NEX, a contraction of nexus (Latin, n.); meaning a connection or series of connections linking two or more things.

Clinical Report: Predictive Infection Surveillance in a 1,200-Bed National Military Hospital


A modified version of a featured talk at ESCMID 2025, Vienna, given by Dr. Vasikasin on April 13th 2025.
Report Date: 10th April 2025

Highlights

A recent clinical study has demonstrated that the NEX Intelligence platform achieves state-of-the-art performance in real-time infection surveillance and early prediction of multidrug-resistant organisms — with an AUC-ROC of 0.89 and 87% sensitivity for predicting CRE acquisition up to seven days in advance.

A Clinical Study Supported by:

Background

The infection prevention and control (IPC) team at Phramongkutklao Hospital, a 1,200-bed military hospital in Thailand, has historically relied on an entirely manual surveillance system. This system required retrospective reviews to identify and collate hospital-acquired infections, making the early identification of infection transmission patterns and rapid response incredibly challenging.

Deployment

In January 2025, the hospital partnered with NEX to deploy an AI-driven infection surveillance system. The project focused on building real-time pipelines for the detection and monitoring of key healthcare-associated infections, including Multi-Drug-Resistant Organisms (MDROs), Bloodstream Infections (BSIs), and Device-associated infections (DAIs).

The deployment was integrated directly with live microbiology feeds and electronic patient records, creating a centralised, automated platform for IPC teams to track, manage, and prevent infections.

Results feed tracking and tagging new microbiological records with MDRO, BSI, and DAI labels for review and surveillance from the NEX Infection Centre.

Application results feed tracking and tagging new microbiological records with MDRO, BSI, and DAI labels for review and surveillance.

Early Results from Hospital Deployment

The NEX system demonstrated rapid impact within weeks of implementation—digitising infection records, enabling real-time data visibility, streamlining workflows, and receiving strong frontline feedback.

"We finally have real-time visibility across the entire hospital. It has dramatically reduced time-consuming tasks and is already helping us prioritise more effectively and prepare faster."

Dr. Vasin Vasikasin, Assistant Professor and Consultant in Infectious Diseases at Phramongkutklao Hospital, Honorary Research Fellow at Imperial College London

John Rush

Predicting Infection Acquisition for Targeted Prevention

We are in the process of validating our early warning models, designed to predict the acquisition of Carbapenem-Resistant Enterobacterales (CRE) — one of the most clinically significant multidrug-resistant organisms in hospital settings (Centers for Disease Control and Prevention).

The system generates live daily risk scores for all inpatients, identifying those at highest risk of acquiring CRE before a positive microbiological result is confirmed. These scores are calculated using live data feeds, drawing on a range of patient-specific variables, including microbiology history, antimicrobial usage, device exposure, and ward movements.

NEX Infection Intelligence identifies high-risk contacts exposed to a source CRE case, with predicted risk of testing positive.

Feature image

Early validation results over six months of prospectively generated data suggest strong predictive performance, with an AUC-ROC of 0.89, sustained across a prospective dataset including both clinical and screening cultures (Vasikasin, 2025). The model enables risk stratification up to seven days before confirmed acquisition.

  • 0.89 AUC-ROC

    Performance in predicting colonisation with CRE among clinical and screening cultures.

  • 7-days ahead

    Providing predictions seven days ahead of confirmed laboratory results.

  • 87% sensitivity

    High sensitivity in the prediction of positive versus negative results.

Conclusions

This collaboration represents the growing transition from retrospective audits and IPC surveillance to real-time, data-driven, more anticipatory infection control. The project has demonstrated the feasibility and utility of automated surveillance in a large, high-volume hospital environment with electronic health records — streamlining IPC workflows, improving situational awareness, and laying the foundations for predictive infection prevention.

NEX continues working closely with frontline IPC teams to ensure the system delivers actionable intelligence — where and when it’s needed most!

References

  1. Centers for Disease Control and Prevention. About CRE: Carbapenem-Resistant Enterobacterales. 3 Aug. 2023, https://www.cdc.gov/cre/about/index.html. Accessed 10 Apr. 2025.
  2. Vasikasin, D. Prospective Evaluation of a Machine Learning-Based Real-Time Multidrug-Resistant Organism Prediction System. Featured talk, ESCMID Global Conference 2025, 13 Apr. 2025, Vienna, Austria.

Curious how it all works?

Explore the technology, research, and real-world impact—and see why the future of infection prevention starts here.

Feature image
Built on Unicorn Platform
English 🇺🇸🇬🇧