Health Messaging and Napkin Epidemiology in the Netherlands

Halfway through the pandemic thriller Contagion (2011), one scene exposes some of the challenges of health communication in the digital age. In a live television interview with Ellis Cheever, head of the CDC, the crooked blogger Alan Krumwiede takes the American government to task over its unwillingness to provide clear numbers or explain epidemiological models on an outbreak of the fictional disease MEV-1:  

Tell them what an R-nought of two really means, Dr. Cheever. Teach them some math. No? I’ll do it then. On day 1, there were two people with it. And then there were four, and then it was sixteen, and you think you’ve got it in front of you, but next it’s 256, and then it’s 65,000. In thirty steps, it’s a billion sick. Three months. It’s a math problem you can do on a napkin.  

Napkin math, back-of-the-envelope math, or what we might facetiously refer to here as ‘napkin epidemiology’ refers to a rough and simplified calculation. It arrives at a ballpark figure; something to work with. In this case, Krumwiede has provided clear answers to two questions that preoccupy the pandemic subject , i.e. an individual who constitutes themselves as a particular kind of subject in response to a pandemic situation (or at least the media). How many? And: how long?  

We may appreciate Cheever’s hesitation to answer such questions. There are too many variables, too much uncertainty about this new disease, too many facets to an epidemic obscured when boiling things down to mere numbers. Meanwhile, however, Krumwiede’s math is off: intentionally adding to an infodemic of false and inaccurate information. His sequence either skips steps or squares the outcome of the R₀ (R-nought): the number of new people any one person is likely to infect. In this case, R₀=2, meaning:  

2 x 2 = 4                      not                   2 x 2 = 4

2 x 4 = 8                                              4 x 4 = 16

2 x 8 = 16                                            16 x 16 = 256

2 x 16 = 32                                          256 x 256 = 65,536  

Next it is 64, then 128, and in step thirty there would be 1 billion infections and two billion ‘sick’ (Table 1) (i.e. the number of previously infected + the number of newly infected), as Krumwiede partially corrects himself for his final figure. The R₀, however, is not a ‘fixed number’. We witness how the reluctance of a government official to cause unrest by providing preliminary and incomplete (but ‘official’ and disconcerting) numbers, collides with fearmonger and distortion by a social media presence keen to promote (as it turns out) his own financial investments in an experimental drug.  

Present

Today, in the midst of a pandemic of the novel disease COVID-19, health communication in the Netherlands has been considerably more forthcoming than appears to be the case in Contagion – yet remains evasive. The present ‘infodemic’ is defined not so much by rumour and misinformation, but by an outpour of numerical and epidemiological information that is continuously updated and revised. Its outcome is the same: confusion.  

News bulletins, current affairs shows, radio broadcasts, and newspapers are saturated with numbers, models, explanations, and forecasts of the virus, its spread, government control measures, and social responses. Social media push banners to guide users to reliable sources of information. The most recent figures for the death toll, hospitalisation rate, and the number of confirmed cases are announced at a daily press conference. Weekly briefings to parliament by Jaap van Dissel of the Institute of Public Health and Environment (RIVM) are livestreamed. For days, the Dutch watched with bated breath to see if ICU capacity in hospitals could be cranked up fast enough to cope with the anticipated peak in COVID-19 patients. In short, the country has been gripped less by panic than by the spectacle of real-time surveillance.  

Many of these figures are either unclear or a work in progress. The daily recovery rate of confirmed cases, for instance, is unavailable to the RIVM as collecting such data is said to constitute an ‘administrative burden’ without ‘influencing policy’. This viewpoint is surprising, given the ongoing debate about the possibility of herd immunity and the centrality of this concept to the Dutch control strategy. On April 16, meanwhile, van Dissel stated that ‘by looking through the eyelashes’ (i.e. a preliminary analysis) at a random sampling study conducted by blood banks, about 3 percent of the Dutch population had thus far developed antibodies to the virus – far exceeding the number of confirmed cases (38,416 on April 28, 2020). About the same time, Statistics Netherlands published its second report noting an unusual uptick in the number of deaths since early March that were presumably attributable to COVID-19 as well. Speaking of which, it has become common knowledge that the death rate for the disease is heavily skewed by the inability to test all suspected cases. Currently standing at an unlikely 11.8 percent for the Netherlands, according to the Johns Hopkins database, references to the death rate have prudently disappeared from much Dutch reporting.  

What these brief examples help to show is that with so many epidemiological models and numerics being published, compared, revised, questioned, and retracted, the clarity of health messaging and the transparency of governance have not necessarily improved. The key message – stay home! – sank in, but, by Week 3 of our so-called ‘intelligent lockdown’, the overriding question had become: for how long? To answer such and other questions, we can all play at napkin epidemiology. In order to examine this problematic a bit closer, one could look at some of the discussions surrounding our formal entry into this pandemic event, and ongoing discussions on our ‘exit strategy’.  

