Chapter 10 Clinical Implication

10.1 ML in clinical settings

We could have the following uses of ML methods in the clinical settings:

  • prescreen patients to identify high-risk patient pool.
    • warn patients about imminent risk
    • helps manage clinical workload
  • help diagnose a disease better with high accuracy
    • could be based on radiology or pathology images
    • could prevent mis-diagnose by giving a second opinion
    • could indicate suspicious regions, assisting clinicians to focus on the most important considerations
  • Monitoring vulnerable patients
    • monitoring devices (e.g., fall detection)
    • ethical, moral and transparency considerations

10.2 Model updating

  • Updating risk prediction model based on new data
    • could be automated given access to continuously collected data
  • policy could change, requiring the update of the model