AI could be used to predict the type of health conditions a patient is likely to develop in the future, according to a new study.
The AI tool, named Foresight, has been trained using information from NHS records and could be used by doctors to aid in diagnosis decisions.
The AI tool was developed by researchers from University College London, King’s College London, St Thomas’ NHS Foundation Trust and King’s College Hospital NHS Foundation Trust.
Foresight is built on OpenAI’s GPT-pipeline and uses a patient’s health status and general clinical history to forecast and simulate future events.
Researchers trained three separate Foresight models using data from over 811,000 patients. The data was taken from two NHS trusts and a publicly available dataset in the US.
“The proposed purpose of Foresight is not to enable patients to self-diagnose or predict their future, but it could potentially be used as an aid by clinicians to make sure a diagnosis is not missed or for continual patient monitoring for real-time risk prediction,” said Zeljko Kraljevic, researcher in health informatics, biostatistics and health informatics at King’s College London.
Based on the NHS and US records, the models were given 10 possible disorders a patient may likely experience next.
According to the study, the AI tool correctly identified the health condition 68% to 76% of the time using NHS data. When using US data, the AI tool was accurate 88% of the time.
“It is an exciting time for AI in healthcare, and to develop effective tools we must ensure that we use appropriate data to train our models and work towards a shared purpose of supporting healthcare systems to support patients,” said senior author Richard Dobson, a professor of medical informatics at King’s College London and University College London.