Marion Campbell

CCR Down Under 2025 Critical Care Reviews Meeting Down Under - The Edge, Federation Square, Melbourne, December 9th - 10th, 2025

🚩The 3rd trial result for CCR Down Under 2025 is ARISE-AFRICA🚩

➡️Standard low flow oxygen vs nasal high flow oxygen vs CPAP in hypoxaemic respiratory failure

➡️Presenter Arthur Kwizera

➡️Join us in Melbourne
criticalcarereviews.com/meetings/ccr...

@precordialthump.bsky.social

CCR Down Under 2025 Critical Care Reviews Meeting Down Under - The Edge, Federation Square, Melbourne, December 9th - 10th, 2025

CCR Down Under 2025 - december 9th & 10th 2025 Melbourne is the place to be! Registrations are open 😁: criticalcarereviews.com/meetings/ccr... See you there! @criticalcarereviews.com

CCR Down Under 2025 Critical Care Reviews Meeting Down Under - The Edge, Federation Square, Melbourne, December 9th - 10th, 2025

The first trial for CCR Down Under 2025 is RSI from the Pragmatic Critical Care Research Group

➡️Ketamine vs etomidate for emergency tracheal intubation
➡️n= 2364

Join us in Melbourne - Dec 9 & 10
criticalcarereviews.com/meetings/ccr...

CCR Down Under is run with the Alfred ICU & Monash Uni

It is helpful to show the impact of borrowing by presenting trial results with and without the borrowed information. Presenting sensitivity analysis depending on the level of influence placed on the borrowed information is also useful. 6/6

One recent trial that used Bayesian borrowing is the RESCUE-3 trial (trial of transvenous diaphragm neurostimulation), which borrowed information from an earlier trial (RESCUE-2) – it downweighted with a propensity-weighted power prior approach. 5/6
www.atsjournals.org/doi/abs/10.1...

Whilst the potential of have increased information is good, one must take care when borrowing information from other sources. You need to be reassured that the external data and the current study data are sufficiently similar so that it cannot bias the analysis. 4/6

A common approach is to “borrow” (ie use) historical data, such as from an earlier trial, to inform the Bayesian prior for your new trial. You can downweight the influence of the previous data (from 0 – no borrowing to 1 - full borrowing) – this is often called a power prior 3/6

Bayesian borrowing refers to using  information from external sources to create an informative prior to use in the Bayesian analysis of your trial. Using this additional information can increase the overall power of a study and potentially help give more precise estimates of effect. 2/6

We often now hear the term “Bayesian borrowing” in trials. But what is Bayesian borrowing and what are the pros and cons around its use? 1/6
#MethodologyMonday #122

20 seconds of the sounds of the North Atlantic this morning (by Upper Loch Torridon).

Little by little one travels far.
Writing one sentence or paragraph seems so little but that's how papers and chapters and books get written.
Take the next step.
Write the next sentence.
Do the next thing.
#PhDchat #ECRchat #postdoc #gradschool

See below for our Greener Trials Carbon Footprinting Drop In clinics...

Do you experience chronic or recurrent UTIs, or care for those who do? Take this survey powered by surveymonkey.com. Create your own surveys for free.

@jameslindalliance.bsky.social priority setting survey for UTI, looking for responses from patients, clinicians and stakeholders. Pls help share to ensure a representative sample #UTISKY

@jennyrohn.bsky.social @angelahuttner.bsky.social @gpollara.bsky.social

www.surveymonkey.com/r/UTI_PSP

Here’s a new interesting thing for a Monday morning…

What's the word for that feeling when you realise that something you thought was a strange coincidence in your particular trial is actually a well described phenomenon?! 😅

Fascinating thread of an instance of what we behavioural scientists call the "planning fallacy" with reference to clinical trials.

Little considered causes:

- Lack of a proper pre-assessment of the workload for clinicians (medics, nurses). Workload starts being overwhelming & recruitment stops

- Clinicians being asked to do additional, technical roles for the RCT: data entry (apart from the usual clinical entries), lab work

NB - autocorrect has been working hard this morning - all should be Lasagna (not Lasagne!) 🙄

Trial staff can also have early inklings that certain sites may not recruit well. Potential red flags have been identified incl. poor site engagement, strong site preferences etc. Ensuring sufficient high performing sites are included is essential 7/7
trialsjournal.biomedcentral.com/articles/10....

Undertaking thorough feasibility work prior to the trial is needed to minimise the possible effects of Lasagne’s Law. Estimates of throughput, consent etc are often optimistic - objective data from real-life piloting of the trial processes are invaluable 6/7

As Bogin notes in a paper discussing this … Lasagna’s Law is a dish best served early! You want to find out as early as possible what the true rate of recruitment to your trial will be. 5/7
www.sciencedirect.com/science/arti...

There can be many reasons for this - throughput estimates were too high, inclusion criteria are more restrictive than anticipated, eligible patients can be missed, some site clinicians may not have equipoise, the trial processes may not be as acceptable as hoped etc 4/7

It is named after Louis Lasagna - a pharmacologist - who had observed that the number of patients who are actually recruited to a trial is often nearer 10%-30% of the initial numbers usually estimated as available 3/7

“Lasagna’s Law” describes the often observed phenomenon that as soon as you start your clinical trial the number of potentially eligible patients you expect to see/recruit suddenly appears to plummet. 2/7

Over recent weeks I have heard about some trials struggling with the phenomenon known as “Lasagne’s Law.” I’ve flagged this issue before, but always good to revisit it and discuss what can be done to minimise it 1/7
#MethodologyMonday #121

Excellent #MethodologyMonday post on composite outcomes. Everyone knows why they’re used but wishes they weren’t.

As such care needs to be taken before deciding whether to use a composite and, if so, which one. A neat summary of the pros & cons is outlined in this paper by Sue Ross 8/9
sciencedirect.com/science/arti...

Another specific issue is when the observed direction of effect in the separate components is inconsistent - say an increase in one but a decrease is another (leading to the composite returning an incorrect null effect as the effect of one component “cancels out” the other) 8/9

Potential solutions such as the “win-ratio” method have been proposed to accommodate for the situation where there is a hierarchy in the composite components (however this method is not without its own problems) 7/9
doi.org/10.1093/eurh...

This is exacerbated if the events in the component of greater importance is small compared to the number of events in the components of lesser importance 6/9
bmj.com/content/334/...