reduce both the longer term population effectiveness but also the shorter term cost to the public purse.
So, something like an NHS adoption of these drugs as obesity and diabetes analogues to statins would be expensive, but wouldn’t necessarily be unaffordable. However it would likely be expensive enough that it would have to be a political decision more so than a decision made by the NHS alone. Its adoption could be a quintessential example of spend-to-save, with reductions in incidence and treatment of downstream consequences of obesity and diabetes potentially leading to it becoming cost saving to the NHS within a few years. Unfortunately, both the current Conservative government, and the overwhelmingly likely next government, appear similarly committed to further austerity, and to a somewhat naive and simplistic way of thinking about public spending that fails to distinguish between spending on investment - which are likely to either save money in the longer term, or reduce the ratio of public debt to GDP by encouraging growth - and other sources of spending, even though even Andy Haldane, the former Chief Economist at the Bank of England, believes the government needs to make this distinction, and spend much more on investment.
A political commitment to increase the NHS budget by a tenth in the short term, with a reasonable expectation that this increased short term cost could become cost neutral for the NHS within a decade, and have substantial and positive growth spillover effects in the broader economy through improved working age health and productivity, could be a very wise commitment to make. However, this commitment seems far from what currently on the cards, and instead a continuation of further de facto privatisation of UK healthcare, due to ever poorer service quality leading to those with the means to go private to avoid waitings lists that are now in the millions, sadly appears more likely.
All the above speculation depends on other unknowns and unknowables too, however. Perhaps after the hype has collapsed it’ll turn out these drugs are much less effective, and much more harmful, than currently believed. More wickedly, perhaps the drugs do turn out to be effective, greatly increasing satiety and so reducing demand for ultra processed food, but as a consequence the economy shrinks rather than grows, because more of it (e.g. Domino’s and Deliveroo) depends on feeding people the junk food that makes them sick, than benefits from the population being well?!