Entry Strategy

On March 16, Prime Minister Mark Rutte gave a rare televised address to the nation on the outbreak and control of COVID-19. The World Health Organization had declared a pandemic three days earlier and social distancing protocols were developed at speed. At a time of great uncertainty, Rutte was praised for his clarity, reassuring tone, and promises of financial relief. The country’s strategy, we learned, was to exert ‘maximum control’ over the spread of the virus within the population best able to withstand it, with the aim of building up ‘herd immunity’ around vulnerable groups. What was missing from Rutte’s talk were numbers, but numbers were hardly in short supply in other media. All anyone needed to divine the consequences of this policy, was a napkin.  

The following evening, in the current affairs programme Nieuwsuur, the university virologist Marion Koopmans was being interviewed about the Dutch strategy by host Jeroen Wollaars. Inevitably, their discussion turned to the numbers involved in attaining herd immunity. What followed was an exchange that bore a curious resemblance to the scene between Krumwiede and Cheever:  

Wollaars: How are things looking in the Netherlands if 50 to 60 percent of the population became infected?  

Koopmans: There are numbers for that, there are models in England and the United States that are also used in the Netherlands, both nationally and regionally, I don’t know those numbers exactly […]

Wollaars: Yes, you say you don’t know these numbers exactly, but we called the RIVM today. They don’t want to elaborate on them in front of the camera, but on the telephone they said the following – I would like to go through them with you – they say, if you assume 50 percent infections, so 17 million people divided by 2, is 8.5 million, more than half of them will become ill, well let’s be careful and say 8 million people, than we can say 4 million people will really become sick. 

Koopmans: Yes.  

Wollaars: Is it known then what percentage of them will die?  

Koopmans: Looking at the numbers around the world, they fluctuate between 1 and 2 percent […]

Wollaars: So, if we translate that to numbers, that would mean between 40,000 and 80,000 people.

Koopmans: Exactly, so that would be a high figure. 

(28:00 – 31:00)  

Bearing in mind that Koopmans is an expert virologist (and not a government spokesperson), her assertion to not know the figures ‘exactly’ may strike us as odd, but not as odd of the subsequent breakdown of the figures to her by a journalist (which she is then forced to confirm). Meanwhile, we have learned that the RIVM is similarly disinclined to elaborate on the figures involved in the strategy announced by Rutte ‘in front of the camera’. We may ask ourselves: is Wollaars’ napkin maths journalistic sensationalism? Fearmongering? Or forcing a discussion on numeric projections that were simultaneously public but blanketed by jargon?  

Again, we witness a clash between what we might think of as an expert attempt to nuance and avoid causing unrest, and a more popular desire to have the figures out in the open. If this is indeed the best way forward, what will it cost us? The question regarding Rutte’s speech, however, is not: ‘why were these numbers not made clear?’ But rather: ‘why were these numbers not made clear given that the data with which to calculate them was circulating widely?’ From either a political or a public health point of view, one might question the wisdom of letting people ‘figure things out’ on their own. This question becomes increasingly pressing as the public debate shifted towards discussions of the Dutch ‘exit strategy’.  

Exit Strategy

Three weeks after social distancing protocols were implemented in the Netherlands, there appeared to be a growing realisation that ‘maximum control’ towards ‘herd immunity’ was a process not of weeks or months but of years. Politicians, economists, social scientists, health workers, and members of the public began to reflect on the question: how do we go forward? Unsurprisingly, the efficacy of various suggestions to ‘just do this’ tended to suffer from oversimplification, or failed to address the associated questions of ethics and solidarity. All too often, these commentaries posited ‘health’ and ‘economy’ as diametrically opposed values.  

These tensions were disturbingly evident in a short commentary on the website of business magazine Quote of April 17, in which real estate tycoon Klaas Hummel urged the Dutch to ‘revolt’ against the so-called ‘1.5-meter society’.[i] Using the latest available figures, he engaged in a bit of napkin epidemiology to argue that the consequences of imminent socio-economic collapse far outweighed the impact of COVID-19.  

