This repository includes code from Farah Shamout's DPhil titled "Machine Learning for the Detection of Clinical Deterioration on Hospital Wards" completed at the University of Oxford.
The work involves three types of models:
- Threshold-based Early Warning Score
- Deep Interpretable Early Warning Score
- Multi-modal Early Warning Score
Steps to run the code:
-
create_dummy_dataset.py in datasets folder- this script generates a dummy dataset of the most routinely collected vital signs (HR, SBP, SPO2, TEMP, AVPU, & masktype). masktype indicates the provision of supplemental oxygen
-
navigate to the score of interest and run its respective scripts
Note that this repository will be constantly updated and is not final yet...
Please cite: [coming soon]
Thanks to my supervisors Prof. David Clifton and Dr. Tingting Zhu and clinical collaborator Dr. Peter Watkinson.