The Adelaide Score: prospective implementation of an artificial intelligence system to improve hospital and cost efficiency

Joshua G. Kovoor, Brandon Stretton, Aashray K. Gupta, Alexander Beath, Mathew O. Jacob, John M. Kefalianos, Gavin J. Carmichael, Ammar Zaka, Gerry O'Callaghan, Shrirajh Satheakeerthy, Andrew Booth, Thomson Delloso, Thomas J. Hugh, Weng Onn Chan, Guy J. Maddern, Eva Balan-Vnuk, Michael Cusack, Toby Gilbert, John Maddison, Stephen Bacchithe Adelaide Score Advisory Group, Matthew Arnold, S. George Barreto, Tarun Bastiampillai, Jayme Bennetts, Niranjan Bidargaddi, Jonathan Clarke, Christopher Dobbins, Lukah Dykes, Kerrie Freeman, Ashley Hopkins, Sarah Howson, Hubertus Jersmann, Matthew Marshall-Webb, Nasim Nematzadeh, Robert Padbury, Wendy Rogers, Minh Son To, Andrew Vanlint, Kathryn Zeitz

Research output: Contribution to journalArticlepeer-review

Abstract

Background: The Adelaide Score is an artificial intelligence system that integrates objective vital signs and laboratory tests to predict likelihood of hospital discharge. 

Methods: A prospective implementation trial was conducted at the Lyell McEwin Hospital in South Australia. The Adelaide Score was added to existing human, artificial intelligence, and other technological infrastructure for the first 28 days of April 2024 (intervention), and outcomes were compared using parametric, non-parametric and health economic analyses, to those in the first 28 days of April 2023 (control). Artificial intelligence evaluated inpatients admitted under 18 surgical and medical teams, and patients of high likelihood of discharge were provided, on working shifts between Thursday to Sunday, to the Supportive Weekend Interprofessional Flow Team (SWIFT) comprising a senior nurse and pharmacist. 

Results: Two thousand nine hundred and sixty-eight admissions were included across intervention and control periods. Relative to the control group, use of the Adelaide Score in the intervention group resulted in significantly shorter median length of stay (3.1 versus 2.9 days, P = 0.028) and significantly lower seven-day readmission rate (7.1 versus 5.0%, p = 0.02). The 0.2 bed-day reduction in median length of stay produced a cost saving of $735 708.60 across the 28-day period, or $9 564 211.80 across a 52-week year. There was no significant difference between intervention and control groups in median length of stay for patients discharged on weekends, in-hospital mortality, or discharge to non-home destinations. 

Conclusions: The prospective implementation of the Adelaide Score was associated with improved hospital and cost efficiency, alongside lower readmissions, for patients across surgical and medical services.

Original languageEnglish
Number of pages8
JournalANZ Journal of Surgery
DOIs
Publication statusE-pub ahead of print - 3 Jan 2025

Keywords

  • artificial intelligence
  • cost saving
  • efficiency
  • the Adelaide score

Fingerprint

Dive into the research topics of 'The Adelaide Score: prospective implementation of an artificial intelligence system to improve hospital and cost efficiency'. Together they form a unique fingerprint.

Cite this