Hospital Using AI To Handle Some Tasks Usually Done By Doctors And Nurses

Posted on May 30, 2018 I Written By

Anne Zieger is veteran healthcare branding and communications expert with more than 25 years of industry experience. and her commentaries have appeared in dozens of international business publications, including Forbes, Business Week and Information Week. She has also worked extensively healthcare and health IT organizations, including several Fortune 500 companies. She can be reached at @ziegerhealth or www.ziegerhealthcare.com.

One of the UK’s biggest facilities has announced plans to delegate some tasks usually performed by doctors and nurses to AI technology. Leaders there say these activities can range from diagnosing cancer to triaging patients.

University College London Hospitals (UCLH) has signed up for a three-year partnership with the Allen Turing Institute designed to bring machine learning to bear on care, a project which could ultimately spark additional AI projects across the entire National Health Service. The NHS is the body which governs all healthcare in the UK’s universal health system.

UCLH is making a big bet on artificial intelligence, investing what UK newspaper The Guardian describes as a “substantial” sum to develop the infrastructure for the effort.

UCLH officials believe — like other health organizations around the world — that machine learning algorithms may someday diagnose disease, identify people at risk for serious illness and more. Examples of related projects abound. Just one case in point is a project begun in 2016 by New York-based Mount Sinai Hospital, which launched an effort using AI to predict which patients might develop congestive heart failure and offer better care to those who have already done so.

Professor Bryan Williams, director of research at University College London Hospitals NHS Foundation Trust, said the move will be a “game changer” which could have a major impact on patient outcomes. “On the NHS, we are nowhere near sophisticated enough,” Williams told The Guardian. “We’re still sending letters out, which is extraordinary.”

UCLH’s first AI effort, which is already underway, is intended to identify patients likely to miss appointments. Using existing data, including demographic factors such as age and address plus outside factors like weather conditions, researchers there have been able to predict with 85% accuracy whether the patient will show up for outpatient visits and MRIs.

Another planned project includes improving the performance of the hospital’s emergency department, which, like many NHS hospitals, isn’t meeting government performance targets such as maximum four-hour wait times. “[This is] an indicator of some of the other things in the entire chain concerning the flow of acute patients in and out of the hospital,” UCLH chief executive Professor Marcel Levi told the newspaper.

The hospital envisions solving its wait-time problem with machine learning. Drawing on data taken from thousands of patients, machine learning algorithm might be able to determine whether a patient with abdominal pain suffers from severe problem like intestinal perforation or a systemic infection, then fast-track those patients. This kind of triage is generally performed by nurses in hospitals around the world.

That being said, the partners agree that machine learning performance must be incredibly accurate before it has any major role in care. At that point, it will be ready to support clinicians, not undercut them. According to Professor Chris Holmes of The Alan Turing Institute, the whole idea is to let doctors do what they do best: “We want to take out the more mundane stuff that’s purely information driven and allow time for things the human expert is best at.”