Not even a global pandemic can stop the feverish hysteria of diet culture! As COVID-19 wreaks havoc across the world, there’s a sh*tload of truly hideous media articles and speculative research editorials proclaiming that higher weight people get sick more often, more severely, and even die at a higher rate than thin people. This narrative is largely being accepted as an unquestioned ‘truth’ by most media outlets. But where did this idea come from, and what does the data say? This week on All Fired Up we’re travelling around the world on a virtual COVID Contiki tour, visiting the COVID-19 hot spots, finding out where these narratives came from, and digging deep into the statistics to see what’s actually going on. I am joined by my fellow tour guides Fiona Willer, anti-diet dietitian and weight science expert, and Jess Campbell, anti-diet nutritionist and medical student, in an intrepid mission to uncover the truth!
This is an epic 2-part series which is ESSENTIAL LISTENING. In such challenging times, we need objective and transparent information. It’s simply not OK to keep serving up weight biased BS. But be warned - this is obviously a very distressing topic, and this episode contains multiple uses of the “O” word (there was no way around it), plus we’re going into explicit discussions of disease, death, and BMI. Look after yourself!
Show Notes
- This week I have multiple guests in an incredible 2 part series! We are super fired up about COVID-19 and the BS weight-related outcomes being speculated about left, right and centre. Fiona Willer, fearless anti-diet dietitian and weight science expert, alongside Jess Campbell, anti-diet nutritionist and medical student join me to unpack the COVID-covert-crap!
- We’re in the midst of an epidemic, with a terrible virus taking over the world. On the day of recording (26th April 2020) there’s something like 200,000 deaths across the world and so many countries in lockdown and in crisis. And as usual, we’re pissed off - we’re hungry for data, to understand and unpack the fast-moving science behind COVID-19 and what’s making it tick.
- This whole podcast is about the impact of how weight stigma and weight bias impacts how we understand the world, and it’s clear now how weight bias is shaping how we understand COVID-19. So, I’m so happy we have two of the finest brains on the planet to join me on this massive research rabbit hole that we’ve been in for the last few weeks.
- Today we’re going to unpack the science and the published data that we’ve come across. Not all of it - there’s mountains of it - but we’ll try and do a bit of a whistle-stop tour around the planet of what kind of data is coming out of different countries, and compare and contrast that with what is being said in some of our media and our journals. What fun! A “COVID Contiki tour!”
- So, what are we seeing? There are some really hideous media articles around that are basically claiming this idea that higher weight people are at increased risk of catching Coronavirus, that they have more severe symptoms and a higher death rate. So our starting questions were about whether this is true, and what is it based on?
- I wanted to start with some examples of the really shitty media articles that are coming out that have really damaged and upset people in larger bodies. Two of them came out on April 16th, 2020 - there was an article in European Scientist from Dr Aseem Malhotra called “COVID 19 and the elephant in the room”. Fiona reckons that the use of this idiom is a bit of a litmus test for uninterrogated biases.
- A warning that this episode will likely mention the ‘o’ word several times, as we discuss these articles.
- The first page of Dr Malhotra’s hysteria-raising article says in capital letters “OBESITY - THE REAL KILLER BEHIND COVID”. Dreadful.
- It basically goes on to claim that being in a larger body is a risk factor for catching COVID-19, and then puts the blame straight back on people by claiming that it’s because of poor diet and body size that people are getting sick, and if only everyone went low carb, high fat immediately, everyone would be okay. He’s also insulting Boris Johnson by saying he got sick because of his weight. A complete ‘sharticle’.
- The very naive belief that somebody’s body size has any reflection of their current eating habits or physical activity habits needs to die. That’s the problem here.
- He also comes from a place of stating that the UK is the unhealthiest it’s ever been, based on body weight statistics.
- The whole sharticle is full of frightening stats, really running on fear, with little to back it up. He does this lovely sentence - “a recent commentary in Nature states that patients with type 2 diabetes may have 10 times greater risk of death when they contract COVID 19”. These are speculative, these articles - they’re commentaries, not based on actual data. But it still raises fear and gives a very suspicious so-called solution as well, because he’s basically recommending everyone changes their diet and they’ll be okay. And lo and behold, when you do a bit of digging into Dr Malhotra, he has a low-cal diet book! Completely unbiased, nothing to sell here! He got such a big platform for writing this article in European Scientist, giving him so much sciency-sounding cred. Articles like this do so much damage to people.
- And European Scientist is not a journal - it’s a magazine. Sneaky.
- Think of him as like the UK version of Dr Oz - he’s got the cardiology background but he’s deep into the ‘woo’ science.
