2021 03 18

Notes on meeting

• working on hon contract; DW chasing
• discussed about rough methods plan for 2021; aiming to recreate the ICNARC risk prediction model for UCLH
• this will then help think about making more specific risk prediction models for obs etc.

Notes on the last week

• Bart's already collecting data on REDCap
• Someone already working on this

Risk prediction model

1. Replicate risk prediction model for ICNARC in UCH critical care
2. Write code for live risk prediction for patients coming into critical -- Use that for m&m meetings, unexpectedly good or bad outcomes -- Also for patients not normally accounted for: -- 50% from age and 50% from physiology numbers --- E.g. young with cancer ---- 25 year old post transplant that gave 5% mortality -- E.g. maternal heart disease
3. Can bootstrap general critical care risk model as a PRIOR -- Then adjust that for maternal parameters