Sign in

Sam Abbott

@seabbs.bsky.social
1.5K followers 785 following 1.1K posts

Real-time infectious disease modelling. Developing methods for outbreak response, surveillance, and pandemic preparedness. samabbott.co.uk Come join me on the epinowcast forum: community.epinowcast.org/latest

PostsRepliesMedia
Sam Abbott @seabbs.bsky.social · 01/10/2026
Thank you!
000
Sam Abbott @seabbs.bsky.social · 01/10/2026
See the Github for further issues and in progress work and modify.
000
Sam Abbott @seabbs.bsky.social · 01/10/2026
In a follow up I also aim to tackle spatial mixing more deeply as this is curently a modulated gravity model for both provinces and then within health zones. The two options I am exploring are increasing the flexibility of the modulation or backing the central estimate to be flowminder based.
100
Sam Abbott @seabbs.bsky.social · 01/10/2026
Note in this release I see a spike in Rt just before detection. I don't think this is real and I think is driven by an artefact in the symptom onset data (see the Rt estimates by data set for this). There are likely other issues that I will aim to deal with in future releases.
100
Sam Abbott @seabbs.bsky.social · 01/10/2026
The model code and analysis were drafted by a language model, then reviewed and revised under human oversight. The named authors are responsible for that oversight.
110
Sam Abbott @seabbs.bsky.social · 01/10/2026
This work adds an external view of the current situation, drawing on our understanding of real-time infectious disease dynamics and the infection process behind the observed surveillance counts. We are developing it and encourage feedback, so please get in touch. We support reuse and adaptation.
100
Sam Abbott @seabbs.bsky.social · 01/10/2026
As in our previous work this work has a lot of limitations: epiforecasts.io/BVDOutbreakS... It is also essentially an unfunded side project of mine with all the limitations that brings both in focus and resources.
epiforecasts.io
Limitations | BVDOutbreakSize
Documentation for BVDOutbreakSize
100
Sam Abbott @seabbs.bsky.social · 01/10/2026
This release contains quite a bit of refactoring and attempts to make tools for the Bayesian workflow accessible. It continues to run in Github CI and so that constrains what can be done. Here we have added a Markov-Melded health zone patch model to give lower-level outputs.
100
Sam Abbott @seabbs.bsky.social · 01/10/2026
New release of our external modelling of the BVD outbreak in the DRC which uses public Sitrep data only and is experimenting with very LLM augmented infectious disease outbreak modelling. This release adds health zone level estimates. epiforecasts.io/BVDOutbreakS...
epiforecasts.io
Dashboard | BVDOutbreakSize
Documentation for BVDOutbreakSize
262
Sam Abbott @seabbs.bsky.social · 30/09/2026
Looks pretty great: cdcgov.github.io/AlgebraicEpi...
cdcgov.github.io
Stratified models - AlgebraicEpiModels
021
Sam Abbott @seabbs.bsky.social · 30/09/2026
Ah interesting. Thanks for the detail!
010
Sam Abbott @seabbs.bsky.social · 30/09/2026
Ah interesting! Will check it out.
010
Sam Abbott @seabbs.bsky.social · 29/09/2026
I think the centred and one day window version (thinking back to our paper from Sang Woo park) is unbiased on the mean with some bias maybe on the sd?
110
Sam Abbott @seabbs.bsky.social · 29/09/2026
I don't think there is another approach out there for this so I can see why you had to make this approx! If riskdistributions can take a custom likelihood it could potentially be used in this way as well?
110
Sam Abbott @seabbs.bsky.social · 29/09/2026
Yeah so I think that this will induce bias in the same way right? The work in epidist has been to make a double interval censored model that can fit to summary stats (which are potentially from a biased model): epidist.epinowcast.org/articles/met...
epidist.epinowcast.org
Fitting to published estimates with the meta model
Fitting to summarised, potentially biased, published estimates jointly with individual level data
210
Sam Abbott @seabbs.bsky.social · 29/09/2026
I would guess it is in the uncertainty and the changes over time though for small timesteps the approximation error will be small.
101
Sam Abbott @seabbs.bsky.social · 29/09/2026
It’s interesting you used a convolutito onsets but kept the renewal on expected onsets (is my read). I can see why but the imputation requirement is a bit of a pain. did you explore using infections as a latent pop as we do in epinow2? Do you see this approach as favourable?
