New goal achieved for The Adelaide Score

Patients at the Lyell McEwin Hospital were less likely to be re-admitted when their likelihood of discharge was evaluated by an artificial intelligence system, new research has found.

Developed by researchers from the University of Adelaide, the Adelaide Score integrates vital signs and laboratory tests to predict the likelihood of a hospital discharge.

The system was trialled over a period of 28 days in April 2024 at the Lyell McEwin, with patients recording a significantly lower seven-day re-admission rate (five per cent) when compared to April 2023 (7.1 per cent).

More than 2900 patients attended the hospital across the intervention and control periods, with the trial also revealing patients had a shorter median stay (from 3.1 days to 2.9 days) for the same period.

According to findings, published in the ANZ Journal of Surgery from the Royal Australasian College of Surgeons, the reduction in admission lengths saved $735,708 across the 28-day period.

“In this study, the Adelaide Score helped identify patients who may be close to discharge and saves time looking through the electronic medical records to find suitable patients,” says University of Adelaide Medical School affiliate and 2025 Fulbright Future Scholar Dr Joshua Kovoor and first author of the paper.

“By optimising the discharge process, we believe it could create savings, which can then be used on treatments and improving outcomes.

“The work involved in organising a suitable discharge can involve many areas. For example, organising a suitable discharge can involve processes like organising transport, discharge medications, wound plans, and follow-up appointments.

“All of these processes take time, and we can see from the successful intervention that using the Adelaide Score improved hospital inflow.”

Each day of the trial, the Adelaide Score and previously implemented Rosetta System (which analyses the language used in ward notes), evaluated the electronic records of inpatients in 18 surgical and medical teams and identified individuals with a high specificity cut-off for being near discharge. Using the results, patients were ranked by probability of discharge.

“The results were sent to the Supportive Weekend Interprofessional Flow Team (SWIFT) of a highly experienced nurse and pharmacist who work to improve and facilitate weekend discharges on the days they work. This helped them identify patients who they might be able to facilitate the discharge of in the next 24 hours,” says Associate Professor Stephen Bacchi, University of Adelaide Medical School affiliate and 2024 Fulbright Future Scholar, and senior author of the paper.

“The final decision however, went to the respective clinical teams, who made the decision to discharge patients independent of the AI analysis.”

The Adelaide Score is now being considered for implementation across the eastern states.

“With the use of additional weekend discharge support like the SWIFT and AI tools such as The Adelaide Score, we could see the discharge of longer-term, more complex patients becoming more feasible on weekends,” says Associate Professor Bacchi.

“While health is a 24/7 issue, hospital staffing is inherently influenced by factors such as the time of day, and day of week.

“It is well known that aspects of care may vary on weekends, including reduced numbers of discharges. There are likely to be multiple and complex reasons for these differences, but often there are fewer staff rostered on weekends so extra information has the potential to make the discharge process more efficient.