Now it's the turn of a new AI-based tool to demonstrate its ability to read doctors' assessments and accurately anticipate the risks of death, hospital readmission and other possible complications.

Created by a team at the Grossman School of Medicine in Langone, New York, the software is now being tested in several of the university's partner hospitals, with the goal of making it a common practice in the medical community in the future.

A study on its possible interest was published Wednesday in the scientific journal Nature.

Its lead author, Eric Oermann, a neurosurgeon and computer engineer at New York School of Medicine, explains that while non-AI-based predictive models have been around for some time, they are rarely used in practice because they require a lot of data capture and formatting.

But there is "one thing that is common to medicine everywhere, it is that doctors take notes on what they see, what they talk about with patients," he said in an interview with AFP.

"So our basic idea was to know if we could start from medical notes as a source of data, and build predictive models from them," he continues.

The predictive model developed, called NYUTron, was formed from millions of medical observations from the records of 387,000 patients treated between January 2011 and May 2020 in hospitals affiliated with the New York university.

These observations included written reports from physicians, patient status notes, X-rays and medical imaging, and recommendations given to patients upon discharge from hospital, all forming a corpus of 4.1 billion words.

One of the main challenges for the software was to successfully interpret the language used by physicians, which varies greatly among professionals, especially in the abbreviations used.

They also tested the tool in real conditions, training it to analyze reports from a hospital in Manhattan and then comparing the results to those of a Brooklyn hospital, with different patients.

By looking at what happened to patients, the researchers were able to measure the number of times the software's predictions turned out to be accurate.

Not a substitute

Disturbingly, NYUTron software identified 95% of patients who died in partner hospitals before a discharge permit, and 80% of those who were readmitted within a month of discharge.

Results that exceeded the predictions of most doctors, as well as those of the non-AI-based computer models currently in use.

But, to everyone's surprise, a very experienced doctor, highly respected in the medical community, gave forecasts "even better than that of the software," said Eric Oermann.

The software also successfully predicted 79% of patients' length of hospitalization, 87% of cases in which patients were denied reimbursement of care by their insurance, and 89% of cases in which the patient suffered from additional pathologies.

Artificial intelligence will never replace the patient-doctor relationship, says Dr. Oermann. But it may make it possible "to provide more information (...) to physicians to enable them to make informed decisions."

© 2023 AFP