Other Titles

PechaKucha Presentation

Abstract

Introduction/Aim: The size, distribution, and skill mix of the nursing workforce are recognised drivers of patient outcomes, yet limited evidence exists on how workforce, clinical, and environmental variables interact to prevent harm. Despite extensive reporting systems, metropolitan hospitals continue to see rising rates of falls, medication errors, and violence, events typically identified only after they occur. This underscores the need to shift from retrospective surveillance to predictive, data-enabled harm prevention. The PreHaRM project, ‘Predictive Harm and Response Management’, aims to deliver this shift by implementing a predictive harm algorithm across South Australia’s two largest health networks, with potential for national scale-up. Its purpose is to equip clinicians with early risk insights that support proactive intervention.

Methods: Sub-project 1, now completed, developed the predictive risk model by identifying priority harm outcomes and mapping contributing factors across clinical, workforce, and environmental domains. Drawing on both conventional data (safety learning systems, adverse event datasets) and non-conventional sources (nursing rosters, ward-level staffing, patient flow), the project used data mining, machine learning, and multivariate modelling to generate and validate the algorithm. The model demonstrated capacity to detect emerging risk patterns before harm occurs.
With the algorithm established, Sub-project 2 focuses on real-world implementation. Current activities include algorithm refinement and co-design of an interactive clinical dashboard to support workflow integration. The dashboard will embed predictive outputs directly into clinicians’ daily practice, enhancing situational awareness and enabling timely, data-driven decisions. Evaluation will assess usability, predictive accuracy in live environments, and impact on harm reduction.

Results: The completed modelling phase has produced a validated predictive harm algorithm capable of identifying risk signals across workforce and clinical variables. The next phase will translate this into an operational digital tool through clinician-led dashboard co-design.

Conclusion: By embedding predictive analytics into routine care, PreHaRM enables a shift from reactive reporting to proactive harm prevention, with potential to reduce falls, medication errors, and violence in healthcare settings.

Notes

Presenter notes available in attached slide deck. To see the notes in Adobe Acrobat, go to Tools > Comment or look at your Layers Panel. If the notes were saved as comments or layers, you can toggle them visible. 

References:

Choi, J.-H., Kim, D.-W., et al. (2023) ‘In-hospital fall prediction using machine learning algorithms and the Morse Fall Scale in patients with acute stroke: a nested case-control study’, BMC Medical Informatics and Decision Making, 23:246. doi: 10.1186/s12911-023-02330-0.

Dobbins, N.J., et al. (2024) ‘Deep learning models can predict violence and threats against healthcare providers using clinical notes’, npj Mental Health Research. doi: 10.1038/s44184-024-00105-7.

Marlow, N., Eckert, M., Sharplin, G., Gwilt, I. & Carson-Chahhoud, K. (2023) ‘Graphical User Interface Development for a Hospital-Based Predictive Risk Tool: Protocol for a Co-Design Study’. JMIR Research Protocols, 12: e47717. doi: 10.2196/47717.

O’Connor, S., Gasteiger, N., Stanmore, E., Wong, D.C. & Lee, J.J.J. (2022) ‘Artificial intelligence for falls management in older adult care: a scoping review of nurses’ role’, Journal of Nursing Management. doi: 10.1111/jonm.13853.

Description

Embedding predictive analytics enables nursing practice to be supported by emerging technology, by enhancing situational awareness, supporting safer workloads, strengthening clinical decision-making, and empowering nurses to prevent harm proactively rather than respond retrospectively. This technological development will result in fewer adverse events, optimisation of data utilisation and reduced harm to the consumer and staff.

Author Details

Prof. Marion Claire Eckert, PhD; Greg Sharplin, MPsych (Org); Lachlan Darch*, MEXSc; Prof. Georg Grossman*, PhD

*Listed in Sigma event system, but not listed in attached slide deck.

Sigma Membership

Psi Alpha at-Large

Type

Presentation

Format Type

Text-based Document

Study Design/Type

Other

Research Approach

Other

Keywords:

Implementation Science, Workforce, Acute Care, Adverse Health Care Event, Adverse Health Care Event--Prevention and Control, Artificial Intelligence

Conference Name

37th International Nursing Research Congress

Conference Host

Sigma Theta Tau International

Conference Location

Toronto, Ontario, Canada

Conference Year

2026

Rights Holder

All rights reserved by the author(s) and/or publisher(s) listed in this item record unless relinquished in whole or part by a rights notation or a Creative Commons License present in this item record. All permission requests should be directed accordingly and not to the Sigma Repository. All submitting authors or publishers have affirmed that when using material in their work where they do not own copyright, they have obtained permission of the copyright holder prior to submission and the rights holder has been acknowledged as necessary.

Review Type

Abstract Review Only: Reviewed by Event Host

Acquisition

Proxy-submission

Date of Issue

2026-08-30

Click above link to access the slide deck.

Share

COinS
 

An AI Driven Harm Response Management Tool to Prevent Adverse Events in Healthcare

Toronto, Ontario, Canada

Introduction/Aim: The size, distribution, and skill mix of the nursing workforce are recognised drivers of patient outcomes, yet limited evidence exists on how workforce, clinical, and environmental variables interact to prevent harm. Despite extensive reporting systems, metropolitan hospitals continue to see rising rates of falls, medication errors, and violence, events typically identified only after they occur. This underscores the need to shift from retrospective surveillance to predictive, data-enabled harm prevention. The PreHaRM project, ‘Predictive Harm and Response Management’, aims to deliver this shift by implementing a predictive harm algorithm across South Australia’s two largest health networks, with potential for national scale-up. Its purpose is to equip clinicians with early risk insights that support proactive intervention.

Methods: Sub-project 1, now completed, developed the predictive risk model by identifying priority harm outcomes and mapping contributing factors across clinical, workforce, and environmental domains. Drawing on both conventional data (safety learning systems, adverse event datasets) and non-conventional sources (nursing rosters, ward-level staffing, patient flow), the project used data mining, machine learning, and multivariate modelling to generate and validate the algorithm. The model demonstrated capacity to detect emerging risk patterns before harm occurs.
With the algorithm established, Sub-project 2 focuses on real-world implementation. Current activities include algorithm refinement and co-design of an interactive clinical dashboard to support workflow integration. The dashboard will embed predictive outputs directly into clinicians’ daily practice, enhancing situational awareness and enabling timely, data-driven decisions. Evaluation will assess usability, predictive accuracy in live environments, and impact on harm reduction.

Results: The completed modelling phase has produced a validated predictive harm algorithm capable of identifying risk signals across workforce and clinical variables. The next phase will translate this into an operational digital tool through clinician-led dashboard co-design.

Conclusion: By embedding predictive analytics into routine care, PreHaRM enables a shift from reactive reporting to proactive harm prevention, with potential to reduce falls, medication errors, and violence in healthcare settings.