2024‐03‐28

Meeting with Matt and Abi after upgrade

Abi will share today or tomorrow.
- I will share with everybody, including Steve and Matt.
- Ask for comments back, propose meeting. Need Nicolette + Daniel only.
- Agenda: Agree journal. Agree final changes.
- Also email Otavio, ask if he'd like to be author. In parallel email on main thread, say he helped me, and ask if ok to include.

I've already shared slides with a clean version of the objectives.
But we need a plan for when I should pause the CCAA work and try the analysis on CC-HIC. Ideally, I should do this after completing a piece of work. I will email about when we should revisit this (ideally middle of April or end of May).

I will try to improve my descriptions of methods by including a running appendix of methods and reasoning, which may not need to go in the main paper.
For the prediction model, I will go with the three group outcome as per Karla's suggestion. We will also think about timepoints, and the possibility of multiplying model probabilities together for sequential days.
Karla also suggested we incorporate the prediction and target trial emulation task into a single target trial. We think this won't work because the outcome will be ‘reducing duration of prolonged ventilation’, rather than ‘identify patients at risk of prolonged ventilation, and then try to reduce it’. This will have the same issues as existing RCTs.
The other suggestion was to do the target trial work first. I'm planning to do that, since the first emulated trial does not depend on the MV model.
I can then work on improving the MV model structure more slowly.

Matt will share some material. I will read, and try to propose a trial structure on a new document (with an appendix!).