ICU doctor. @wellcometrust PhD researching personalised care using physiological models, Bayesian statistics & machine learning. Jazz piano enthusiast.
Joined March 2015
- Tweets 265
- Following 387
- Followers 404
- Likes 1,186
7 Photos and videos
Pinned Tweet
9 May 2024
1/4 Could AI personalise sepsis resuscitation? Our new paper in @CritCareExplore identifies technical and translational barriers, and recommends solutions to turn promise into patient benefit.
journals.lww.com/ccejournal/…
2
4
14
3,260
2 Oct 2024
A very generous write-up from the @VPH_Institute following my award vph-institute.org/news/vphi-… Already looking forward to VPH2026!
1
3
331
Finn Catling retweeted
9 Jul 2024
Me after learning about collider bias
2
22
120
51,652
9 May 2024
1/4 Could AI personalise sepsis resuscitation? Our new paper in @CritCareExplore identifies technical and translational barriers, and recommends solutions to turn promise into patient benefit.
journals.lww.com/ccejournal/…
2
4
14
3,260
9 May 2024
3/4 For a deeper dive into the merits of individual AI systems for sepsis treatment, and a clinician-friendly intro to how they work, take a look at our recent scoping review.
medrxiv.org/content/10.1101/…
1
1
3
338
9 May 2024
4/4 Thanks to my brilliant coauthors @MyuraNagendran @PFestor @ZuzieBien @drstevok @AnalogAldo @agordonICU @matkomorowski
2
172
Finn Catling retweeted
23 Oct 2023
In a Bayesian trial investigating a safe and inexpensive treatment (statins in COVID-19), what probability of a beneficial effect would be enough to change your practice:
a) 21-day organ support-free days, incorporating mortality as the worst outcome
@dfmcauley @agordonICU
47%
>75%
28%
>90%
19%
>95%
7%
>99%
75 votes • Final results
3
5
8
5,053
21 Oct 2022
2
3
14
13 Jun 2022
What an informative (and pretty!) visualisation: steady-state CVP modelled as a function of position in the cardiac and respiratory cycles, plus the interaction between the two
13 Jun 2022
We can present the sum of these three effects as a plane (mainly for fun).
2
2
5
Finn Catling retweeted
10 Jun 2022
New paper published in @npjDigitalMed highlighting uncertainty in emergency bowel surgery prediction models from
S&C's @FinnCatling @willtube4food Prof Stephen Brett and UCL's @rmoonesinghe
Find out more ⬇️
imperial.ac.uk/news/237135/h…
5
7
Finn Catling retweeted
9 Jun 2022
A very elegant and clinically important modelling paper. “At work” we rely on the point estimates derived from the NELA score, making clinical decisions. But how does it influenced by missing values and how those result might change the point estimate? @mbmuthuswamy @SCCM_Anesth
8 Jun 2022
I'm super proud of the paper myself and @FinnCatling co-first authored along with excellent supervision from @rmoonesinghe and Professor Stephen Brett.
Over 127,000 patients from @NELANews used to create a novel mortality risk calculator. Open Access.
nature.com/articles/s41746-0…
1
2
3
Finn Catling retweeted
8 Jun 2022
Great thread. Open code. Open tools. Clinicians who understand data, and how to put it into action, outside the narrow outdated 1990s model of publishing an epidemiology paper as a PDF and then walking away.
8 Jun 2022
You can view the final model at laparotomy-risk.com. There's a web calculator and an API, something we wanted from the outset. We couldn't compare our work with the ACS NSQUIP calculator because it didn't have one.
I want to highlight a few things I'm particularly proud of.
1
5
26
Nice paper, highlights challenges of risk prediction and communication.
The visualisation that this model allows could really help in discussions with patients and colleagues.
Admirable commitment to open source and data sharing.
Congratulations to all involved!
8 Jun 2022
Using data from 127134 emergency laparotomies, our new study nature.com/articles/s41746-0… moves beyond point predictions to show the uncertainty around estimates of mortality risk.
Work with @willtube4food @rmoonesinghe and imperial.ac.uk/people/stephe… in @npjDigitalMed @Nature_NPJ
1
2
Finn Catling retweeted
8 Jun 2022
I'm super proud of the paper myself and @FinnCatling co-first authored along with excellent supervision from @rmoonesinghe and Professor Stephen Brett.
Over 127,000 patients from @NELANews used to create a novel mortality risk calculator. Open Access.
nature.com/articles/s41746-0…
11
21
109
Finn Catling retweeted
8 Jun 2022
This is exactly why I think "personalised medicine" (whatever that means) can't ever precisely predict how *individuals* will fare. We can only say how a group of patients that look like the patient in front of us do, on average.
1
1
4
8 Jun 2022
Using data from 127134 emergency laparotomies, our new study nature.com/articles/s41746-0… moves beyond point predictions to show the uncertainty around estimates of mortality risk.
Work with @willtube4food @rmoonesinghe and imperial.ac.uk/people/stephe… in @npjDigitalMed @Nature_NPJ
3
14
39
Finn Catling retweeted
6 Dec 2021
Massive Congratulations to @FinnCatling @benebenve and @NeunMax, today’s Transfer Team for completing NCL-ACTS 800th Critical Care Transfer!! Thank you so much to all our staff who’ve made this incredible milestone possible! @ArndtMelzer @RiazAziz_ @uclh @GBellingan @NECCTARTeam
2
8
46
13 Oct 2021
Look forward to presenting, and listening to some great speakers, tomorrow afternoon!
2
Finn Catling retweeted
26 Sep 2021
Great summary by @ioannides_alex about the power of statistical modeling vs non-interpretable ML models: linkedin.com/feed/update/urn… Based on "General Pitfalls of Model-Agnostic Interpretation Methods for Machine Learning Models" arxiv.org/abs/2007.04131
3
44
167