An AI-Enabled Predictive Harm Response Management Algorithmic Tool to Reduce Adverse Events in Health Care Settings (PreHaRM): Protocol for a Three-Phase Model Development and Pilot-Testing Study

Description

Marion Eckert and colleagues publish a 2026 protocol for PreHaRM, an artificial intelligence tool meant to predict patient harm before it happens. Hospitals usually learn about adverse events only after someone files a report, which is too late for prevention. The project aims to turn administrative data already flowing through two South Australian health networks into real-time risk profiles for nurses and midwives.

The work runs in three phases: generating the predictive model, evaluating it, and building a working prototype with an interactive dashboard. The protocol spells out the ethics and governance approvals that linked data use requires. The model will hunt for clinical, workforce, and environmental factors that move harm risk up or down, so that the dashboard points to actions, not just warnings. The pilot will test whether real-time prediction can shift care from reacting to harm toward preventing it.

Catalog record

Source: https://europepmc.org/articles/PMC13254504