110
Sam Abbott @seabbs.bsky.social · 29/09/2026
Particularly interested as I added a model like this to epidist last week!
010
Sam Abbott @seabbs.bsky.social · 29/09/2026
I’m not familiar with rriskDistributions does it allow for adjusting for double interval cesnoring? My read of your methods is this is a separate step from your discretisation scheme but could easily be wrong
310
Sam Abbott @seabbs.bsky.social · 29/09/2026
I would think it would have the biggest impact at the weekly time step
110
Sam Abbott @seabbs.bsky.social · 29/09/2026
Just skimmed looking forward to reading. For interest did you consider using an exact discretisation approach? We have tools for this to inject into stan models but as far as I am aware are the only people that use them! primarycensored.epinowcast.org
primarycensored.epinowcast.org
Primary Event Censored Distributions
Provides functions for working with primary event censored distributions and Stan implementations for use in Bayesian modeling. Primary event censored distributions are useful for modeling delayed rep...
110
Reposted by Sam Abbott
Andrei R. Akhmetzhanov (吳亞克) @aakhmetz.bsky.social · 29/09/2026
Data, code and all Stan models are on GitHub, with a link to the full model output: github.com/aakhmetz/Bun...
012
Sam Abbott @seabbs.bsky.social · 28/09/2026
Thank you - will forward to Nyall!
000
Reposted by Sam Abbott
IDDjobs @iddjobs.org · 25/09/2026
Postdoc (London, UK) Develop new methods for pathogen genomic epidemiology with @ceciletk.bsky.social at Imperial College London @mrc-outbreak.bsky.social More details: iddjobs.org/jobs/2596
026
Sam Abbott @seabbs.bsky.social · 28/09/2026
i.e. "This decomposition mirrors within-host viral dynamics: stochastic viral replication early in infection creates variation in the onset of detectable virus (the latent period), while subsequent shedding kinetics are more consistent across individuals (the burst)"
110
Sam Abbott @seabbs.bsky.social · 28/09/2026
Who also has this really nice preprint (which I am honestly still digesting) that looks like it could have viral loads shoved into it as a bursy mechanism.... bsky.app/profile/skis...
100
Sam Abbott @seabbs.bsky.social · 28/09/2026
Ah, thank you! It is probably a bit me centric as its hard to capture what everyone else is doing at the same time.
010
Reposted by Sam Abbott
Stephen Kissler @skissler.bsky.social · 29/07/2026
I'm excited to share a preprint that's been years in the making: the idea took shape during the COVID pandemic. Anecdotally, people would sometimes spread COVID at work but not at home, and vice versa. Why? One explanation would be if infectiousness is fleeting...
medrxiv.org
How bursty infectiousness shapes epidemic dynamics
An epidemic’s expected course is determined by the magnitude and timing of a typical person’s infectiousness — captured, in turn, by the basic reproduction number and the generation-time distribution....
24417
Sam Abbott @seabbs.bsky.social · 28/09/2026
woops skissler.bsky.social not doing a good job of finding people today!
skissler.bsky.social
Stephen Kissler (@skissler.bsky.social)
Infectious disease epidemiologist / computer science professor @ CU Boulder. Trying to figure out how to keep us all healthy and fed.
100
Sam Abbott @seabbs.bsky.social · 28/09/2026
This has been a big collaborative effort with a huge amount of work from Nyall in particular. Some cool follow ups to this to come including thinking about applying the idea to othe diseases and joint fitting with viral load data. Please reach out if interested in being involved!
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
An obvious question from all of this and one we are still exploring is what happens when you jointly model with viral load data. That is still in the works but for now here is the impact of different prior choices (i.e. informed by viral load data and not).
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
We also found that in a few case studies a viral load based distribution fit the data better than the distribution used as the best fit in the original study and that this changed the estimated parameters in ways that might have impacted downstream modelling and hence decisions.
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
yes you can recover true parameters under simulation
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
The TLDR is that yes you can do this, yes it is fairly elegant (in my opinion), yes you can connect this specialised distribution to common ones with different assumptions,
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
So we did! www.medrxiv.org/content/10.6...
medrxiv.org
220
Sam Abbott @seabbs.bsky.social · 28/09/2026
period we could develop an epidemiologically motivated model for the serial interval, generation time, and all its component delays.
