Truace
Subscribe

Artificial Intelligence Prognostic Models for Healthcare-Associated Infection Prediction in Adult Intensive Care Units: A Systematic Review Protocol

Background and aims Despite widespread Infection Prevention and Control (IPC) measures, Healthcare-Associated Infections (HAIs) remain a major concern in Intensive Care Units (ICUs), leading to increased mortality, prolonged hospital stays, and additional healthcare costs. Prognostic prediction models support identification of patients at high infection risk, and Machine Learning and Artificial Intelligence (ML/AI) methods offer the potential to improve such predictions. However, no systematic review has specifi…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1846· cite & verify
Artificial Intelligence Prognostic Models for Healthcare-Associated Infection Prediction in Adult Intensive Care Units: A Systematic Review Protocol
US Navy 030423-N-6967M-088 The Intensive Care Unit (ICU) night shift aboard USNS Comfort (T-AH 20) goes over the history of each patient by U.S. Navy photo by Photographer's Mate 1st Class Shane T. McCoy.. Public domain

In brief

Researchers have registered a plan for a systematic review of machine learning and AI models that estimate which adult intensive care patients are likely to develop infections acquired in hospital. The team will search three major databases and have three reviewers independently screen studies and extract data.

The work matters because infections in intensive care are linked to higher mortality, longer stays and extra costs, and AI predictions are proposed to help target prevention. Because this is only a protocol, no findings exist yet, so how reliable or ready these models are remains unknown.

Main points

  1. Healthcare-associated infections in ICUs are tied to higher mortality, longer hospital stays and added costs despite infection prevention measures.
  2. The review will cover supervised ML/AI models predicting any HAI occurring after adult ICU admission.
  3. Risk of bias will be assessed with PROBAST + AI, and reporting quality with the SPIN-Prediction Models framework.
  4. The protocol is registered on the Open Science Framework and uses a narrative synthesis.

The problem

Healthcare-associated infections persist in adult ICUs despite prevention measures, and no systematic review has yet examined the quality or readiness of ML/AI models meant to predict them.

The rundown

The protocol states that no earlier systematic review has focused on ML/AI prognostic models for HAI prediction in adult ICU populations, which is the gap it aims to fill. It will identify models and examine their methodological characteristics, risk of bias and reporting quality.

The search will use PubMed, Scopus and Web of Science. Three reviewers will work independently on screening, data extraction and bias assessment, collecting details such as predictors, sample size, missing data, analysis methods, performance, and model presentation and interpretation.

The authors expect the findings to inform researchers, clinicians and infection prevention decision-makers about the reliability, applicability and readiness for clinical use of these models.

Sources

  1. Peer-reviewedHealth Science Reports2026-10-05

The debate