Finn Catling

4

14

3.2K

Finn Catling@FinnCatling·2 Eki

A very generous write-up from the @VPH_Institute following my award vph-institute.org/news/vphi-best… Already looking forward to VPH2026!

English

0

3

326

Finn Catling retweetledi

Anne Scheel@annemscheel·9 Tem

Me after learning about collider bias

English

22

121

51.6K

Finn Catling@FinnCatling·9 May

4/4 Thanks to my brilliant coauthors @MyuraNagendran @PFestor @ZuzieBien @drstevok @AnalogAldo @agordonICU @matkomorowski

English

2

169

Finn Catling@FinnCatling·9 May

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/10.110…

English

3

334

Finn Catling@FinnCatling·9 May

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/ful…

English

4

14

3.2K

Finn Catling retweetledi

Critical Care Reviews@CritCareReviews·23 Eki

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

English

3

5

8

5K

Finn Catling@FinnCatling·21 Eki

Braving the rain on the #artery22 run this morning! @slonold42 @ArterySociety @VascAgeNet

English

3

14

0

Finn Catling@FinnCatling·23 Ağu

@JohsEnevoldsen @PeterHCharlton @BeharJoachim @mmejiaeli @prof_mant @RCBE_City We have been using AMPD for finding ABP systole onset in github.com/UCL-Chimera/si… , but will look at switching to MSPTD in light of your results Pete! Especially as the memory overhead is smaller

English

2

0

Johannes Enevoldsen@JohsEnevoldsen·23 Ağu

@PeterHCharlton @BeharJoachim @mmejiaeli @prof_mant @RCBE_City This looks like a great resource. Seems like it should work for ABP as well (?)

English

0

Peter Charlton@PeterHCharlton·22 Ağu

Detecting beats in the photoplethysmogram: a new paper including: - Assessment of PPG beat detectors - Recommended best open-source beat detectors - Toolbox of datasets and code Paper: doi.org/10.1088/1361-6… Datasets + code: ppg-beats.readthedocs.io

English

4

11

0

Finn Catling@FinnCatling·20 Haz

😍

ART

Finn Catling@FinnCatling·18 Haz

@dannyjnwong This is a nice approach: bpspsychub.onlinelibrary.wiley.com/doi/abs/10.111…

English

1

0

Danny Wong (黄永年)@dannyjnwong·18 Haz

Anybody work with 5 star ratings in their #DataScience #RStats work? Eg google ratings. Just looking for a quick primer to account for the skewed distribution and ordinal values. I know the distances between 1 star and 2 stars may not be the same as between 4 star and 5 star.

English

Johannes Enevoldsen@JohsEnevoldsen

0

Finn Catling@FinnCatling·13 Haz

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

We can present the sum of these three effects as a plane (mainly for fun).

English

5

0

Finn Catling retweetledi

Surgery and Cancer@ImperialSandC·10 Haz

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/hi…

English

Jakob Mathiszig-Lee@willtube4food

5

7

0

Finn Catling retweetledi

Tamas Szakmany MBE@iamyourgasman·9 Haz

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

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/s4174…

English

Jakob Mathiszig-Lee@willtube4food

2

3

0

Finn Catling retweetledi

Ben Goldacre@bengoldacre·8 Haz

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.

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.

English

5

26

0

Finn Catling retweetledi

John SW@DrJohnSW·9 Haz

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!

Finn Catling@FinnCatling

Using data from 127134 emergency laparotomies, our new study nature.com/articles/s4174… moves beyond point predictions to show the uncertainty around estimates of mortality risk. Work with @willtube4food @rmoonesinghe and imperial.ac.uk/people/stephen… in @npjDigitalMed @Nature_NPJ

English

1

2

0

Finn Catling retweetledi

Jakob Mathiszig-Lee@willtube4food·8 Haz

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/s4174…

English

11

21

110

0

Finn Catling retweetledi

Danny Wong (黄永年)@dannyjnwong·8 Haz

@FinnCatling @johngeorgedon @willtube4food @rmoonesinghe @npjDigitalMed @Nature_NPJ 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.

English

4

0

Finn Catling@FinnCatling·8 Haz

@johngeorgedon @willtube4food @rmoonesinghe @npjDigitalMed @Nature_NPJ Thanks for your question! Point estimates can accurately show the *average* risk for a group of patients with similar values of the model input variables. The problem is that risk for individuals within the group can be quite different from the average

English

2

4

0

John George K@johngeorgedon·8 Haz

@FinnCatling @willtube4food @rmoonesinghe @npjDigitalMed @Nature_NPJ Congrats @willtube4food! Is your model a visual risk probability density calculator compared with point estimate predictors like the NELA risk calculator? If the confidence intervals are large as you depict, can any point estimate calculator ever be considered accurate?

English