2024‐01‐26 - aasiyahrashan/PhD-Aasiyah GitHub Wiki

Meeting with Abi and Matt about the prolonged MV model

Most of the actions are on the document. https://docs.google.com/document/d/1L6F5gK6I0ihrMBOutkTZxfVCGbzy1CWSHZ4Apx8hSds/edit#heading=h.2owqv0fyypeq

• I need to get availability of variables from the source dataset.
• The diagnosis for apache ii probability of mortality should be whatever is used for covid.
• I need to make sure I'm not including patients with something other than covid. Eg, RTA patients who happened to have covid.
• Each entity in the logic model should have a corresponding CCAA variable. If >1, pick the one with best availability
• Fit a forward stepwise logistic regression without doing a training/test split.
• Use complete case analysis for now - can consider multiple imputation later on
• The methods section needs better reasoning. Use D's slides as an example.