Can AI help nurses improve patient safety?

3 minute read

A new study from the AIHI at Macquarie University is testing a simulated AI model to provide nursing decision support to eliminate adverse events.

Researchers from the Australian Institute of Health Innovation are looking for registered nurses to spend 45 minutes answering an online survey about the role AI could play in nursing decision support. 

Professor Farah Magrabi told HSD the study was centred around the question of inpatient adverse events. 

“Essentially, one in 10 patients that goes into hospital experiences harm in the process of delivering care,” she said. “And around about a quarter of those safety events are preventable. 

“When a nurse has a patient in hospital, they need to assess risks for that patient. They need to decide, first and foremost, which are the most appropriate risk assessments to do – they pick and prioritise. 

“Then they need to complete the risk assessment form. Each risk has its own form. And when they are completing it, they often need to enter the same information in different forms.  

“The other issue is that a patient’s status changes as they’re in hospital – they don’t stay constant. So often the nurses need to remember to go back and reassess risks.  

“That’s another aspect that we’re looking at – when the patient status changes, is it possible for an AI tool to remind nurses to complete those risk assessments.” 

The study is recruiting now and can be found online here

The Third Degree with Dr Priit Tohver

Big wins for healthcare workers in two states

NSW Health secures 1100 nursing and midwife jobs

To be eligible, the nurse must have worked in an Australian hospital as an RN in the last two years (adult, non-critical units only) or be a final-year student of a nursing degree program or graduate who was awarded an Australian nursing degree in 2022 or later. Nurses working in critical care, emergency and paediatric settings, and year 1 and year 2 nursing students are not eligible. 

Participants are asked to review and complete six patient scenarios using an AI-based simulated decision support system. The system uses cookies to accumulate the data entered. The study takes around 45 minutes to complete and is not a test of clinical knowledge or computer skills.  

“We are interested in assessing how this AI system performs,” said Professor Magrabi. 

“Because it’s an online study, it’s doing it in a patient-free and a risk-free setting.  

“Technically, it is possible to design these kind of AI systems, but we don’t know whether it’s going to be useful to nurses. Essentially, we are looking to see whether having something like this would actually affect decision making of nurses.  

“The study is driven by the problem on the ground. If something like this works, then it’s not hard to build it technically.” 

Do you have a story tip for us, or a topic you would like to see us cover? Contact the editor at

End of content

No more pages to load

Log In Register ×