- Then we have a New York Times article that came out on the same day, by Roni Caryn Rabin titled “Obesity Linked to Severe Coronavirus Disease, Especially for Younger Patients”. And it’s accompanied by the tried-and-true headless fatty shot. This article, it should come with a health warning for weight bias and weight stigma. Reading this article is actually what got me started on this podcast episode, because the article does reference some studies and data that were starting to come out which started the rabbit hole for us. This article implies a level of certainty that is not backed up when we actually look at the studies it is talking about.
- So, in addition to the media stuff, we also see this narrative being built in the academic journals that is really troubling. On April 1st 2020, there was a letter to the editor in the Obesity Journal from William Dietz, called “Obesity and its implications for COVID 19”. This letter claims a strong relationship between COVID and weight.
- On the same day in the same journal, The Editors Speak Out was published. It tries to make the case that this is a weight related problem and says speculative things like “we are likely to see a collision of the two public health epidemics in the US, with obesity and COVID 19 interacting to further strain our health system”.
- What’s interesting with that is that this is April 1st 2020, that this journal is posting with a very strong level of certainty that because COVID 19 is a weight related problem, that we need to focus our attention on people’s weight when we think about this illness.
- They’re pulling on information from previous pandemics too, from H1N1, SARS and MERS, and more broadly from others. They’re reporting lots of animal study outcomes for those, and then trying to project them onto humans - but in real life that needs a lot more clarification. Lots more study is needed to find out whether what we find in an animal model is also found in humans.
- The reliance on acute respiratory distress data, when we started to realise that COVID was presenting with an atypical ARDS - it’s apples and oranges, the data that they are relying on for those commentary pieces. So, it’s pulling stuff that’s not related and saying ‘it’s definitely going to happen in the same way’.
- It’s also ignoring the body of evidence of the protective mechanisms of a high BMI for ARDS as well - a lot of those articles say that increasing BMI is associated with worse outcomes with ARDS, and on balance that’s not what the evidence is either. It’s not a fair representation of the research.
- Whenever provided information about BMI, the tendency for journalists or research teams is to present it in a negative light no matter how borderline it is in real life. The bias goes unchecked.
- There’s also a perception that talking about body weight is ‘newsworthy’. It’s a topic that is exhaustingly everywhere all the time.
- And there’s this perception in the media that a negative result - where you test something and find no relation between the things you were looking at - people don’t want to publish that. They feel it’s not a ‘real’ result, and those studies don’t make it into the literature as much as they should. A ‘no relationship’ result should be making it into journals more often.
- These two opinion letters … let’s remember first of all that the name of the journal is Obesity, with a lens of ‘obesity is bad’. William Dietz is very well known for his weight bias agenda, and is very well paid off by Novo Nordisk and WW (Weight Watchers).
- So, shall we get started? We have a lot of ground to cover! We thought we’d do it by going around the world to figure out who is doing what research, and what is actually being found.
- First stop - China. We’re looking at “Obesity and COVID-19 Severity in a Designated Hospital in Shenzhen, China” by Qingxian, Cai, et al. posted on April 1st 2020. This study looks at 383 people. From the title, you can see we’re going after the issue of obesity.
- This paper is amongst the earliest written about weight and outcomes for COVID, and in the time line it’s early in the world’s experience of coronavirus but far along in China’s experience of it. Something that’s different to other papers like it is that they talk about the types of treatment people have received, and the progression from hospitalisation through to intensive interventions like ventilation. It’s more information than we get from most of the other papers.
- This paper was the canary in the coal mine for the claiming a relationship between BMI and more severe outcomes. This was the first point at which the media stopped pontificating about what they ‘felt’ about BMI and had something concrete to use. But they found that there was not a relationship between BMI categorisation and severity of illness except in men. When we look at the weight information, we have 383 patients and only 41 have a BMI that China categorises as ‘obese’ (side note - China’s BMI cut off for ‘obese’ is 28).
- Of those 41 patients, 8 of them had liver disease. Liver disease is a condition where you can develop another condition called ascites, where there’s a lot of fluid sitting around the organs and guts and you’re not necessarily very good at getting off extra fluid. So, it may well be that a number of that 41 had a condition where they were fluid overloaded all the time. They may not then necessarily be in the ‘obese’ bracket if they did not have this liver condition at the same time. It’s not just ‘well’ humans who are in this bracket - and there are more people with liver disease in this bracket than the other BMI brackets. The liver disease is a much more serious disease to have co-occurring with COVID 19. Statistically significant in terms of progressing to severe disease is only seen in this small group of men with higher BMI.
- Even for a first-year undergrad, you get the message that statistical significance is great from a mass perspective, but it may not actually be clinically significant. If we know that there’s a statistically significant difference, what does that mean? What do we do with that information? If it’s not body weight, if it’s not a condition that’s highly stigmatised, then we say “oh, we should be doing better screening…” but with weight it’s like, “oh well”.