According to van Dissel’s briefing the day before, 3 percent of the Dutch population was thought to have developed antibodies to the virus – just over 500,000 people. There had been 3,315 confirmed deaths, supplemented by ‘several thousands’ that emerged through the mortality data collected by Statistics Netherlands. This, said Hummel, amounted to no more than 6,500 COVID-19 related deaths. If so, Hummel was on the verge of determining a not insignificant death rate (1.3 percent) but instead he forged ahead to emphasise how 97 percent of these deaths were among people above 60 (‘but really above 70’) with underlying health conditions. With an average daily death toll set by Hummel at 140, this ‘target audience’ of 4 million people stood to lose 40,000 people (or 1 percent) to COVID-19 by the end of 2020. All this was truly terrible, but one had to realise that these were mostly people with short life expectancies, few remaining ‘efficient’ years, and poor enjoyment of that remaining time if social distancing protocols had to be adhered to. Indeed, 3.25 percent of this group were expected to die from other causes anyway. The trick was to substitute part of these deaths for COVID-19. Thus, mortality for this group would not go up by a full percent (to 4.25) but only by a little to, say, 3.5 or 3.6 percent: an uptick (Hummel left unsaid) that translated into an additional 16,000 to 20,000 deaths.  

There are moral arguments to be made here, but let us concentrate on the use of data instead. The provisional nature of some numbers (3 percent exposure of the population), the provenance of others (6,500 deaths in total; an average of 140 deaths a day), and the calculations applied by Hummel (notably his ballpark estimate of 3.5 to 3.6 percent mortality among people over 60) are not made clear. But this, to an extent, is irrelevant. Even if they are perfectly correct, the major flaw in Hummel’s reasoning is his suggestion that the numbers would remain constant once the Dutch rebelled against social distancing protocols. What he – and other commentators – either forgot or ignored is that the figures with which he worked were the product of a tremendous health intervention. If these measures were lifted outright, infection, hospitalisation, and mortality rates would undoubtedly rise at an alarming rate. Hospitals and healthcare staff would be overwhelmed, with severe consequences for the provision of ‘regular’ healthcare for people across all age brackets. Every ‘exit strategy’ will have a cost.  

Conclusion

The key point of these reflections is to stress that our knowledge of the disease COVID-19 and its causative virus SARS-CoV-2 is still very much in flux. In the Netherlands, seemingly contradictory information crowds the media every day. Sometimes this is particularly evident, such as when a newspaper notes the RIVM will ‘think about’ making face mask wearing mandatory in particular situations on the same page as an article stating that much-maligned droplets were ‘mostly not’ to blame for the transmission of this disease. There is a similar back-and-forth on the matter of herd immunity, the susceptibility of children, and the correlation of COVID-19 to particular medical and environmental conditions. We are gripped by the spectacle of pandemic real-time surveillance, but equally absorbed by an unceasing flow of scientific revelations. The author Leon de Winter captured this sentiment eloquently (albeit in an article engaged in a lay epidemiological commentary in the same vein as Hummel’s):  

Every day, the numbers of the Institute of Public Health are published. Hungrily I absorb them as I look up from my lay study of epidemics, with which I have occupied myself this last month.  

In order to cope with the incessant flow of information on COVID-19, we have all had to become amateur epidemiologists to an extent. We partake in the feast of numbers to determine how we might practise our trade in the near future, or how to meet friends and loved ones in safety. We see how napkin epidemiology might help us get some figures out into the open, but also how our imperfect understanding of them may lead to oversimplification or distortion. Scientific transparency (on which the government advisory panel was challenged at the time of writing), clarity of numbers, and precision about the nature of the ‘diabolical dilemmas’ facing us are key as we move forward. This is increasingly apparent as government officials and health workers are increasingly called upon to justify their decisions and clarify: why don’t we ‘just do this’.[ii]

---

Maurits Bastiaan Meerwijk is an interdisciplinary historian of medicine and colonialism in Asia. His research considers the impact, representation, and control of vector-borne and zoonotic diseases against the backdrop of scientific development and environmental change – in particular dengue fever and plague. He is an affiliate research scholar with the Centre for the Humanities and Medicine at the University of Hong Kong.

---

Notes

[i] Cabinet now speaks of the ‘1.5-meter society’ and the ‘1.5-meter economy’, in relation to the distance people are being recommended to stay apart. This distance, however, is set at 1 meter in France, 2 meters in the United Kingdom, and 6 feet (1.80 meter) in the United States. Such causes further confusion. Is ‘distance’ itself the goal and we take whatever distance sounds best in our respective languages, or is there a specific, scientific standard to uphold?

[ii] The questions asked of Rutte and van Dissel at a press conference announcing the extension of most control measures on April 21, 2020 are illuminating in this regard. E.g. ‘Mr. Rutte, for weeks you have said the Dutch are behaving admirably, and what do people get in return? Another three weeks prolongation. How do you square that?’ (Answer: ‘The virus. The virus. We are not protected against the virus.’) followed up with the question ‘Mr van Dissel, you have always insisted that relaxation of measures can only take place when ICU’s are less full and we have enough testing capacity. Does this relaxation [the partial opening of primary schools] not go against your own guidelines?’ (Answer: ‘No.’)