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
In conversations (via Epiengange part of the CDC funded insight net project) Nyall, Laren Meyer and I realised that maybe if we thought about the generation time (and its component distributions) as being driven by within host processes in the same way Nyall had been thinking about the incubation
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
So nothing else to do? Not quite. Alongside all of this a big gap that remains in using common outbreaks models is understanding what the generation time (i.e time from infection to infection) is and how to model it.
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
Coming back to the Charniga work it would have also enabled us to give some advice beyond try a few things and cross your fingers (i.e your a distribution motivated by the within host mechanisms of your disease).
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
Some really nice recent work linking these ideas from Christopher Boyer, Stephen Kissler and @mlipsitch.bsky.social arxiv.org/html/2605.20...
arxiv.org
Inferring infectiousness: a joint model of the within-host viral kinetics of SARS-CoV-2
220
Sam Abbott @seabbs.bsky.social · 28/09/2026
Why? Well we ( Timothy Russell, @adamjkucharski.bsky.social and co) had just been modelling viral kenetics jointly with sypmtom onsets (assuming a lognormal distribution) so a distribution that was viral load motivated would have allowed for very elegant linkage. journals.plos.org/plosbiology/...
journals.plos.org
Combined analyses of within-host SARS-CoV-2 viral kinetics and information on past exposures to the virus in a human cohort identifies intrinsic differences of Omicron and Delta variants
The emergence of successive SARS-CoV-2 variants of concern during 2020-22 created a need to understand the drivers of such growth. This study uses a Bayesian model to reveal how a set of key covariate...
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
As someone who doesn't love Burrrs I did not initially find this that exciting but after reading more closely I realised that the main motivation for the modified Burrr here was an argument that the incubation period is an emergent property of within host viral kenetics which got me very excited!
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
At about the same time as we were doing our thinking Nyall Jamieson, Ian Hall, and co were thinking deeply about how much they loved Burrr distributions for modelling incubation periods. journals.plos.org/ploscompbiol...
journals.plos.org
The Burr distribution as a model for the delay between key events in an individual’s infection history
Author summary In public health, it is important to know key temporal properties of diseases (such as how long someone is ill for or infectious for). Mathematical characterisation of properties requir...
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
The lack of mechanism in that approach is potentially problematic both in that it will struggle under sparse data but also that it can struggle in settings with compound and complex delays such as generation times.
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
This didn't make me very happy. We have been thinking about this in various ways since with for example epidemiological bias corrected non parametric delay estimation methods. primarycensored.epinowcast.org/articles/fit...
primarycensored.epinowcast.org
Fitting non-parametric delay distributions
A guide on how to fit non-parametric delay distributions using primarycensored, with both MLE and Bayesian approaches.
130
Sam Abbott @seabbs.bsky.social · 28/09/2026
A lot of that work solid and I hope has moved the field forward a bit. However, there was a gap when it came to how to select a distribution for fitting to your data. We said try a few, report all of them, and give an indication of which fits best (via best fit metrics etc. etc.)
120
Sam Abbott @seabbs.bsky.social · 28/09/2026
A few years ago we (Kelly Charniga @sangwoopark.bsky.social et al.) did some thinking on how we should estimate and report epidemiological delay distributions (i.e incubation periods etc. etc.). journals.plos.org/ploscompbiol...
journals.plos.org
Best practices for estimating and reporting epidemiological delay distributions of infectious diseases
Epidemiological delays are key quantities that inform public health policy and clinical practice. They are used as inputs for mathematical and statistical models, which in turn can guide control strat...
21510
Sam Abbott @seabbs.bsky.social · 25/09/2026
backed = backend
000
Sam Abbott @seabbs.bsky.social · 25/09/2026
This could all be another round of LLM coding agent induced madness which I find strikes every month or so.
100
Sam Abbott @seabbs.bsky.social · 25/09/2026
I have been trying to work out which is the AD backend to go all in with in Julia as various community and technical tradeoffs involved. Currently thinking ease of understanding when a custom rule would help and writing that rule could be important.
101