I took some leave earlier in 2020 around the time I presented some Dengue work in Annecy at the Fondation Mérieux (check spelling). When I got back I was asked if I could help Billy Quilty (who I’m now supervising as a PhD student) with some analysis he was doing about traveller screening for the novel coronavirus that had appeared in China a few weeks prior. This ended up kicking off my involvement with the CMMID COVID-19 Working Group (formerly nCoV-2019 Working Group).

It’s been a bit of a whirlwind over the last few months. I’ve been writing thousands of lines of code, writing papers, internally reviewing others’ papers, getting brought on to others’ papers, checking others’ code, arguing about visualisation styles, trying to nail down terminology, and building apps to demonstrate our models. There’s too much to cover but I thought I’d just highlight a few things I’ve been involved with.

Basically the entirety of my research time is spent on COVID at the moment.

Syndromic screening

Billy J Quilty, Clifford, et al. (2020a) was the first piece of work that came out of the coronavirus working group. Using the best available information at the time, which wasn’t much, we looked at how likely syndromic screening (thermal scanning and looking for symptoms) was to pick up travellers infected with SARS-CoV-2. We built an app that would allow users to vary the sensitivity of the syndromic screening to see how many passengers would arrive with an active infection. One of the big challenges with this work was, and still remains, accurate characterisation of the proportion of infections which are asymptomatic.

Conclusion: syndromic screening is insufficient to reduce the number of imported infectious arrivals.

Travel restrictions in China

Billy Quilty and Charlie Diamond, two early career researchers at CMMID, did some outstanding work looking at the impact of internal travel restrictions around Lunar New Year in China (Billy J. Quilty et al. 2020). This involved a lot of work and assumptions regarding travel patterns and branching processes for transmission.

Conclusion: sustained transmission was likely occurring in four major cities prior to the introduction of restrictions and these restrictions likely only helped minor cities.

Delaying an outbreak

We updated the syndromic screening and combined it with some work by our colleagues on contact tracing (Hellewell et al. 2020) to build a Poisson process for traveller arrivals (Clifford, Pearson, et al. 2020). Here we explicitly accounted for the increase in cases at the travel origin and considered how long it would take to reach the outbreak threshold under varying syndromic screening and contact tracing scenarios.

Conclusion: syndromic screening and effective contact tracing can buy you less than two weeks early in the outbreak.

Risk of importation

More recently I contributed some plots and code to work that looks at estimating the number of local cases in a country and then using OpenSky flight data to estimate how many infectious travellers are moving between countries (Russell et al. 2020). We then calculated the number of imported infections as a fraction of local cases to determine which countries were at risk of triggering a second wave based on imported cases.

Conclusion: countries considering flight restrictions from countries where they have a lot of travel from and/or there is high prevalence of infection should also consider their local outbreak situtation and act accordingly.

Of course, flight restrictions aren’t the only non-pharmaceutical intervention you can apply when looking at risk of importation. The next step was to build a model of quarantine effectiveness to determine whether the use of PCR testing could give the same sort of performance as a 14 day quarantine period in the UK (Clifford, Quilty, et al. 2020). This involved combining CAA air travel data with prevalence estimates (as in the Russell work) and looking at the proportion of travellers who were asymptomatic, applying models to published data on PCR testing of nasopharyngeal swabs, and building clinical progression profiles for simulated travellers. This was a mammoth amount of work and is still under review but forms the basis for further work.

Conclusion: Double testing doesn’t give you much benefit over single testing, and quarantine can be reduced by 5-7 days if you do a test a day before travellers exit quarantine.

Local quarantine

We are adapting our traveller quarantine work to look at the UK’s Test and Trace programme to see if it’s possible to reduce the amount of time that traced secondary contacts spend in quarantine. We’ve got a preprint that’s been presented (Billy J Quilty, Clifford, et al. 2020b) but are continuing to look into a few issues around adherence waning.

Conclusion: So far we have found that (if you assume perfect compliance and adherence to quarantine rules) you can reduce quarantine duration by a few days if you test at the end of quarantine.

Other work

In addition to work around traveller screening and quarantine I’ve also been involved with some work on hospital length of stay (Rees et al. 2020) and this has flowed through to work on bed occupancy (Jombart, Clifford, et al. 2020). This is mostly with Thibaut Jombart, with whom I did some work looking at projecting early case numbers from deaths (Jombart, Van Zandvoort, et al. 2020). We are attempting to finish off a paper on the bed occupancy forecasting work.

