Artificial intelligence to improve outcomes after surgery
$1,179,424
Coordinating Principal Investigator: Professor Graham Hillis
Investigator Team: Professor Ferdous Sohel, Dr Janis Nolde, Professor Markus Schlaich, Dr Tim Bowles, Associate Professor Frank Sanfilippo, Mr Adam Lloyd
Associate Investigators: Mr Ben Horgan
Every year more than 200 million people undergo major non-cardiac surgery worldwide. These operations save and improve lives, but they are not without risks. The most common complications after surgery are cardiovascular (problems with the heart or the blood vessels). About one in three patients suffer some damage to the heart after non-cardiac surgery. Whilst this is often very minor, it is associated with an increased risk of death and long-term health adverse outcomes for the patient.
Complications after surgery are generally difficult to predict and hard to prevent. This research project aims to address this difficult problem using an artificial intelligence technique called machine learning, where computers 'learn' from the data they receive and improve their predictive accuracy. These techniques can continually update risk estimates as new data becomes available.
Using state-of-the-art technology based within Royal Perth Hospital’s Health in a Virtual Environment (HIVE), the team will use HIVE’s machine learning and monitoring technology capabilities to test and improve these innovative and novel risk prediction techniques. The team will develop a system that not only predicts a patient's risk before surgery but also continuously updates the patient’s risk during and after their operation. This will also help identify patients whose risk is increasing and warn doctors and nurses looking after them.
The team aims to produce a user-friendly and clinically useful software program that can be adapted for use in any hospital, and which automatically predicts the risk of heart and other complications after surgery, updates this continuously as new information becomes available, and warns clinicians if the risk is increasing. Most importantly, this first-of-its-kind research could provide an effective and low-cost means to improve the outcomes of the many millions of patients undergoing surgery every year and help save lives.