- Fiona is very keen to get a copy of the data in this study so she can run the numbers herself!
- The article concludes with the statement that “compared to individuals with so-called normal weight, obese persons were more likely to progress to severe pneumonia due to COVID 19”. And that statement hardly fits the data - they themselves only found that ‘obese’ men progressed to a more severe version! They’re spinning their own data. And THAT’S the only thing that we will hear out of that study.
- At one point, that paper does mention that people with higher body weight were treated later than the other patients, which might have impacted outcomes. But you couldn’t see clearly from the data what that actually did.
- We can’t assume that there isn’t a lot of jumbling in who got what treatment underneath the surface of these numbers, compared to a study of a more mature condition.
- Interesting that this study didn’t control for smoking, when something like half the population of men in China smoke?
- These things don’t make it into the sexy headline - this early Chinese data is the beginning point for the COVID and weight relationship. When we dig into it, it’s not so straightforward.
- No paper is complete - all we can do is look at the data from all the places we can get it, and look at the summary papers that have been published so far. We look for a pattern that keeps being repeated across different locations to see if it’s a real thing or not. Findings from these smaller papers could be a quirk of stats, or influenced by the author’s particular perspective. If a phenomenon is a real thing, it’s going to be repeated across many locations over and over again - it’s going to be really obvious that the effect is there. That’s what we’re searching for. More than just hysterical headlines - we need the ‘what then can we do to improve outcomes for this subgroup?’.
- We’d also expect dose-response with BMI, not just categories presenting with associations that aren’t held across other categories.
- The way that we treat BMI is nonsense in that we have a continuous measure (going from zero to infinity in a straight line). It doesn’t naturally lend itself to categorisation. Those cutoffs in BMI are to a degree arbitrary, because a human has decided them. They’re not based on something magical that happens overnight when you go from a BMI of 29 to a BMI of 30. Without a set of scales, you couldn’t measure that difference. So, the cutoffs are dodgy from that perspective. And different studies group BMI together in different categories, which means we end up with very different findings across studies. When you lump a group of people in a category, such as BMI of 30-40, we lose all the significant data of the people along the continuum within that group. They’re all treated as if they’re the same, whereas in real life they’re probably quite different in terms of their experience and health conditions. It’s not a fair treatment of the information - especially considering the study has the detail of each study participant’s BMI. Stats should be treating BMI as a continuous measure, otherwise it’s not a fair assessment.
- Another thing - is BMI being measured, or eyeballed, or estimated? There’s huge amounts of missing data in many studies.
- The Italian data (coming up!) only had BMI data on 8.8% of the people in their studies - and yet are extrapolating based on BMI.
- An early paper looking at cardiovascular outcomes from China got the world spinning on the BMI and COVID connection - a small cohort study of 112 patients. The study reported that of the 17 patients who died of COVID, 15 had a BMI of over 25. 15 of those 17 patients who died also had one or more of the following: hypertension, coronary heart disease or heart failure. So, whilst the paper observed an association, it did not establish causality nor did it tease out the relative contribution of each of those conditions to the outcome. This study was really used to push a direct link between BMI and COVID, despite it not standing BMI as an independent risk factor. We also have to remember that the relationship between cardiovascular disease and BMI is complex in its own right, often leaving out of the conversation the impact of weight stigma, weight cycling and medical marginalisation on folks that have heart disease and are also at a higher weight.
- Next stop - the USA. Published on April 8th 2020, it’s titled “Factors associated with hospitalization and critical illness among 4,103 patients with COVID-19 disease in New York City”. What a sexy title!
- As you might suspect, this paper looked at COVID patients in New York City. 7,700 people got tested, with 4,103 testing positive for COVID. This paper investigates what happened to them after testing - who got hospitalised, and who of the hospitalised became ‘critically ill’. We do find out that from this sample, 292 people have died and 417 are still hospitalised with nothing in particular happening to them.
- This is a situation, again, where things are being reported with a whole lot of data underneath that Fiona Willer would love to get her hands on to run the numbers. It’s interesting to find out from their sample who got hospitalised, and who became really sick - because the decision to hospitalise someone is a decision made by another human being. There’s bias built into the system. What we see from this paper is that people who are more likely to be hospitalised are people over 65, and with a BMI of over 30. Then you think okay, that’s because the people doing the admitting are clearly most worried about those two kinds of cohorts. There’s a lot of overlap - in the population we have a situation where BMI’s do increase with increasing age, so we can’t really unpick age and BMI from the other. Age is always going to be there behind other factors. Another part of that phenomenon is that it’s a particular cohort that’s aging - they had particular experiences and exposures in their childhoods, teens, adulthoods, etcetera. Today’s 65 year old is very different from a 65 year old in the 1950’s.