I’ve also done some work with Yang Liu on her paper about surveillance in hospitals and the community, determining where testing effort is best spent in terms of being able to pick up a signal for an outbreak (Liu et al. 2020). This was some really interesting work, and we had to revise the model a few times, and should be useful for countries or regions who have managed to get their outbreaks under control and are planning regular testing for detection of a second wave.

Other news

Additionally, my wife and I found out she was pregnant around March this year, and we are expecting a daughter in the next few weeks. So interacting with the NHS has been difficult at times, as appointments get cancelled and rescheduled, and birth partners aren’t allowed at routine scans.

In other happy news, I’ve been promoted to Assistant Professor and have just finished all the assessment for the first module of PGCILT. PGCILT deserves its own write-up. Maybe I’ll just upload my 3000 word reflection.


Clifford, Samuel, Carl A B Pearson, Petra Klepac, Kevin Van Zandvoort, Billy J Quilty, Rosalind M Eggo, and Stefan Flasche and. 2020. “Effectiveness of Interventions Targeting Air Travellers for Delaying Local Outbreaks of SARS-CoV-2.” Journal of Travel Medicine 27 (5). Oxford University Press (OUP).

Clifford, Samuel, Billy J Quilty, Timothy W Russell, Yang Liu, Yung-Wai Desmond Chan, Carl A B Pearson, Rosalind M Eggo, Akira Endo, Stefan Flasche, and W John Edmunds and. 2020. “Strategies to Reduce the Risk of SARS-CoV-2 Re-Introduction from International Travellers,” July. Cold Spring Harbor Laboratory.

Hellewell, Joel, Sam Abbott, Amy Gimma, Nikos I Bosse, Christopher I Jarvis, Timothy W Russell, James D Munday, et al. 2020. “Feasibility of Controlling COVID-19 Outbreaks by Isolation of Cases and Contacts.” The Lancet Global Health 8 (4). Elsevier BV:e488–e496.

Jombart, Thibaut, Sam Clifford, Carl A. B. Pearson, Erees, Esnightingale, Mert0248, and Gwen Knight. 2020. Zenodo.

Jombart, Thibaut, Kevin Van Zandvoort, Timothy W Russell, Christopher I Jarvis, Amy Gimma, Sam Abbott, Sam Clifford, et al. 2020. “Inferring the Number of Covid-19 Cases from Recently Reported Deaths.” Wellcome Open Research 5. The Wellcome Trust.

Liu, Yang, Wenfeng Gong, Samuel Clifford, Maria E Sundaram, Mark Jit, Stefan Flasche, and Petra Klepac and. 2020. “A Modelling Study for Designing a Multi-Layered Surveillance Approach to Detect the Potential Resurgence of SARS-CoV-2,” June. Cold Spring Harbor Laboratory.

Quilty, Billy J, Sam Clifford, Stefan Flasche, and Rosalind M Eggo and. 2020a. “Effectiveness of Airport Screening at Detecting Travellers Infected with Novel Coronavirus (2019-nCoV).” Eurosurveillance 25 (5). European Centre for Disease Control; Prevention (ECDC).

Quilty, Billy J, Samuel Clifford, Stefan Flasche, Adam J Kucharski, and W John Edmunds. 2020b. “Quarantine and Testing Strategies in Contact Tracing for SARS-CoV-2,” August. Cold Spring Harbor Laboratory.

Quilty, Billy J., Charlie Diamond, Yang Liu, Hamish Gibbs, Timothy W. Russell, Christopher I. Jarvis, Kiesha Prem, et al. 2020. “The Effect of Travel Restrictions on the Geographical Spread of COVID-19 Between Large Cities in China: A Modelling Study.” BMC Medicine 18 (1). Springer Science; Business Media LLC.

Rees, Eleanor M, Emily S Nightingale, Yalda Jafari, Naomi Waterlow, Samuel Clifford, Carl A B Pearson, Thibaut Jombert, Simon R Procter, and Gwenan M Knight and. 2020. “COVID-19 Length of Hospital Stay: A Systematic Review and Data Synthesis,” May. Cold Spring Harbor Laboratory.

Russell, Timothy W., Joesph Wu, Samuel Clifford, John Edmunds, Adam J Kucharski, and Mark Jit. 2020. “The Effect of International Travel Restrictions on Internal Spread of COVID-19,” July. Cold Spring Harbor Laboratory.

comments powered by Disqus