Thread by @BarnardResearch

My research (with @_nickdavies @cmmid_lshtm @markjit & John Edmunds) is published in @NatureComms 🥳

This is the culmination of many months of work (my funding from @Epipose)🍾🎉

nature.com/articles/s4146…

Until I write a book on this (publishers/agents hmu😝), here is a 🦣🧵

The code to reproduce this work is available at: github.com/rosannaclaireb… 🛠️⚙️🧩🏗️

This implementation of 'covidm' builds on work by @_nickdavies (the OG creator) in the early #COVID days

Since joining @LSHTM in July 2020, I have developed covidm with @_nickdavies

👩‍💻👩‍🔧👩‍🎨👩‍💼👩‍🔬

Early 2020 discussions @cmmid_lshtm led @_nickdavies to build #covidm

CMMID members @LSHTM (e.g. @rozeggo, John Edmunds, & @adamjkucharski) contributed to these discussions

Many other colleagues contributed to model development (e.g. @cap1024 @Lloyd_Chapman_ @KevinvZandvoort)

The first work I was involved in with #covidm explored the effect of regional tiered #restrictions and the second #lockdown in England (work with @_nickdavies @jarvisc @timwrussell @ProfCalumSemple @markjit & John Edmunds, ofc!)

sciencedirect.com/science/articl…

@cmmid_lshtm @CCPUKstudy

I wrote a Twitter thread 🧵 on that paper here:

https://twitter.com/BarnardResearch/status/1342087036241666049

Not long after, I was involved in work led by @_nickdavies using #covidm (and other methods) to explore the impact of the Alpha B.1.1.7 variant of concern (#VOC) on #SARSCoV2 transmission in England 🦠❤️‍🩹🏴󠁧󠁢󠁥󠁮󠁧󠁿

This is now published in @ScienceMagazine ⤵️

science.org/doi/10.1126/sc…

At this point #covidm was implemented as a two-strain or two-variant model with vaccination (see attached image)

This allowed us to model the Alpha VOC taking over from previous #SARSCoV2 variants in England, as well as the mitigating effect of #COVID19 #vaccines on transmission

In January 2021, I built extra compartments into #covidm so that we could explore the effect of vaccination in England in more detail, such as the timing and number of vaccine doses received.

The model now included:

2⃣ vaccine products (viral-vector; mRNA) &
2⃣ vaccine doses 💉

Next, we (me @_nickdavies @markjit & John Edmunds) used #covidm to produce scenarios assessing the potential impact of relaxing restrictions (the so-called #roadmap out of lockdown in England) 🔒👎

These scenarios provided evidence to decision makers ⤵️

gov.uk/government/pub…

After the (somewhat traumatic) experience of rushing to model the Alpha B.1.1.7 VOC, and living through the resulting lockdown, in March 2021 I decided to add some latency into #covidm…

This took the form of a third SARS-CoV-2 variant (use in case of emergencies!🦺)

2⃣➡️3⃣

Over the next few months, we (cc @_nickdavies @markjit & John Edmunds) produced lots more roadmap modelling scenarios, coinciding with each staged relaxation 👐☺️

In May 2021, the third modelled variant was already in use for hypothetical VOC scenarios:

gov.uk/government/pub…

Roadmap step 3⃣ went ahead as planned

By June 2021, the Delta B.1.617.2 VOC was spreading in England ... 🦠

We fitted our model to genomic sequencing data for Delta, explored scenarios for characteristics of Delta & considered step 4⃣ of the roadmap ⤵️

gov.uk/government/pub…

Roadmap step 4⃣ was delayed from its original planned implementation date.

That meant we needed another round of modelling looking at the impact of Delta B.1.617.2 and delayed step 4⃣

I wrote another Twitter thread 🧵 then (which links to the report ⤵️)

https://twitter.com/BarnardResearch/status/1414636107548045317

Whilst posting our research *output*, I want to note how hard it was in reality...

We (shoutout @_nickdavies @markjit John Edmunds) were often working late and under a lot of pressure 😑🗜️🤯‼️

[video recorded at 01:19 on 7th July 2021; I finished the report around 4am😪]

So after all of that, I took some time to rest, reset, and figure out my next steps 🌴🌊😎🌄👪🐈‍⬛

In the Autumn of 2021, we spent more time on #covidm development:

▶️ adding Va2w & Vb2w compartments (allowing more accurate modelling of boosters & waning)
▶️ updating @ONS population estimates
▶️ new method for historic time-varying transmission adjustments to better fit data

Then we provided more evidence to UK decision makers: autumn/winter 2021-22 scenarios & contingency ("Plan B") measures...

📌booster vaccines, waning vaccine protection
📌vaccination of 12-15 year old children
📌& assuming the delta VOC remains dominant

gov.uk/government/pub…

Finally, it was time for me to focus on the metrics that #academia cares about: first-author peer-reviewed publications 🥁📜🏆🏅🍾👑💥

I started to wrap up the aforementioned work into a paper exploring key factors influencing dynamics of #SARSCoV2 transmission in England, projected forwards to September 2022 🔐

The key factors turned out to be

🍏behaviour change,
🍏booster vaccinations, and
🍏waning immunity

I sent the paper off to peer-review and submitted the preprint to @medrxivpreprint on 29th November 2021

👏🤞👇🤘

medrxiv.org/content/10.110…

Six days later, we (me @_nickdavies @cap1024 @markjit & John Edmunds) had to start looking at #Omicron scenarios 🤯

(we firstly called these 'Nu' scenarios, as we started them before the @WHO named Omicron!)

My Twitter thread links to that report here:

https://twitter.com/BarnardResearch/status/1469646833563615237

Switching back to my paper under peer-review in late 2021 🔙

Our peer-review reports were returned later in December ...

TLDR: the reviewers liked the work, but the Omicron variant was appearing likely to have a significant impact on the relevance of our modelling results

🤐🫠

In 2022 me, @_nickdavies, @markjit, & John Edmunds decided to overhaul the paper, exploring the same questions

i.e. what are the factors influencing projected #SARSCoV2 transmission dynamics in England

& considering the huge wave of Omicron transmission that begun in late 2021

By 24th March I had everything reproduced (data built, model fitting routines run, model projections generated, Figures produced, paper rewritten, response to peer-reviewers written) 😅

But..... Omicron BA.2 😰

bbc.co.uk/news/health-60…

This threw another spanner in the works 🔧

(Wow, just realised there is a 25-tweet drafting limit 🤯)

This rise in infections driven by #Omicron BA.2 contradicted our results in March, which assumed that #Omicron BA.1 would remain dominant, & projected a reduction in #COVID19 transmission in England until September 2022

John Edmunds thought the work was becoming "increasingly anachronistic" (a beautiful turn of phrase which I couldn't forget 😂)

So we buckled up and worked to:

🧰incorporate @sangerinstitute data to inform BA2 taking over from BA1
🧰extend modelling projections to December 2022

By May, the paper was finally ready 🍾🪩🥳🥲

We resubmitted the paper to be peer-reviewed and uploaded the pre-print to @medrxivpreprint

(the pre-print version is here medrxiv.org/content/10.110…)

I took some time off and got to fangirl over @_Warpaint at @dlwp 😊

After revising our paper in June to adhere to editorial requirements, we received news that @NatureComms had formally accepted the paper on 25th July 🎉🆒

I've linked the paper in tweet number 1 of this 🦣🧵

Thanks to @_nickdavies @cmmid_lshtm @markjit John Edmunds & @EpiPose