Yangyang Dai

[bensoltoff]

• Are the complications a result of being transferred to ICU or an artifact of the underlying reason for transfer?
• Nice modeling strategy
• Seems like it doesn't quite tackle the more important question. Who should be transferred to ICU, rather than who was transferred to ICU? The target variable is still based on actual transfers based on the old model.

[tonofshell]

What unique challenges did you encounter while trying to obtain the patient data used in your study?

[rickecon]

• Is this to get you into the ICU sooner or is it to avoid going to the ICU?
• Why 10-fold cross validation? Did you try bootstrapping?
• Nice horse race among models.
• Nice comparison slide to the PEWS model.
• Very good variable importance plot.

[hanjiaxu]

I have a similar question. Does the model predict infants to transfer to ICU or transfer out of ICU?