The scientists hope the system could enable better end-of-life care for hospital patients with a terminal disease.
The computer program uses AI to achieve its high-level of accuracy and after an analysis of electronic health records.
The AI program looked at 160,000 adult and child patient files from Stanford and Lucile Packard Children’s hospital.
To improve the systems accuracy it looked at patient information like diagnosis, what procedures were performed, patient scans and what medicines had been taken.
Anand Avati, a member of the AI Lab at Stanford University, said: “The scale of data available allowed us to build an all-cause mortality prediction model, instead of being disease or demographic specific.”
When the algorithm was asked to predict which of 40,000 patients would die in next three to twelve months, it was correct in predicting when the people would die in 90 per cent of the cases.
The Stanford researchers published their findings in a paper titled “Improving Palliative Care with Deep Learning”.
The AI system performed well in its pilot run but researchers want to continue to improve the program before it could be rolled out to hospitals.
For this next stage to happen, the program would need to have access to more data to produce better and more accurate results.
Kenneth Jung, a research scientist at Stanford University, said: “We think that keeping a doctor in the loop and thinking of this as ‘machine learning plus the doctor’ is the way to go as opposed to blindly doing medical interventions based on algorithms… that puts us on firmer ground both ethically and safety-wise.”