- We start seeing some more objective measures in this paper - who is progressing on to particular measures once hospitalised. What we want to see is that there’s an increase in the likelihood that people who are hospitalised progressing to ‘critical illness’ in those two categories. And we want to know what is associated with ‘critical illness’ so we can direct resources there.
- When things don’t go through peer review, we end up with tables that are all higglety-pigglety and driving Fiona Willer to distraction.
- We can see that yes, men are more likely to progress, and older people are more likely to progress to severe illness. But with the BMI categorisation, the way that it’s presented is not clear. The effect is much smaller for the relationship between BMI and progression to severe illness than it is for hospitalisation.
- There’s the potential for circular reasoning in this paper - weight/BMI being double-counted. And this paper is all over the place - if Fiona and Jess have to spend days getting their heads around it, the media isn’t going to have much luck.
- This paper is painful! Where did those 48 people go? What happened to them?
- A paper should present information in a way that means other research groups can understand it, and attempt to replicate it. That another group of researchers could go and run the same trial - because it’s not a ‘thing’ unless it’s replicated in multiple places. The whole point of a research paper is to illicit replication in another location, and there’s not enough information here in this paper. How did they do their multivariable regression? What did they add in?
- You can hear the pain in Fiona Willer’s voice here, there’s some sort of traumatic injury forming here with this paper!
- For those of us without a huge amount of statistical knowledge - this paper shows that you have a higher risk of being admitted to hospital if you have a higher body weight, but once you’re in hospital there’s no link between BMI and severity or risk of death. That’s clear in this data. But going back to that New York Times article we discussed earlier, the lead author of this paper is quoted as saying “obesity also appears to be a factor for higher risk of death from COVID 19”. And she’s saying that to the New York Times without any evidence in her OWN paper. This research article did not even discuss deaths. It did have death statistics noted in the flow chart, but no data was presented on who was more likely to die - let alone specific data to show a relationship between higher body weight and risk of death.
- This is a key example of how scientific data gets translated into a media message. And this is a pretty scary message - that if you’re higher weight you are more likely to die from COVID 19. And it’s not a message backed up by the science here.
- We can give these researchers a lot of leeway - there’s no peer review process here, they’re under clinical pressure, there’s a horrifying death rate in New York. There’s a push to get data out so they can work out what ‘best practice’ is for this virus. There’s no thought here that this is malicious - but it shows how important it is to be careful with what we print and to be careful with our data, and to be careful with what we say to the media. This is affecting human lives. At the very least, include the complete data set that you’re analysing so others can use it and go through that review process.
- Highlighting an article in Wired from Christy Harrison that talks about how troubling these articles are. She points out that this research was not controlling for any factors that we know have a massive impact on people’s health, such as cardiovascular health, diabetes, hypertension - factors like socioeconomic status, weight bias, race. She points out that BMI is quite a lazy tool and we can’t see all that inequality that happens with health risks and how we’re treated in health systems.
- BMI is used as a scapegoat for the disparities in African-American communities too - instead of digging deep into marginalisation, colonisation, socioeconomic status to see how they’re affecting outcomes.
- It’s easier to blame it on individual food choices - and easy to say that BMI is a simple, numerical measure compared to the messiness of factors such as marginalisation.
- There’s a Centre for Disease Control (CDC) report that is being used to back up this relationship between weight and COVID risk/severity. It looks at clinical data for people being admitted during March 2020, the first month that the US started surveying data. There’s 1,482 patients in hospital - 74.5% were over 50, 45% were male, rates were highest amongst people over the age of 65. They had data on people in this cohort for people with underlying conditions - the comorbidities these people might have - but they only have it for 12% of their dataset.
- Of those 12% - 89% have one or more underlying condition, most common being hypertension (49%), what they call ‘obesity’ (48% - in the US ‘obesity’ is classified as a disease). 34% had chronic lung disease, 28% had diabetes, and 27% had cardiovascular disease.
- What they’re concluding from this is that older people have higher rates of being hospitalised, and also that the majority of people being hospitalised with COVID have an underlying medical condition - but remember, we only have that data for 12%.
- Something else to take away - 48% of the 12% had what they call ‘obesity’. In the US, the prevalence nation-wide of ‘obesity’ is 42%. So, it’s only marginally higher in terms of bodyweight across the American population rather than being something statistically alarming.
- In the UK, we saw as early as the 23rd of March that we saw the first report out of the ICU units that ran in the media as ‘60%, 70% of admissions to ICU were in higher weight bodies’. But of course, this is not news - people showing up in the ICU were only reflecting the population distribution of BMI.
Resources
Hosted on Acast. See acast.com/privacy for more information.
