AI will be used to reduce patient wait times at University College London Hospitals

University College London Hospitals (UCLH), one of the largest hospitals in London, announced today that it will recruit artificial intelligence to carry out some tasks currently undertaken by doctors and nurses, with the goal of improving emergency room admittance rates, follow-up appointment attendance, and speed for routine tests.

Machine learning algorithms supplied by the Alan Turing Institute will pore over admittance data to track how doctors and patients move through the hospital and identify potential bottlenecks. According to a March survey published by the U.K.’s National Health Service, just 76.4 percent of patients requiring urgent care at London hospitals were treated within four hours — the lowest proportion since 2010, when records began.

UCLH CEO Marvel Levi told the Guardian that a future version of the software might prioritize patients based on the severity of their symptom, such as fast-tracking a person suffering from abdominal pain who is likely to have appendicitis, kidney disease, or another critical ailment.

A second project, which was developed by UCLH clinical research associate Parashkev Nachev, will flag patients who are most likely to miss appointments, taking into account factors such as age, address, and weather conditions, and will automatically text reminders or even reschedule visits. In preliminary trials, the system was 85 percent accurate at predicting whether patients would show up for MRI scans and outpatient clinic appointments.

Finally, UCLH will apply AI to CT scan analyses of 25,000 former smokers and to automatic cervical smear tests.

“Machines will never replace doctors, but the use of data, expertise, and technology can radically change how we manage our services — for the better,” Levi told the Guardian.

The hospital’s effort comes months after London-based Google subsidiary DeepMind, in partnership with the National Health Service and London’s Moorfields Eye Hospital, trained AI to diagnose eye diseases from 3D retinal scans.

In the U.S., hospitals are using AI to aid in the treatment of certain diseases and life-threatening conditions. In January 2017, the Mayo Clinic’s Center for Individualized Medicine teamed up with startup Tempus to analyze the scans of 1,000 patients participating in immunotherapy studies for melanoma, lymphoma, and other forms of cancer. And in September 2016, the Cleveland Clinic announced a collaboration with Microsoft to monitor intensive care unit patients who are at high risk for cardiac arrest.

Industry analysts estimate that the AI health market will reach $6.6 billion by 2021, up from $600 million in 2014. In the U.S., AI is projected to create $150 billion in annual savings by 2026.