The world is now weirdly full of epidemiology. Which is both a blessing – as pretty well everyone now knows what it is and why it matters – and a curse – too much simplification and desire for categorical answers to analogue questions.
As a Director of Public Health, I now get telephone calls and messages asking me for R values. Yesterday evening, Kirsty Wark was grilling Newsnight guests on their calculation and use.
Last week’s publication of the government’s COVID recovery strategy has mostly been notable for people’s confusion over how many of their parents they are allowed to speak to in the park at once. But it also flagged the establishment of a ‘Joint Biosecurity Centre’ tasked with “an independent analytical function that will provide real time analysis and assessment of infection outbreaks at a community level”. 1
This, obviously, makes a lot of sense. Testing still captures only a fraction of those who have the disease and admission to hospital is variable as an indicator. Death data are more robust, but infection with the COVID virus can take a week to manifest, after which the illness may grumble on for some time, and those unfortunate enough to die of the disease will typically do so around 3 weeks after onset of symptoms. Registration of death may take several more days, and analysis of dates of death then takes some more time after that. Moreover, it is probable that not all COVID deaths are recognised and recorded as COVID on the death certificate.
As a consequence, collation of other, softer data sources is necessary in judging what is happening now rather than what happened a few weeks ago. Many of these are likely to be proxy measures – analysis of internet search terms, for example, or measures of public mixing fed into risk models.
So, here is an option that may be worth exploring as a more direct potential measure in real-time.
Anosmia, loss of sense of smell, even if not unique to this coronavirus, is increasingly recognised as a key symptom of COVID-19. 2,3 During the pandemic, its acute onset verges on pathognomonic. According to Klopfenstein et al. “mean duration of anosmia [in our series of 114 patients] was 8.9 (± 6.3 [1–21]) days and 98% of patients recovered within 28 days”. 2 The Royal College of Pathologists has incorporated it into its diagnostic algorithm. 4 Although the frequency with which anosmia happens is not precisely determined for patients with milder disease, we know anecdotally that it is commonly observed in community as well as in those hospitalised, and it seems reasonable to assume it will have some degree of consistency across the spectrum of severity.
My hypothesis, then, is that tracking the proportion of people currently suffering anosmia would allow us a real-time measure of trend in COVID infections across the community. Which is appealing because it might be done on a mass scale with simple messaging and a text return.
Essentially, we would send a message regularly to thousands of people saying: “Please take the top off a bottle of vinegar – text N if you can’t smell it”. A perfume or some other pungent household substance would serve equally well, but most homes will have vinegar and it offers a degree of standardisation if we focus on one substance. Then we count the number and proportion and, crucially, the trend of responses.
It would take a huge sample to narrow the confidence intervals in a one-off population survey, but regular (weekly? every few days?) repeats would more robustly accumulate trend data. And though fatigue might set in over time, reinforcement and communication of the safety message could sustain responses – particularly as the experience of infection would, for many, be a significant motivating factor.
The beauty of this is that we could potentially track the disease without new apps or gizmos – the technology is already in our pockets and our kitchens.
So, who wants to run the COVID-R-vinegar trial?
1 OUR PLAN TO REBUILD: The UK Government’s COVID-19 recovery strategy. London, HM Government, 2020.
2 Klopfenstein T, Kadiane-Oussou NJ, Toko L, et al. Features of anosmia in COVID-19. Med Mal Infect 2020; : 4–7.
3 Carrillo-larco RM, Altez-fernandez C. Anosmia and dysgeusia in COVID-19 : A systematic review [ version 1 ; peer review : awaiting peer review ]. 2020; : 1–8.
4 Pathologists RC of. Algorithm for symptomatic staff, symptomatic household testing and further actions. 2020.https://www.rcpath.org/uploads/assets/7f828428-ca0a-4d1b-82a9c48a06d5fb32/G215-RCPath-Algorithm-for-symptomatic-staff-household-testing-and-further-actions.pdf (accessed 16 May2020).