Even a simple treatment that seems safe may have a small chance of going wrong. The risk, though, is different for each person.
The Singapore General Hospital (SGH) has made a risk prediction tool that uses artificial intelligence (AI) which can tell how much care a patient needs and if there could be problems after surgery, or if it would be better to put off the surgery.
CARES-ML, which stands for Combined Assessment of Risk Encountered in Surgery-Machine Learning, has been tested and has shown that it can make accurate predictions more than 90% of the time. It pulls a patient’s medical history, physical state classification, and investigative test results like x-rays and blood tests from the hospital system to make a surgery risk report and score. The risk of something going wrong after surgery goes up with the score.
Associate Professor Hairil Rizal, Senior Consultant, Anaesthesiology, SGH, said, “CARES-ML helps the anesthesiologist and surgeon evaluate each patient and improves the clinical team’s decisions and recommendations about the patient’s perioperative plan of care.” In the end, this makes patients safer and improves their health.
When a patient is set to have surgery, they must first be checked to see if they are ready. This review is usually done about 10 days before surgery by an anesthesiologist and a surgeon. This is done so that the right steps can be taken at the right time, such as treating anaemia that the patient may not be aware of. Patients in this situation are more likely to get sick, have a stroke, or hurt their kidneys.
CARES-ML may be used to supplement the pre-surgery assessment in the future. It is based on the Hospital’s CARES calculator, which was established in 2018 to assess surgical risk in the local community and has been updated and validated once again using a local dataset of approximately 100,000 surgical patients from 2015 to 2022.
While the tool is being evaluated for its efficacy and impact on surgical outcomes, the team of anaesthesiologists, surgeons, and biostatisticians is already working on expanding the model to predict additional outcomes such as duration of hospital stay, risk of pneumonia, and stroke.
They are also considering using generative AI to assign a patient’s physical state classification level, which CARES-ML also uses for prediction. In another study based on 10 standardised hypothetical patient scenarios, Prof Hairil and his team discovered that the AI chatbot was comparable in terms of accuracy, but more consistent than human assessors.
Prof Hairil received the National Medical Research Council’s Health Promotion, Preventive Health, Population Health, and Health Services Research Clinician Scientist (Investigator) award this year for his work on CARES-ML. In July 2023, the outcomes of his team’s study on using generative AI for physical status classification will be published in the British Journal of Anaesthesia.
Further, AI algorithms can analyse large amounts of medical data to identify patterns and indicators of diseases at an early stage, enabling timely intervention and improved patient outcomes.
AI algorithms can analyse patient data to predict disease progression, treatment response, and potential complications, enabling healthcare providers to make informed decisions and improve patient care. Also, AI can analyse genetic and patient-specific data to develop personalised treatment plans, allowing for tailored and targeted therapies that increase the chances of successful outcomes.
